All Episodes

July 24, 2023 116 mins

Imagine transforming your understanding of Network Attached Storage (NAS) from a complex, jargon-filled mystery to an easy-to-grasp concept. Our special guest, Robbie Andrews from the YouTube channel NASCompares, is here to help us navigate this complex landscape. Robbie's knack for breaking down complex topics in an engaging, whimsical way permeates through this long-form conversation.

From the intricacies of storage types and hard drives, we navigate towards the optimal configurations for a NAS,  the integration of NVMEs in NAS systems, and then the conversation takes an intriguing turn towards the shadowy, dark world of ransomware attacks and the notorious Deadbolt ransomware attacks on QNAP.

Wrapping up this enlightening chat, we pit Synology and QNAP against each other, debating the merits and drawbacks of both. So, gear up and join us for a deep dive into the exciting, sometimes daunting, yet always fascinating world of NAS with Robbie of NASCompares.


Send us a text

Digital Disruption with Geoff Nielson
Discover how technology is reshaping our lives and livelihoods.

Listen on: Apple Podcasts   Spotify

Other Ways to Connect with the Uncast Show


Mark as Played
Transcript

Episode Transcript

Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Ed (00:08):
Hi there, robbie.
Thank you very much for joiningus on the Uncast Show.
It's great to have you here.
The first thing I'd like to askyou really for people who don't
know you, you run a YouTubechannel called NAScompares.
Can you share how your interestfirst started into network
attack storage and how youdecided to start up the website
and the YouTube channel?

Robbie (00:30):
I mean.
Well, much like most coolpeople out there, it all stems
from gaming.
Originally, I would sort ofrepurpose I mean we're talking
like late teens, early 20s.
We would repurpose old WindowsXP systems as working as
Counter-Strike source andSoldier of Fortune, jedi Knight
on a rough day, and turn it intolike servers and so from there

(00:52):
that was where a lot of myinterest came from, the kind of
networking and storage bandedtogether.
And then around 2007, 2008, Istayed working for an e-retailer
that dealt with storagepredominantly, and from there
that was kind of where the nutsand bolts of it was organically
sort of thrown at me and a lotof it was coal face type stuff,
and then there was a lot ofbuild stuff until eventually it

(01:12):
sort of gave way to justmarketing it just because one of
my keenest interests.
So I know it sounds a bit outthere, but this is an incredibly
technical subject for the mostpart that those who don't know
what they're talking about.
They would find it too dull tocare or it's just not profitable
for them to care.
And half the motivation at thebeginning, the other half being

(01:33):
being on camera, hello.
The first half of themotivation was always to take an
incredibly complex subject andmake it crayon and chewable
friendly.
So that's kind of where it allstarted with the blog.
It was taking even somethinglike a raid and trying to talk
about the different kinds ofstorage pools and volume
allocation and I scuzzy,lonesome and basically turning

(01:53):
it into crayon Lego English andthat's kind of how it started.
And then we suddenly realizedhow much of that could actually
be translated into easy andthat's what effectively every
video has become.
Every time we tackle a subject,we try to keep it as chewable,
easy and understandable aspossible, because there are lots

(02:13):
and lots of platforms out there, bigger and, I would argue,
smarter than me, that tacklethis subject with a sense of
technical nuance I do not have.
But if you're starting at stepone, two, three or four, all
these other platforms that startat platform nine, 10, 11, 12
are intimidating beyond measureand whether you are a small

(02:34):
post-production facility, awedding photographer, or you are
a multi-site deployment, tryingto hire a system admin to learn
these things is tough, andthat's basically where this all
started and that's where we arenow just taking incredibly
difficult to understand subjectssometimes and making them
chewable.

Ed (02:52):
Yeah, definitely.
It's always nice to be able tolook at a subject and not always
have the person explaining thesubject assume that you know 50%
about it already.
So being able to kind ofactually come into it and be
able to learn from the ground up, really I think that's really
awesome.

Robbie (03:11):
I mean a lot of that.
I would say I've just interjecteven my early understanding of
unraided.
A lot of that comes from yourown videos, because not a
dissimilar approach that youhave to unraided.
If you look online aboutunraided, there is extensive
guides and stuff online but alot of them miss out steps one
through seven very early doorsand even some of your earliest
stuff.
I remember one of the firstvideos that yours overwatched

(03:33):
was the one where you werediscussing how the parity works
within the storage scalabilityof unraided and it was
incredibly user friendly and youhad these lovely graphics on
screen that I swear a five yearold could follow, which very few
people could say that aboutunraided tutorials online.

Ed (03:51):
So you know what came first for you.
Was it the website or theYouTube channel?

Robbie (03:55):
The website came first when I was working for the said
e-retailer.
At that point I was working intheir marketing team and they
wanted we were working sort ofwith them to talk about the
product.
But there were subjects Icouldn't talk about and couldn't
cover within that spectrum thatI wanted to cover but you
couldn't cover within ane-retailer perspective.

(04:16):
So then that's where I wentinto kind of blogging there on
the side and then that sort ofgave way to everything else that
came afterwards and that's kindof where it all began.

Ed (04:26):
We were talking a couple of weeks ago when we were arranging
to do the podcast, you were offto.
Was it Taiwan?

Robbie (04:31):
you were off to yeah, taipei for Computex.
Second time I've went, I washoping to go more frequently to
cover that event and that's kindof that's R3, if you will.
I mean, ces is the world's kindof access to a lot of the
technology and the new stuffwould they bring up?

Ed (04:47):
If I can just kind of butt in sorry for one moment the
people that don't know whatComputex is.

Robbie (04:52):
It's.
It's effectively it's a tradeshow that shows off a lot of the
new hardware and technicalinnovations coming out from the
East and again, a lot of thoseare either brand specific or
kind of the back end things,your real techs, your envidia's,
all these ones that work withso many different partners and
they're showing off theirinnovations, which will

(05:12):
eventually be taken up andworked forward with afterwards.

Ed (05:16):
So did you see any kind of really cool new things that
really took your eye while youwere there?

Robbie (05:21):
Professionally and personally.
Personally, I got served adrink by a robot which, I'll be
honest, 13 year old me lost hismind, but a lot of it was just
kind of the integration of Wi-Fi7.
It's like real, actual kind ofprogression on that.
Then there was a lot of stuffto do, is obviously AI is king,
you know from your chat GPTs toMPU processors being integrated

(05:45):
into more and more devices andthere was a lot more real world
applications being shown off,kind of model home stuff and
it's obviously the pandemic andsemiconductor shortages and a
myriad of socio politicalfactors have slowed down this
subject a lot.
But Computex this year andInovex running at the same time,
we just saw a lot more of thatintegrated home and professional

(06:10):
AI integration come to theforefront and actually be shown
off in a relatable fashion.
So that was a privilege, thatwas a standout for me.

Ed (06:19):
Well, switching gears a bit and kind of going back to
talking about NAS.
I'm sure there's some listenersout there who maybe don't even
know what a NAS is.
I'm sure some of my old friendsstill think NAS is a 1990s
rapper.
So can you basically explainwhat a NAS drive actually is?

Robbie (06:39):
At its most basic level and I mean fantastically basic
it's a hard drive you accessover the network or the internet
, and even then I've not made itthat basic.
It's ultimately a storage devicethat can be accessed via any
device in your home or businessand also has the potential to be
accessed remotely when you wantit.
But more often than not theword NAS is more heavily

(07:01):
associated with a turnkeysolution and although obviously
there is open sourcealternatives and build your own
alternatives, those are moreoften referred to colloquially
as servers, even though it willbe a DIY server, diy NAS,
whereas at the flip side, nasmore often than not gets

(07:21):
associated with paid turnkeyhardware, software, combined
solutions, and of course there'sa huge element of ambiguity
there in the middle.
But ultimately that's what aNAS is.
My cat has decided to jump inbecause she's a fame to log, but
yeah, for me that's what a NASis and I think at its entry
point it's a backup device.
After that, everything else,from surveillance to virtual

(07:45):
machines to multimedia they'rejust extras, but ultimately it's
a single storage device that'saccessible via a myriad of often
wireless or wired means.

Ed (07:56):
And why would someone prefer to have a NAS than use
something like OneDrive orDropbox?

Robbie (08:00):
It's a very, very good point.
A lot of that come.
I mean, there's lots of smallreasons that are and we moved
the cat.
There are lots of reasons thatare more specific in users, like
performance, but a lot of itcomes down to ownership.
So I know that a NAS in mystudio, much like my cat, I can

(08:20):
get hands on it and, before Ifire her into the sky, I know
that a cloud although it mightbe cloud storage device, which
has its place, and I do thinkNAS and cloud synchronization
and having both as a cloudgateway is useful to a lot of
users.
And NAS is 100% disconnectablefrom everyone.

(08:41):
Cloud drive, like OneDrive,dropbox, google Drive, I can be
99.9% certain that'sdisconnected.
The same goes for deletion.
I can be up to, if I've got ahandy volcano nearby, 100%
certain of destroying the dataon there, whereas that's
something that's lacking on acloud service provider.
And finally, when users do buya cloud device by a cloud

(09:05):
subscription, they're kind ofplaying this game where you get
a Google Drive and you know thatin three years of that
subscription you are going topay X amount of money.
That data has to be somewhere inthree years anyway, so then you
have to continue to pay, butyou're increasing your data all
the time.
So then you may have to go fora larger tier and you're more in

(09:26):
the hole and eventually youneed to remove that data from
the cloud.
So when you remove it, where'sit going to be?
On a storage device like a NAS.
So if that's what's going tohappen in a three to five year
forecast, just buy the NAS inthe first place, because you've
just spent all of this money inthe interim anyway and there are
downsides.
You know electricity costs, youknow can cost a penny here, a

(09:48):
pound there, and then on top ofthat, replacement drives and
raid functionality is going tocost more up front.
But I want to leave it to costthat you acknowledge you're
going to have to engage with inthree to five years.
Then a NAS becomes the betterchoice and also, of course,
control of your data.

Ed (10:06):
That's a really, really interesting thought.
I've never thought of the factthat when someone has got all
their stuff on Google Drive orOneDrive, one day they are going
to have to move it somewhereelse or continually pay forever,
and so the time they actuallywant to have the data themselves
, they are going to have to putit somewhere or just say, well,

(10:28):
I don't want that data anymore,I'll just delete it.
But I've never actually thoughtabout that as a reason not to
use them.
My reasons have always beenkind of privacy and just not
wanting, like you say.

Robbie (10:39):
You know, when you delete the data has it really
been deleted and even legallywith a lot of businesses you're
kind of in many regards you'reon a bound to keep hold of data
for an extended period of time,so you're kind of locked in that
you can't leave.
You can't just walk away andleave your data.
And indeed at the Excel Centrein London they do an event there

(11:01):
called IP Expo, and there's acompany that all they do is
destroy hard drives.
They give it to do a mini EMPon a selection of drives.
They then wipe it, index it andthen crush it in front of you
and then turn it into likerecycled toys and stuff.
And this industry exists becauseof so many businesses, can't?
They have to be 100% certaintheir data is destroyed or

(11:23):
completely disconnected.
And there's all thesebusinesses that are in this.
They're stuck constantly inthis subscription model because
not only if they created so muchdata that even just removing it
from the cloud is troublesomein terms of scaling, but also
the sheer volume of data meansthey've lost track of what is
integral, what is missioncritical and what is chaff.

(11:44):
And again, within a NAS youhave a great degree of control
over that very early doors.

Ed (11:51):
And talking about differences between NAS, how
about direct attached storage?
Can you explain the differencesbetween a NAS and DAS is it
called, I don't know and a SAN?
Can you explain the kind ofdifferences in these type of
storage technologies please?

Robbie (12:08):
Well, when it comes to, I mean, nas, we've sort of gone
through their direct attachedstorage is kind of a one to one
and you can't really multi-useraccess it, and SAN lives
somewhere in the middle.
Plus there's this phenomenalperformance potential with the
way the block transfers are andthe way you can kind of scale
the area of storage to startwith to get higher performance

(12:30):
in most regards than you wouldin traditional NAS protocol.
But you also have themulti-user access and the
multi-client connection, be theyan actual physical client, a
user or just a software openingup an exchange.
That is just not possible on aDAS.
So that's why, when you get tothe enterprise level tier that's
why SAN becomes more desirable.

(12:50):
It's slightly not quite as highfor loot as it used to be
because NAS becomes increasinglymore efficient in its exchanges
.
But definitely SAN in themiddle between the two of them
appeases a lot and ultimatelywhy it becomes considerably more
desirable at the enterprisetier.

Ed (13:08):
If I've got it right, so direct attached storage would be
something like a USB hard driveplugged into your laptop, so
that's basically block storageplugged into a computer, and SAN
is block storage that can beshared over a network to a
computer and it thinks it's ahard drive plugged in.
Am I correct?

Robbie (13:29):
Yeah, I mean again, there's a lot of things you can
do when it comes to SAN levelconnectivity to dupe your local
system into or dupe is a strongword but to let your client apps
and devices treat it like alocal storage device.
And then, of course, on theactual I guess it kind of the
lun in the SAN end of the serveryou're utilizing you can be

(13:52):
incredibly creative about howyou want the storage to be built
.
So if you're going to be havinga large volume and large
frequency database, so that's,you know, a million squillion
tiny files, you're going to wantto make sure that your
provisioning is built aroundthat in a way that you can't
really do with an external drive, because with external drives

(14:13):
the amount of customization andcontrol you know, is it going to
be fat 32?
Well, it isn't.
Well, there you go.
I've got two options left andit's kind of like that.
You've got so much moreflexibility on SAN than that.
But unfortunately SAN is alsofantastically complex and if you
set it up wrong at thebeginning, you can't just it's
not easy to do a do over withthat and even with the

(14:34):
scalability of SAN, so you cancreate a SAN target for storage
and then go.
I want that to be a billionsquillion terabytes.
I've only got 20 terabytes butI plan to scale up later and you
can create this imaginarytarget for you to scale up
towards all you can just use tolock it in and then there are
performance benefits anddetriment to either one of them.
But again it just comes down toflexibility at the enterprise

(14:58):
level.
That kind of unified storagetier really benefits from that.

Ed (15:03):
So for us home users and small business users, just
having access to shared filesthrough a NAS, you know that's
kind of ideal for us.
So inside our NAS.
Can we talk a little bit abouthard drives please, robbie,
absolutely.
I wondered if you could help meunderstand the differences
between the different types ofhard drives.

(15:24):
You know, when you go on placeslike Amazon that kind of place
you'll see it might sayenterprise drive, you might see
CCTV drive, performance desktopdrive or a NAS specific drive.
Do we really need NAS specificdrives and what's the difference

(15:45):
between all of these differenttypes and does it matter?

Robbie (15:49):
I think if you'd asked me that question back in maybe
2005 or 2006, the answer wouldhave been daaah, but
unfortunately the real answer isa lot more nuanced now and,
like any, you know what, we'llbring things down.
Caveman, that's my style right.
Look at your cutlery draw.
You have a variety of differentknives, a variety of different
spoons, a variety of you knowdifferent forks.

(16:10):
I don't know how rich you areand from there you know that at
the beginning there was probablyone fork and then they
redeveloped and created thesenewer, different versions of it
that are more tailored to adifferent user case scenario.
You're not going to cut a steakwith a butter knife but at the
same time you're not going touse a steak knife and to spread
butter, unless you know yourdishwasher is not done.

Ed (16:31):
I don't know, robbie, you didn't know me.
When I was a student, I hadalways come to do this.

Robbie (16:34):
Not noodle with a fork gotcha or a pen, but on top of
it, when it comes to hard drives, early hard drive technology,
the difference betweenindividual drives are very, very
small and, as the rest of theinfrastructure, the client
devices that we're going into,could be geared certain ways.
It's the same with cars anddifferent brakes and you know

(16:55):
accelerations and tires andstuff like that.
So, for example, a traditionalPC class hard drive, you know it
, can read, it can write, canprobably hit, you know, 160 to
220 megs depending on capacity.
It's not designed to be on andconstantly being hit 24-7.
It's designed for a nice easyspin up, spin down.
So then we would look at, say,a NAS drive.

(17:16):
A NAS drive has much moreintermittent and irregular power
up, power down.
It has to go into hibernationmode a lot more readily.
Also, it's going to be inlarger groups of drives at the
same time.
So that means sometimesmultiple drives are being
written to in this more rugged,more high working environment
and drives are being written onbut in a very irregular fashion.

(17:36):
So you got like right, right,right, parity, right, right,
right.
So it's the behavior slightlydifferent there and a lot of the
time when you're being you'rereading from the drive, you're
reading one piece of data spreadacross all of those discs,
whereas if you use traditionalstandard class drives, you
wouldn't see the same level ofperformance level.
And then you go to some likesurveillance, where there's
going to be 95 to 98 percentright of footage and very rarely

(18:00):
you're going to access thefootage.
So that's another way in whichthe drive has to be geared
towards that user case.
And then we break down into thewhole pro enterprise thing and
a lot of that comes down to twowords durability, that the drive
is going to be an environmentwhere it's going to be battered
constantly, and then, when youget to the enterprise tier, you
can double down on the endurancefactor there.
But then on top of that therehas to be sustained performance.

(18:24):
So these are drives that aredesigned to reach heights in
performance but maintain thatheight without getting an
oversaturated cache, without youknow the the arm or the
actuator inside gettingoverworked.
And then you get into thesubject of dual arm hard drives
and NVMe hard drive, seagate.
What are you?
What is that?

Ed (18:46):
And the NAS drives.
When there's a lot of them nextto each other, they handle
things like vibration better andthat kind of thing.
Is it mainly kind of firmwarerelated differences, or is it
actually the build of thehardware inside the drive that
are different between thesedrives?
You know where?
Where is the difference?

(19:06):
Or is it a combination of thetwo?

Robbie (19:08):
It's a combination of the two.
I would argue that 10, 12 yearsago they were more physical
differences than firmwaredifferences.
But now we're seeing thefirmware play a bigger part,
particularly with more frequentfirmware updates as the drives
are used more.
And then, particularly now,we're seeing the introduction of
helium into hard drives.

(19:28):
So now we crossed the 12 TBline and we're now getting into
22 and 24 TB hard drives.
The only way that was achievedby multi, the different
methodologies was to take autilize helium in a sealed
casing and therefore theplatters that I wish I had by
fake little hard drive.
To show you my little four yearold welcome to school hard drive
.
It does my nut in, but theplatters got to be thinner.

(19:50):
So when the platters got to bethinner, then after that they
become helium, you know, reducesdrag factor in those casings.
But at the same time you've nowgot a much more crafted device,
which therefore then you'repushing even more into
surveillance, enterprise, nas,desktop use, as the capacity's

(20:11):
got higher and the way they werebuilt became more specialized,
and at that point the harddrives become even very, very
similar.
But then the firmware has to begeared to making the most out
of that.
You could.
You know, if you abandon firmwearing in general and you lined
up all the hard drives next toeach other, they're probably
very, very similar in inside,although some of the enterprise
ones will have a little bit moreonboard cache, and this is how

(20:34):
optin and thing with WD to e-cata few more terabytes, but
ultimately the firmware and morerecent hard drive releases is
where the drivers being givenbetter control of the crafted
elements inside, more tailoredto the end use and inside.

Ed (20:50):
Inside NAS and enterprise drives there's a feature TL ER
time limited error recovery.
Wondering if you could talk alittle bit about that and what
that is and what it does well, Ibelieve that's to do with the
amount of time it the drive isleft after.

Robbie (21:07):
Well, calling an error will be a bit strong.
But the recurring revisit ofthe drive and I think that's
another example of the drive,some drives having support of
that within the NAS structureand again you've caught me off
with that one I'm thinking aboutthe TL ER as the drive how
frequently a re attempt to readfrom the drive is made, and I
think more modern hard driveshave really upped that and got

(21:30):
the drive to effectively, evenas soon as it hits a wall, to
already be beginning the nextattempt by even a TL ER.
The last time I was reallyfocusing on that was around
about 2019 2020, and a lot ofthat was with WD and the whole
SMR business before.

Ed (21:46):
So again, perhaps things have moved forward how I
understand it works is you'requite right it's how many times
it will continually try andreread an error, a desktop hard
drive.
I believe it will keep ontrying and trying and trying.
But the problem is is when youhave that it can make the drive
drop out of a raid.
So the TL ER will stop ittrying so much so the drive

(22:10):
won't drop out of a out of araid system.

Robbie (22:14):
I mean as we move.
I might be wrong about becauseas we move more and more into
SSD populated server systemsbecause the cost of NVME SSDs
have come down, there was aspike, cheer, semiconductor
shortages and the like.
As we're seeing more and moreof SSD utilization being cashing
or in a hybrid tiered storagefashion, more and more users are

(22:35):
becoming less bothered aboutrecurrent error attempts because
SSDs by their nature alreadyhave that.
They've got that wholeblueprint ideology of access to
hold drive at once, whereas adrive has obviously got a more
linear approach to it so it canovercome or ignore errors or
work around or we'll get back tothat error in just a few
milliseconds.

(22:56):
Be right back.
But yeah, we're seeing, we'rekind of hearing about that less
and less at the moment as SSDoccupation increases and also in
drives, the very large harddrive.

Ed (23:08):
Some of them are SMR, as it called SMR hard drives.
I'm shingled magnetic recording.
What is that and should be?
Do you think we should avoidhard drives that have made that
way, or are they as good asother hard drives that don't use
that technology?

Robbie (23:25):
I mean.
I mean there's there's a fewdifferent outlooks.
Probably one of the morecurrent, one of the more common
outlooks is hate it, hate it,hate it, don't want them, hate
it.
And the real answer is probablya little bit more balanced than
that.
And again, we referenced the WDthing there and we'll come to
that in just a moment.
But Schengel magnetic recordingin a drive is, as the name, as a
roof would suggest, that thewriting action inside is ever so

(23:49):
slightly overlapped and itallows a drive to actually
achieve higher capacities at alower price point.
But it's built around the ideathat has to be a sufficient
amount of idle time for thedrive to recalibrate kind of the
written sectors and the indexinternally.
And if a drive isn't given thatoff at that down idle time to

(24:10):
recalibrate, then that's whenthey become problematic and
therefore unsuitable and a nass.
Now circling back to WD, wd gotin a little bit of trouble
because their WD hard drivescertain capacities I think it
was the two, three, four and 60B drives were SMR drives but not
advertised as such and therewere lots of people putting them

(24:30):
in things like a true, you know, a true NAS or free NAS back
then, and other service systemswhere if you were doing constant
activity with this drive andyou suffered a drive failure or
the drives reached at 100%capacity and then you wanted to
include a new drive and startexpanding your existing rate,
you still weren't given thesystem enough time for downtime

(24:51):
and it increased the potentialfor a rate for the drive rate
recovery or rate expansion tofail.
And that was why utilization ofSMR drives or drive managed and
host managed SMR and anotherthat in a bit it came incredibly
unpopular.
However, smr drives are stillvery common these days and, as

(25:13):
we generally find with a newcapacity tier I think 2060 B was
the latest one the earliestcapacity deers that get rolled
out are SMR and if you have ahost managed SMR drive, there
are ways and means to regulatethe drives a little bit more to
allow and actually profit by theuse of SMR drives in a larger

(25:34):
raid environment.
I personally still don'tmassively recommend it unless
you're running a particularlyhigh end system with
particularly proficient hostsmanaged SMR capabilities.
But yeah, smr drives, Ipersonally wouldn't recommend
them to most NAS users outsideof very specific environments.

Ed (25:56):
What do you think of the practice of people they're not
using NAS drives, they'relooking for something a bit
cheaper is shucking a drive froman external hard drive?
Are the hard drives insideexternal drives as good as if
you bought, say, the same drivethat wasn't inside a case?
I don't know how much it isnowadays.
I think the drives are prettymuch almost the same price, but

(26:18):
it used to be cheaper to buy thesame drive in an external case
and take the drive out.
I heard that a lot of timesthey were drives that failed
quality tests that they'd put inexternal hard drives.
I don't know what your thoughtsare on that, robbie, about
people using shut drives.

Robbie (26:37):
Well, I had heard that last point you made there and I
wrote an article, I made a videoabout this self plug, self plug
.
I did do a video on an articleon this and I'd heard that, but
because I couldn't find anywhereto back it up specifically
online, I didn't include that.
I have heard not a dissimilarpoint to that.
But mainly you're right aboutthe cost of shucking drives

(26:58):
these days is generallycomparable to standard buying
drives at internal bear and ofcourse, shucking a drive you're
generally throwing your warrantyaway.
You have a drive that has asoldered connector in some cases
or a bridging device that canoften be attached to the top of
the drive.
So, despite my efforts to tryand create a database of

(27:21):
shuckable drives, I've almostcompletely given it up, because
I created a massive database ofevidence online from like
everything from Reddit toindividual forums to buying
patents to Amazon that showedthe internals of some external
drives and the drive that wasinside.
But then I found out one changeof a serial number later and
it's a different drive, so itbecame incredibly inconsistent.

(27:42):
But on the other end of theseesaw one, if you go for a
device like an external, thathas more than one drive inside.
So it's maybe a RAID 5 devicethat's got four WD reds in drive
.
The saving potential there ishuge, Because those ones where
you generally find the pricesisn't just a case of 10 pounds

(28:02):
off each drive, but somehow thepricing on there because they
accept users are going to lose adrive in the RAID capacity
calculation.
I'm not sure, but it is hugesavings to be made in shucking
in that fashion.
Then you've got some drivesthat when the WD, for example, I
keep I don't know what.
Let's be fair.
Let's say Seagate.
This time Seagate will go aheadand go right.

(28:23):
We have produced 10,000 unitsof these 10 TB drives.
We're going to put three, butwe're going to put 3000 of these
in externals and the other 7000we're going to sell bear.
Now, these bear ones are moresubject to things like inflation
, profit margins and ultimatelytheir price point from after
manufacture is more floaty,totie all over the place.

(28:44):
But the ones that go intoexternal casing go into a
completely different pricingbracket and sale and demand.
And you will often find andagain we saw this enormously
during 2021 with thatcryptocurrency Chia.
I think it was 2021, it mighthave been 2022.
You must have seen loads ofpeople wanting to make an RAID,
chia service and stuff, and withthat people were buying up

(29:08):
external hard drives to shutthose drives, the reason being
because normal hard drives at 14, 16, 18, whatever the TB limit
was at that point were changinghands for silly money and the
e-retailers were aware of that.
But the external okay, the boxesthat were pre-populated DAZ
boxes predominantly, but the WD,mycloud and Seagate Backup Plus

(29:31):
and stuff those, the pricing onthose didn't increase, notably
for lots of different factors.
So they ended up getting, youknow, people would just buy them
.
Sell the empty case, becausethat's another way to recoup
your loss as well.
Selling the empty enclosure andultimately shucking eyes do
still think is very much a thing, but the brands are nailing it
down so much by adding thoselittle bridge, bore things on

(29:54):
the top of them, not onlynullifying your warranty but in
some cases not giving people anyindication of the drive that's
inside, and then white labellingthose drives as well, which is
another big factor, I think,where you will get a drive
that's the same model idea as abare drive, but it's been white
labelled, which again leads tothis question of hmm, is it a

(30:14):
quality control issue?
Did they specifically tailorthis drive for an external or
otherwise?

Ed (30:19):
As well.
Sometimes, when you shut them,didn't you have to kind of
bridge two pins on the?
Hmm?

Robbie (30:24):
I forgot what that's called.
There was a particularconnector on the top I can never
remember what it's called onthe top of a hard drive.
I don't think it's just power,but old servers would have an
actual interface going into thatone and I know a lot of the
time those plugs would either becompletely removed, and
therefore barring their use insome cases, or wouldn't run at
all unless they had that insideto start with.

(30:44):
We don't see as much of thatanymore, due to mass production
technique, I imagine, and alsobecause a lot of that was
restrict was available to IDEand ATA drives that had that
interface built into the pins,but you see it less now on SATA
drives.

Ed (31:00):
Talking about cheaper drives as well as like shucking drives
.
What do you think aboutrefurbished drives?
What's your opinion on thesedrives?
Is it normally the manufacturerthat's refurbishing these when
you see them in places likeAmazon, or is it like a reseller
refurbishing them?

Robbie (31:16):
Well, I have seen refurbished drives as an option
listed on quite a few officialwebsites.
I don't see it as much anymore,but I have seen it like
previously.
Certainly for the right setup,I would not be averse to
refurbished drives.
Indeed, again, I'll have todouble check this, but I did a
video recently self plug, selfplug, arrogant beyond the

(31:36):
extreme.
I did a video about is itbetter to have fewer larger hard
drives or lots of smallerdrives and then use, you know,
ray calculations.

Ed (31:47):
That was a question I wanted to ask you actually that, in
your opinion, what is better?
If I wanted, say, a 20 terabyteof usable space, would I be
better off to have two 20terabyte drives and just have
two, or would I be better off tohave maybe, I don't know three
10 terabyte drives?

Robbie (32:09):
The issue is there is no answer because the user case
scenarios.
There's pros and cons on eitherside that will be fit a
different user.
So if you look at, say, a fouror an eight bay server device,
nas, be it Turnkey or DIY that's, you're going to need a bigger
PSU, you're going to be you openthe door to maybe having a

(32:29):
better CPU to get more out of itand eat more a bit of 10 GB,
because more drives equalshigher performance but it also
leads to more points of failure.
You can save money by going formultiple drives, often over
going for larger drives, whichare often newer, and therefore
the newer drives.
You know we've seen examples inthe past that a lot of people
won't jump on the bandwagon witha new drive because the failure

(32:52):
rates are up there sometimesdepending on the drive.
So there are arguments foreither side.
I think it really does comedown to a user case scenario.
It's not always which one'sgoing to be the best money,
because in a lot of ourcalculations and I believe WDC8
and the like, frotashiba in whynot these brands?
They're aware of thecalculation, the whole price per

(33:14):
terabyte, and they've broughtit closer and closer and closer
together to a point where thesavings one way or the other are
actually quite minimal, butit's the system you're putting
them in.
So if you buy a 20 TB drivescore, but you need two, you
can't have one because you'vejust put all your eggs in one
basket.
So this drive that cost you 5,600 NICA, you've just spent a

(33:35):
grand to 1200 quid and you'venot even got the enclosure.
So the savings and also youdon't get the performance, as
mentioned earlier, whereas ifyou've got a bunch of 4 TB
drives for that same amount ofmoney and you put it in a RAID 5
, you've spread out your pointsof value.
You need a bigger enclosure andeverything, of course, but the
performance you will get whilereading off of 4, 6, 8 drives.

(33:57):
You can max out a 10 gig withthat.
That's fantastic stuff.
You're never going to max out a10 gig with two 20 TB drives.
If you're lucky you're going tohit 4, 450 megs on a good day.
So there's no real one answer.
There's just a huge separationof variables that can apply to
different end users.

Ed (34:15):
Yeah, I find personally when I'm using in my RAID server, in
the main RAID array, I preferto go for as few large drives as
possible because the data's notstriped across on the RAID
array.
I don't get any performancebenefit from having multiple
drives as well.
I haven't got a lot of slots insome of my servers so if I

(34:37):
wanted to actually increase thesize of the array later on down
the road like, as you know, themain array on RAID we can add
drives to it, increasing it inthe future.
But if our parity drive isn'tbig enough we can't add a larger
drive.
So I always like to try andkeep as few drives as possible.
But now ZFS is in RAID 6.12.

(34:57):
I'm in the process of buildinga server next week with some
SSDs in.
I chose to get some 2TB SSDsand I had the choice between
buying 6 2TB SSDs or 3 4TB SSDsand it was cheaper to buy the 6
than the 4.

(35:18):
And I ended up.
I think I ended up having aslightly more capacity and, like
you say, the speed increase,you know with the striped data
over the SSDs made me choose thesmaller drives.

Robbie (35:30):
And actually in the last point, to add to that, I should
have mentioned earlier theincremental cost of failure,
which I believe is a StevenSeagal film, the.
When, when you're running asystem that's got, say, 10 2TB
drives and one of those fails.
Now you should have a hot spareon a shelf of course everyone at
home say it together but at thesame time you can nip out and

(35:51):
buy 2TB drive.
Now if you were running 10 TBdrives and one fails, you're
down to 300 NICA, just like that, and you've got to go out and
get that and there's got to beavailability for that.
And you, despite modern harddrive production getting better
and better and better and thisis a job I have no statistical
research to back this up otherthan every single Black Friday,

(36:12):
every single Cyber Monday andworking in e retail for a long
time there's always more of thesmaller capacities and the big
capacities, and I don't knowwhether it's a resourcing, that
is, brands knowing that they aregoing to sell more of certain
hard drives, or that larger harddrives are purchased on mass
and they're bought in cartons of20 at the data center level or
unified storage level, but allof the smaller drives, none of

(36:35):
the big guys are going for those, so you can go to your local e
shop and then get a bunch of one, two, four, six, eight TB
drives easily.
So that's I think that I'mgoing to update my article after
this.
Definitely a reason that wholekind of incremental cost of
failure coming this summer.

Ed (36:52):
Yeah, that's a really good point.
You know, if a drive fails,it's much nicer.
It costs you a hundred dollarsor a hundred pounds to replace
them, three or four hundred.
Yeah, it's definitely.
I'm going to ask you before wemove on to speaking about SSDs,
because I want to speak to youabout SSDs, about Western

(37:13):
Digital Device Analytics or WDDA.
Now I know.

Robbie (37:19):
Sorry, the sleeves are going up.
Now, the sleeves are going up.

Ed (37:22):
So can you, for people who know nothing about it whatsoever
, explain what it is and why itis so controversial at the
moment?

Robbie (37:32):
Well, you know credit where it's due.
A chat Will over at Space Rex,again another NAS platform over
in the US.
There he was pretty much theone that brought this to
everyone's attention.
It was only after zooming withhim and doing a few emails back
and forth that we startedtalking about looking at this
from a lot of different NASbrands, because he really
doesn't cover too many beyondtrue NAS and and synology and a

(37:56):
few smaller like pocket bitslike that.
But the long story short is andI wish I had a WD drive here to
show you I believe it was2019-20.
Don't quote me on that, youknow.
You know alleged they rolled outtheir alternative to Seagate's
Iron Wolf Health Management,which is a kind of an on.
It's either on the drive or itis an on system utility that is

(38:18):
accessing the drive and taking alot of the values not
dissimilar to that of smart andin some cases, reinterpreting
the smart data in new, more userfriendly metrics to alert the
end user to the health of yourdrive.
So when you do smart tests onany drive, they are tremendously
useful.
Smart tests have been aroundfor donkey's years.
They will continue to be aroundfor a long time to tell you

(38:39):
about you know bad sector spinup, unload cycles, power on how
long, etc.
But where WDDI felt foul a lotof public opinion.
One particular value the poweron hours.
Because users started croppingup saying their WD drive
reported power on hoursequivalent to three years on the

(39:02):
mark.
To say your drive has now beenon for three years, you might
want to consider replacing it.
Now.
This power on hours warning,notwithstanding that it was
included apparently on a lot ofpro series drives which have got
a five year manufacturerswarranty, which is a lib.
But on top of that it was howit was being handled by the

(39:23):
connected host systems that itwas in.
So earliest reports wereSynology NASs that were running
DSM 7.0.
It has since been removed fromlater revisions of DSM.
I should add, had it turned offby default.
But if a user turned on WDDI,western Digital Device Analytics
from within the storage manager, if the drive said your drive

(39:44):
has been on for a certain numberof hours, now you might want to
consider replacement, theSynology would then push a
notification to the end user andput the storage into a warning
which, for anyone who owns aserver with mission critical
data, that is serious browntrousers time.
That is terrifying because allof a sudden, your drives, which

(40:06):
you've purchased at the sametime, remember, which you've put
into the system at the sametime, but therefore power on
hours is gonna be not dissimilar.
Often, you know, one or twohours between is gonna result in
your drives, one by one, samewarning and you are going to
freak out.
And a lot of users went out andbought drives.
Apparently what Will was sayingwith one of his clients as well

(40:28):
, and this was the wholebusiness.
So that WDDI tool and healthmonitoring tool, seemingly I
don't know how much of it isunique data.
A lot of it seemed to bereinterpreted smart data from
what we could see, but presentedin optimal user-friendly
fashion.
That power on hours thing hasrubbed a lot of people up the
wrong way and now, althoughSynology again removed it as of,

(40:50):
I believe, midway through 2021,and it's in no later revisions
of Synology DSM platform, qnaphave just rolled out QTS 5.1,
which has WDDA support includedand you can't turn it off and
it's on by default, but we'restill.
We have no way of modding andforcing a drive to give that

(41:12):
three year warning withoutwaiting three years to see how
the QNAP interprets that.
Qnap have spoken to me and saidthat they, their system
interprets this as information,not a warning, and it won't put
the storage manager into awarning state.
It won't also shout at the enduser by bloody drives.
It won't do any of that.
But we these companies?

Ed (41:33):
do they partner with WD?
Is there any kind of financialincentive?
Like you say, after three yearsit's gonna sell quite a lot of
hard drives if a companyintegrates the WDDA into their
system and saying your harddrives need replacing, that's
quite a lot of sales after threeyears.

Robbie (41:53):
I mean, I'm obviously without knowing for certain, I'm
reluctant to say whether thatis the case, but my instinct
says, with regards tocollaboration at least, I don't
think the NAS brands would havecollaborated with WD to that end
, because Synology ditched itvery early doors and QNAP have

(42:14):
only just added it to QTS fromsomething that was rolled out in
2019 to 2020.
Now QNAP integrated it, fromwhat I could see, as part of
their predicted failuremigration service.
So QNAP have got this thing.
I'll make it very, very briefwhere, when you have a hot spare
drive at the moment, if one ofyour drives falls over and
knackered, whatever you wannacall it, your hot spare jumps in

(42:36):
and your raid starts rebuilding.

Ed (42:37):
If it was ZFS, it would be a lot quicker, but so basically,
it's making you use your hotspare when you don't need to.

Robbie (42:46):
Yeah, but what this might do when, in the case of
predicted failure with the QNAP,is the drive is being monitored
with as many health monitoringtools as possible.
So you got Ion Wolf, wdda theirown drive analyzer, smart, and
when they can see indicationsthat the drive could potentially
fall over soon, then itintegrates this drive and
therefore it clones the driverather than the raid degradation

(43:08):
in the raid rebuild, which canyou know 8, 12, 24 hours, it's
cloning this drive that'sshowing signs of dodginess with
a new drive and then it justpowers down the system and swaps
the new drive in and you've gotno downtime.
The problem lies with if WDDAis influencing that, then, as
you say, the drive failureprediction might replace that

(43:28):
drive on the basis of the drivebeing on for three years, even
though in every other regardsmart tests and just in general,
it's a healthy drive.
I mean, I was thinking aboutyou, but I'm willing to bet
somewhere around you is a harddrive that's been knocking
around for longer than threeyears in operation.
Oh for sure, I've got one nineyears that I walk past every day
.
I want to kiss it if it wasn'telectric.

(43:49):
But yeah, so WDDA, I don'tthink it's a financial incentive
by the NAS brand to integrateit.
I don't think money has changedhands any more than them
integrating Seagate's own wolfhealth management with the date
recovery.
I can't speak for WD, because WDreinterpreting these metrics in
WDDA to make them all userfriendly sounds quite nice.

(44:11):
But if they'd repackaged thishours on warning to your drive
with three years warranty hasbeen on for two years and 11
months it's about to go outsideof warranty.
We just wanted to let you know.
That sounds great to me,because there are people out
there that add 10 years warrantyreplacement spares on their

(44:33):
washing machine.
They get the extended warrantyon everything and when the
warranty is almost up they flogthe thing and use the money
towards something new.
It's quite common.
So if the warning had beeninterpreted like that, I'd like
that.
Actually I really like that.
I wish drive manufacturers didthat.
They won't, because then theyopen themselves up to pointless

(44:53):
RMAs, but I think that wouldhave been a better way for that
message to be delivered.

Ed (44:58):
Do you think, Robbie, that say you had a NAS drive, you had
four drives in it.
For instance, you didn't have ahot spare.
You ran it for six months andyou thought I really should get
a hot spare.
So you're going by another WD,so you've got four of the drives
that are six months older thanyour hot spare.
It comes to the three yearpoint and come up with a warning

(45:23):
and it swaps it for your hotspare.
Your hot spare is already gonnabe two and a half years old, or
even if you've had your hotspare safe for even longer, for
three years as well.
It could literally do a hotspare and go oh actually you
bought this one only two daysafter the other one, so the hot
spare will still be, even thoughit's not really been used on
this power on time.

(45:44):
So it will still be reportingthe same information, I assume.

Robbie (45:47):
I will smash stuff, cry, go to the pub.
I would be absolutely livid.
And the thing is, people onlinehave made much more I would
take eloquent descriptions thanmyself.
As for case scenarios and themost compelling for me which
sort of impacts me, but notreally, but it is, for me, the

(46:08):
most astute argument Imagine youare either a system admin or
you're an installer, a serviceprovider, and you two, three
jobs, maybe five jobs a weekwhere you set up people's
storage for them.
So that's maybe three to fivejobs a week.
You've been using this veryspecific setup that you know is
fantastic for that businessmodel.
They're content creators,they're post-production, they're

(46:28):
an editing suite.
So you're creating this setupfor them and you're doing it.
And then all of a sudden youwake up one morning and because
remember, it's power on hours aswell, so people do power it
down, so it won't be the dominoeffect per se.
But then all of a sudden you'vegot freaked out businesses that
you're on a retainer for or thatyou've signed off on a warranty
period that you take care of,warranties for them, and it's

(46:50):
all going bananas becausethey've all suddenly woke up to
amber lights everywhere and hotspares suddenly getting swapped
out unnecessarily in specificconfigurations and you have
suddenly got a workload thatyou're not really gonna be paid
for.
And all of this down to awarning that has no real bearing

(47:10):
on anything it has.
It's the equivalent of sayingthe metal has become slightly
less ionized.
It might as well be that thedepths of this warning so yeah,
that's the big brouhaha aboutthis, and it's again.
Spacerx, slightly more so thanme, is really pushing for WD to

(47:33):
remove this on their drives,because it's one thing for all
of the NAS and indeed and I'mnot sure about its position in
things like TrueNAS or Unray Iimagine it isn't integrated into
that, but then I just simplydon't know that he would like it
if it was removed utterly fromWD drives, whereas from my point
of view, I just wanna make sure, as many NAS vendors, do not

(47:56):
integrate WDDA or at leastdisable this one warning,
because the rest of the warningsor the information being
presented they reinterpretedsmart data, to my mind, and
therefore they actually have anelement of use, because telling
a slightly more novice end userthat your unload load cycles are

(48:17):
too high, they have no ideawhat that means, whereas if you
say to them this drive has beenspun up, spanned down, powered
on.
Powered on just a little bittoo much.
Frankly, they're gonnaunderstand that, because it's
caveman to me.

Ed (48:31):
Yeah, it's very naughty as well, I think, for them to have
the warning when it's threeyears.
If they have a five yearwarranty on a hard drive, it
should only be five years.
You know, that's really badbecause they're saying, like you
said earlier, that their fiveyear warranty isn't really kind
of worth the paper or the pixelsit's written on.

Robbie (48:56):
I should have mentioned it earlier.
One of the big barriers againthat we've encountered with this
is the information is so scantonline.
Wd has its own page that's beenthere for three to four years,
I believe, and it's incrediblynon-descriptive.
Then the NAS brands themselveshave put very, very little
resources out and there is noindex online what WD drives have

(49:18):
this on board.
There's just nothing onlinedetailing it, so we don't know.
All we've got is conjecture,forum reports and the drives
that we've managed to test.
So a lot of the informationwe've gathered unfortunately has
to come under the heading ofhearsay a little bit, because
we're going for some randomReddit post in some cases.
But there's enough evidence tokind of underline this point, I

(49:41):
think, and it's enough, I think,for a lot of users much like
the SMR business we talked aboutearlier with WD that have made
them go.
Ooh, that's not nice.
We don't like your businesspractice Like we could.
Seagate for a long time, let'sbe frank, were kind of the
dumpster fire for a lot of usersin terms of hard drives.
It was like oh, the Seagatedrives are cheaper.

(50:02):
Oh well, I'll deal with it.
And WD had this spotlessreputation about a decade ago in
terms of hard drives.
I don't know what's going on inthe last decade.

Ed (50:12):
Yes, it sounds a bit of a shame they do that, but I think
your idea of just vendors havingit disabled for that particular
reading and keep all the otherones on there or even just have
it off by default people canswitch it on if they want to and
it explains that it doesn'tmean you need to actually change

(50:34):
the drive.
It's just a warning that it'sbeen in use for a long time.
I was thinking like do you knowat all why?
Talking about smart data, whyhas that never been standardized
?
Like every vendor, their smartdata is slightly different and
is interpreted slightlydifferently.
Any idea why?
I guess it's just never been astandard for it.

Robbie (50:55):
I would have to guess I mean the bits of it that aren't
universal, I think wouldprobably come down to their own
secret recipe for their harddrives they wouldn't want to
give away because even thoughthere's a million squillion hard
drives released every year fromthese brands, a lot of the
production and even the firm wasprobably locked up pretty tight

(51:15):
and I think it would be toughfor them to give away a lot of
their secret source that goesinto it.
I mean we look at the newergeneration of hard drives that
are rocking out with WD, rockingout Optinand to free up some of
the space on the platters toget bigger hard drives.
Then you've got EAMR, hamr andwe're talking about like
companies in the tens ofbillions of dollars they're

(51:37):
always trying to cut ahead.
I think.
Is it the end of this year theywant to have 30 TB drives live
and by the end of the decade 50and 100 TB drives.
They get there by beingsecretive as all hell.
And I think standardization ofsmart data they would probably
fight against it a little bitfor that reason.
There will be some values againpower on load, on load cycles,

(52:00):
blocks that are going to bestandardized for the most part,
and I think when the smart testis performed can differ
depending on systems, whetherit's a physical action, a
scheduled action or every timethe drive is booted.
I think there is disparitythere as well.

Ed (52:17):
I can know the smart data.
Sometimes the lower number'sbetter and sometimes the higher
number and they have this kindof tolerance.
The tolerances can be kind ofdifferent between different
vendors.
The low number with one can bebetter and a high number with
another.
Like it starts at 100, and itgoes below 40 as your threshold.
That's really bad.
But then you can see thethreshold is 40 and you see that

(52:39):
your hard drive is like at 100and you think, oh, does that
mean that's really bad becauseit's over 40, but the bad thing
would be if it's under 40 andthat kind of thing.
I've always found that smartinformation very difficult to
understand.

Robbie (52:54):
But again, that's probably, and again, this is not
me defending WD at all, butWDDI when it was first
envisioned.
And again we've got to adddisclaimer than allegedly all
the information we found onlinewith WDDI was that it was
initially rolled out for WDpurple drives for surveillance.
Now in that arena, a lot ofthese surveillance devices, they

(53:16):
have a surveillance feed, abunch of camera feeds, your pan
tilt zoom and all that luck.
But the actual management ofstorage bordered on caveman,
bordered on terrible, because itjust went where's your storage?
It's over there, okay, and thatwas it.
You couldn't do anything.
So what WDDI in that context wasfor was to allow the drives own

(53:38):
health monitoring andmanagement to be translatable to
your HIC visions, your anchors,your Rio links, your DVR
vendors, and then those systemscould then have that interpreted
data and then splashed it intotheir own GUI.
Which is why when WDDI firstrolled out this is getting
really boring, I'm sorry, butWDDI, when it was first rolled
out, was available as an APK forall of these vendors to

(54:01):
integrate into their own GUI.
So in that context, when yoursystem doesn't have a robust
hard drive or storage media,health monitoring tool and
sophisticated warning system,wdda makes a lot of sense.
But the minute you roll it intoa NAS system of any version of
NAS again we can go Turkey, opensource or anywhere in the
middle.
Those systems have robuststorage management.

(54:24):
They have the skill set and theoptions of the customization
and the control to handle thatand WDDA is not necessary in
that arena.

Ed (54:35):
I think the WDDA, you know it's quite a nice idea in a way,
because being able to interpretsmart data in a human readable
form is quite a nice idea.
But hopefully they may justkind of change that bit and
improve the WDDA going forwardand it was just a bit of a
mistake.
Anyway, kind of moving on, I'dlove to talk to you about SSDs,

(54:57):
if I may.

Robbie (54:59):
Far.
Away.

Ed (55:00):
So, in the context of kind of NAS systems, what do you see
the pros and cons of using SSDsare, please, and also really
what specific use cases wouldyou think for a NAS to have SSDs
as opposed to not?

Robbie (55:15):
Well, up until about two months ago, my perspective was
pretty solid on SSD utilizationin NAS, again, where I've done
that caveat.
You'll see why in a bit.
So SSDs in NAS, you have tohave a NAS system that has a
sufficient CPU that can affordthe number of PCI lanes to make

(55:37):
the most of it.
Now, if you go to SATA SSDs,which again are becoming
increasingly affordable, in thatscenario you don't have to
worry too much about the CPU,because the amount of SATA lanes
that are all paired up for allthe individual bays that you're
laying out, it's the same forhard drives and SSDs.
But SATA SSDs, again, they willadvertise five to 550 megs, but

(55:57):
realistically you're looking atsomewhere about 450 peak,
except if you're using veryspecialized and targeted setups,
whereas hard drives are nowavailable that go up to 285
megabytes per second and we'restarting to see reveals of hard
drives getting to 300 megs.
And then you've got dualactuator drives, a hard drive

(56:17):
with two arms that can hit 440,450 megs, and these are hard
drives that can have 16, 18, 20terabytes.
So the need for SATA SSDs hasrescinded, whereas NVME SSDs in
the NAS is starting to boom.
So NVME SSDs in NAS, becausethe NAS has the slots there or
we're getting better CPUs andthe price of NVMEs coming down

(56:40):
has all resulted in this voguetowards NASs either being fully
or partially populated withNVMEs and up until about two
months ago I would have said itwas largely a waste of money for
anyone underneath enterprise.
Because if you don't have a CPUthat can spread out its
available lines, if you're goingfor like a cellar on that's got
eight lines to play with, eachone of those NVME SSDs that can

(57:03):
hit 3, 4,000 megs can't exceed1,000 megs each.
So you're capping immediatelyand one of the amazing benefits
of M2 NVMEs gets lost intranslation because each one can
only give you 1,000 megs andthe CPU has to go hell for lever
needs at least eight gig ofmemory minimum to really make
the most of it.
So you end up kind of losingout on some of the benefits of

(57:26):
the SSDs.
However, two months ago AsusStore NAS brand nowhere near the
clout of your big names outthere they released something
called the Flash Store and itbroke a lot of the
preconceptions.
It's a six-by NVME Intel cellaron powered system and it's $499

(57:47):
.
It's got four gig of memory inthere.
It's as quiet as a mouse that'sasleep and even though each of
the lanes are limited to threetimes one, for $499 for a six-by
M2 NVME system cellar on custombuilt, tiny, well ventilated.
I struggled to try to build that.

(58:07):
I tried to spec up.
And they include software withit BTRFS, full UI, mobile apps,
desktop, the works.
There's even remote accessavailable as well with their own
domain.
So I was trying to figure outhow could you even build
something like that at thisprice point?
And then they released a 12NVME version which has got 10

(58:30):
gig on board as well.
So that's about $850 for 12NVMEs.
Now, again, they're allrestricted.
But all of my reservations andissues with restricted bandwidth
or limited bandwidthavailability because, remember,
that is a GIN, sorry, that's aneight-lane CPU with 12 NVMEs

(58:50):
inside and they're using alittle bridge.
They're using a PCI bridgeinternally to make it work.
But at that price point itcosts less than most eight-by
hard drive turnkey solutions inthe market and at $499 for the
six-by one.
I'm not a salesman but I justcouldn't figure out how they
could piece it togetherfinancially.

(59:11):
And now, from what I understand,all of the NAS brands are
working on M2 NVME solutions.
So in the last few months I'veseen this change towards
people's attitudes to M2 NVMEsin NAS, because the decreasing
cost of M2 NVMEs and thedecreasing cost of the NAS, the
flash systems that use it, haveremoved that big bottleneck

(59:32):
factor from people's budget andtheir minds.
Because they've realized well,I'm not spending two, three,
four, five grand on a compactflash server or 10 to 20 grand
on a rack mount U.2 server.
I'm spending 500 NICA on thisserver.
That's M2 NVME all the waythrough, which opens the door to
multi-virtual machineutilization, container

(59:54):
utilization, plex Media Server.
I did 8K Plex Media Servertranscoding on a 499-pound NAS,
which is insane for 8K playback,real 8K HVC.
And then when you've got usersthat are running again high
volume, high frequency databaseswhere the utility and benefits

(01:00:16):
of M2 NVME, your multi-useraccess, really come into
fruition, this tiny littlecompact server that's about this
big is just insane and we'reseeing more of these solutions
throughout this year.

Ed (01:00:29):
Even though the bandwidth isn't there to run all the NVMEs
at full speed.
In reality, you would neverhave all of the drives running
at full speed anyway, so for alot of people you wouldn't
really notice the lack ofbandwidth running all of the
NVMEs and, like you say, it isan amazing price point for 499
pounds.
Was it 499 pounds or dollars?

Robbie (01:00:50):
It's 499 dollars, so dollars as well, but again
that's 6M2 NVMEs that arerestricted to three times one,
but they're in a raid, so thereis performance benefits
internally to them all beingafforded together.

Ed (01:01:03):
I think a lot of times companies like you say you don't
know how ASUS could do it socheap.
I think sometimes businessmodels for companies, when
they're not established in themarket, is they come in with a
product and they make a loss onit, and then all the other
brands are like how can you dothat for this amount?

(01:01:23):
And it gives them a reputationand they're happy to lose the
money in order to angle theminto the market.
They're probably making a losson each one of those in order to
basically give them areputation that they don't have
in the NAS market maybe.

Robbie (01:01:42):
I think at the moment we're seeing funny you should
say that this again I didn'tknow if we were talking about
this, I would love talking aboutit anyway this growing middle
ground between turnkey NASsolutions QNAP, synology,
terramaster, asus, to all ofthat luck and at the other end
you've got your pure DIY.
So build it from the ground up.

(01:02:02):
You use Unraid, you use TrueNAS, you use Open Media Vault a
myriad of different open sourceoptions out there and we're
starting to see more and morethis middle ground of companies
that are just producing pre-madeNAS servers with no OS.
They might lump on Unraid in atrial version or True NAS, but
they're releasing these hardwaresolutions in the middle and a

(01:02:24):
lot of the time and a case inpoint, the store AXA.
They're being designed arguablyas loss leaders themselves so
this company can build areputation and then maybe it
released our own OS down theline.
So this middle ground that'sexpanding by people.
Again.
Topten is another one withtheir little mini servers we're
seeing rolling out.
We did a review of one of thosethat had a Penteon processor

(01:02:49):
inside at N600, and that systemis so remarkably cheap.
It's like 250, 300 Nica forthis quad core Penteon system
with M2 slots ready to go and onthe one hand, you have the open
source crew going.
Wait, why am I even I thinkeven Linus Tech Tips made a
video on this very subject whenhe was talking about that AXA

(01:03:09):
store device going?
Why am I wasting time buildingwhen it's costing this little to
have these things now?
And then you have to turn.
Key solution provided.

Ed (01:03:17):
So on the AXA's one, could you run anything on that?

Robbie (01:03:20):
Absolutely.

Ed (01:03:20):
I mean again so I could buy one and run Unraid on one of
those AXA's boxes.

Robbie (01:03:25):
And can I recommend, if you do that, to Google it first,
because you'll find my video.
He said arrogantly, but yeah,100% they do not endorse it.
They don't like it.
But I've installed Proxmox,truenas and Unraid on AXA store
systems and QNAP as well A lotof these NAS brands.
They've a run from an internalDOM little five gig flash module

(01:03:47):
inside but stick a keyboard,stick a KVM keyboard video mouse
.
Go into BIOS, change the orderUSB sticking your USB done and
if you want to turn it back tothe original OS BIOS, reset it.
Remove your USB key and thenthat's it.
And in most cases, unless youdo something particularly
horrific I don't think it youknow they would have to prove

(01:04:10):
you've invalidated your warrantythere, which most NAS brands
probably wouldn't do.
So that's one of the reasons.
I've seen a lot of users whojust they're money rich, time
poor, which I know is incrediblytrite description but rather
than build their own NAS and geta MOBO from scan, go to eBuyer
and get the hard drives and theSATA cable and go to about three

(01:04:32):
different vendors, all of whichhave their own RMAs to build it
, sit on the floor and have thathorrible moment to press the
power button with the fingerscrossed and praying to the gods
of IT.
All of that time is.
You can now just buy a custommade solution from a lot of
these, either middle groundproviders or the turnkey NAS
providers, and they just lump onyour own OS.
It's really interesting the waythings have developed and, with

(01:04:55):
a lot of open source softwareand again Unrayed, probably the
most user friendly of all of thesoftware only NAS software is
on the market you can justsimply download, try for free.
All of those have becomeincreasingly more user friendly
and easy to install on thishardware and I just don't think

(01:05:15):
the NAS brands like it that much, but it's that's what people
are doing.

Ed (01:05:18):
And I think you know, for Unrayed as well, the fact if you
buy one of these turnkeysystems and put another OS on
the fact that Unrayed runs off aflash drive and that flash
drive runs into RAM, you're notwasting one of the disks inside
your NAS with the actual OS forthe NAS, so you're able to use
your hardware a little bit more.
I saw your video on puttingUnrayed on QNAP and I tell you I

(01:05:44):
really liked the part of yourvideo and I think everyone
should go and watch this partand I'm going to put it as a
link in the show notes where youdescribe how to format a.
USB stick, that's larger than 32gigs in Fat32.
Because Unrayed needs Fat32 toboot and you know it's really

(01:06:05):
great to see your way ofactually being able to format.
Did I use disk part and thenRufus?
It was a while ago, I can'tremember what I did In my head.

Robbie (01:06:15):
it was using disk part to format and then I used Rufus
to create the ISO.
You use Rufus just to formatthe format in.

Ed (01:06:23):
Fat32 because Windows will go.
No, that's more than 32 gigs.
You're not allowed to do that.
I'm sorry.

Robbie (01:06:31):
I like it somewhere like Windows.

Ed (01:06:33):
They're like what year?

Robbie (01:06:35):
is it?
It's 2023.
No, we're not formatting Fat32.
Are you serious?
We have XFAT now, you know.
Tell them to use that.

Ed (01:06:45):
Make it a little bit more serious.

Robbie (01:06:48):
Tell them to use that Make it double XFAT Unbelievable
.

Ed (01:06:55):
And with QNAP can most of the ones installed on Raid-On.
The ARM-based ones can't.

Robbie (01:07:05):
Well, the ones that have got an integrated graphics CPU
and therefore have HDMI out.

Ed (01:07:11):
And why do they need that, Robbie?

Robbie (01:07:14):
Because you can't UI into it.
You need the means to make theQNAP allow you to get into BIOS,
because otherwise you need avisual output to get into BIOS,
to switch away from the DOM,because otherwise by default it
will always load from there.

Ed (01:07:28):
Anyway, I'm going to switch gears a little bit and I would
like to ask you about ransomware.

Robbie (01:07:35):
Keep it light right.

Ed (01:07:39):
Can you explain what ransomware is and how it differs
from other kind of cyberthreats like just the regular
virus, and what are the motivesbehind ransomware attacks?

Robbie (01:07:52):
Well, I think a lot of people that are watching this
podcast, a lot of us are, youknow, the old school.
So when we think of viruses andwe think of a PC covered in
pop-up windows and you know thenightmare scenario, what makes
ransomware different isransomware tries its damnedest

(01:08:13):
to not harm the system it's in.
It's one of the few invasivethreat actors that doesn't try
to destroy the system.
All it wants to do is getinside, zip up everything into a
completely inaccessible format,leave a little ransom note,
generally on a digital rolled upscroll, on a knife stabbed in

(01:08:35):
the wall, before it leaves.
That effectively says if youwant to access your data, you've
got to pay us via this specificmeans.
More often than not these daysit is via Bitcoin and a Bitcoin
wallet, because it is incrediblydifficult and impossible to
trace for them to be found outvia those means but then we're
still via.

(01:08:56):
This methodology is unlikenormal malware and major viruses
which target big, big files.
Generally.
They'll kind of see someonebeen able to get in and then
attack the larger file structure.
Ransomware starts at the bottommore often than not and it goes
to the smallest files, thereason being it can do more

(01:09:16):
files and it heightens thepercentage chance of hitting
that payload, that wonderful payfile that someone goes.
They're the fighters of mychild from when they were zero
years old.
They're the ones I'm preparedto pay for because they're
irreplaceable.

Ed (01:09:31):
If I could just just button very, very slightly.
So just just for people.
I'm sure everyone listening tothe podcast or watching it knows
, but what ransomware does isactually encrypt the files and
hold them to ransom.

Robbie (01:09:44):
Sorry, yes.

Ed (01:09:45):
Yeah, Okay, sorry, no, no, no, no no you're right.

Robbie (01:09:49):
But when it comes to when people's perception of
ransomware, whenever we look atbig corporations, they get hit
by ransomware and most people go, yeah, well, they're a big
corporation, they should havepaid for a lot of system admin
and security in about 8,000 VPNs.
And when smaller groups areimpacted by ransomware, there is
this vibe of well, the brandshouldn't have let me down, you

(01:10:10):
know.
But unfortunately the truth ofransomware is that I'm not going
to say the majority, because Idon't have the statistic to hand
, but a significant body ofpeople that are impacted by
ransomware.
It's often because they punchedholes in their firewall or
ignored a lot of the warningsthat are presented to them by
the system when they're settingit up and then later on.

(01:10:32):
Then you come into this road ofhaving regular updates and
there is this quasi relationshipwith ransomwares.
Who is truly culpable?
Obviously, the threat actorsperforming it are their dog dirt
, but it is where the end clientdevice can only do so much.
When we talk about open source,for example I know I'm going on
a slight tangent, but I willreally in I promise when we're

(01:10:55):
looking at, say, open source alot of the open source platforms
they like, true now is afantastic example If you try to
set it up in an insecure fashion.
True, now is just goes.
No, if you really really mustdo this very silly thing in this
very silly way, then click thisbox.
That goes.
I take all the risks.

(01:11:16):
Not a lot of NAS brands do thatand it really winds me up.
They don't do that.
The reason that a lot of theturnkey now solutions have been
hit recently by ransomware andagain we'll talk about deadbolt
I assume Several brands wereimpacted by that was because, on
the one hand, users werehandling their device unsafely,
but a lot of these devices wereadvertised as simple, easy to

(01:11:39):
use, chewable crayon, userfriendly.
You know knobbers like me onYouTube saying as such.
And then the NAS device comesout and it allows these users to
set it up in this insecurefashion.
So yeah, ransomware attacks aregetting more frequent and it's
worse than a virus, because atleast with a virus you don't
know in charges the virus that'sright actor.

(01:12:01):
With a virus doesn't leave youwith an invoice, if you know
what I mean.

Ed (01:12:06):
And what would you say is a good set of things for someone
to do to help prevent themhaving a ransomware attack in
the first place?
Are there certain protocols?
You know best practices?

Robbie (01:12:18):
Well, definitely, these days, port forwarding and just
punching holes in your, in yourrouter and your firewall and
stuff is so much less necessarythan it's ever been, and I think
a lot of users are slightlyunaware of the fact that when
they need remote access, thatwhen they do remote access their
server, they either use the onethat's provided by the NAS

(01:12:41):
brand and the third party brandas well, and, I think, unraided
as well.
I forgot what the unraided oneis called that allows remote
access.
You sign up to something Ican't remember what it was, I'm
sure you're going to correct mein a moment but things like
Towscale now exist and lots ofother VPN type splinter that
actually give a decent amount ofperformance, and installing

(01:13:02):
them allow you to completelynullify unencrypted transmission
between you and your deviceremotely, and a lot of them can
be tailored towards third partyapps.
Now on the subject to thirdparty apps, most third party
apps now everything you need iskind of there, and a lot of
people are installing thirdparty apps on their NAS systems.

(01:13:24):
A lot of the time grant theseapps way too much access, so
when you're installing the appfor the first time, some NAS
brand will say what directoriesdo you want access to?
What privileges?
Do you want this to have?
Much like your mobile phone,when you open an app for the
first time and then it goes holdup.
Do you want to give access tothe camera or hold up?
Do you want to give access toyour contacts?

(01:13:44):
Not, a lot of apps on NASplatforms I include all the NAS
platforms actually open sourceor turnkey provide that, but it
is there so you can installpretty much anything, first or
third party now go straight intothe control settings and remove
all of the access, and the samething goes with some NAS
devices.
Now have an isolation mode.
You click it and any thirdparty app, anything that uses

(01:14:07):
remote access, php type stuffany of that can be disabled in a
heartbeat and after that,really everything else is just
small increments of protection.
So, disabling your adminaccount, randomized ports these
are things that now we'restarting to see NAS brands
actually say during setup, wewon't allow you to use the admin

(01:14:31):
account, you have to disable it.
Or when you set up the devicefor the first time, it won't
allow you to run certain apps orservices without tick boxing
all of those protections.

Ed (01:14:44):
Basically, it's keeping the server secure from, obviously,
the internet.
So that would also, I guess,include things like your router
and firewall, keeping thefirmware on that up to date as
well.
So it's not just, it's not justyour, your NAS or your server
you need to keep secure.
You need to keep the gateway ofthe internet coming into your

(01:15:06):
house secure.
You don't want to be using oldfirmware on a 10 year old router
.

Robbie (01:15:13):
That was another big area.
That deadbolt impacted peoplebecause a lot of the I think it
was three, three and a halfthousand users I'm sure someone
will correct me that were hit bydeadbolt and a lot of those
devices were running super oldfirmware because people took the
device home, set up in thecorner and never updated the
firmware.
But I mean, I don't know about,I don't know how much console

(01:15:33):
gaming you do, but if you own aPlayStation four or five, if you
don't update the firmware for acouple of revisions, you can't
access the internet.
It just won't let you, youwon't allow you to use online
services.
Now there are users that wouldhate to have that on a NAS
server, but I personally thinkthat would be great.
If you're not updating tolatest firmware, you can't
access remote level services.
Same goes for two stepauthentication.

(01:15:55):
You should be set up that.
No, you can't do this muchunless you enable two step.
But these are very strongpositions to take.

Ed (01:16:04):
Yeah, I can.
I can totally see your pointthat you're only as secure as
your last, your most recentpatch or firmware update One
step.
But I I'm the opinion thatpeople should be able to kind of
choose.
I look at it as like, say, Iwant to go and buy a Ferrari and
Ferrari put a limiter on so Ican't go above 70 miles an hour.

(01:16:24):
I shouldn't go above 70 milesan hour, but it should be the
kind of user's choice.
But, like what you were sayingearlier, I think all of these
things should be in every kindof situation.
If you're not doing the rightthing, you should be highly
warned and you should have toliterally tick a box to say I
know, if I do this, this couldgo horribly wrong.

(01:16:48):
But I know what I'm doing andhopefully it shouldn't do.

Robbie (01:16:52):
It's like when you set up a PC for the first time to
use a reference from earlier.
I think anyone that's everbuilt a PC DIY or does it more
regularly how often have you hada motherboard on a table
connected to a PSU on a table,connected to a hard drive on a
table and the case Well, thecase is still in the shop, the
case is still in the shop andtherefore in that scenario,

(01:17:12):
anyone else you saw doing thatyou would lose your mind.
But because you're doing it,it's okay and I think it's on
the back of deadbolt, when QNAPdid effectively a pushed update
on lots of users as a reactionto deadbolt because their
handling of the situation wasnot great.
But they did a push update andupdated lots of people's servers

(01:17:34):
to the latest server revisionto stop further users being
impacted by deadbolt.
That weren't the first time.
But the problem is, if you wererunning lots of VMs in your
business, if your server isn'tone of these five minute
turnaround reboot.
It's actually a quite complexreboot scenario where you had
ice guzzies suddenly lost, youhad dynamic IPs that you

(01:17:55):
probably shouldn't have had onDHCP, that all of a sudden have
changed our identity, and youhad VMs and more that hacked off
even more people the way thatwas handled.
But again, it comes down tothat whole should you update or
not, and should you be forced,should you be allowed, to do it
your own way?

Ed (01:18:11):
Why do you think people don't like to update?
I love doing an update.
As soon as there's an update, Iget quite excited and I like to
click and do an update.
Why do you think there's a lotof people who don't like to
bother to update things?

Robbie (01:18:24):
Well, I think part of updates it's that whole business
of security updates and featureupdates.
And I say that as a man who, inthe bottom right of my screen
right now, is still beingoffered Windows 11.
And I'll tell you right now,it's still going to be quite a
long while before I upgrade toWindows 11.
Now, why is that?
Because I'm used to Windows 10.

(01:18:44):
I didn't want to give upWindows XP.
If I'm honest, I still thinkabout it.
I'm not sure if it's under mypillow, but it's the idea that
there'll be change.
I feel like there'll be thingsthat don't work.
There'll be adaptions andchanges or maybe even a little
bit of modded script that Ithink will not work in this
newer version.
And in terms of NAS, we'vealready seen from all of the

(01:19:05):
major NAS players I can onlyspeak on behalf of Turnkey.
I've not followed Unraid orTrueNAS Edwin near as close but
there has been certain featuresthat are only available in old
revisions, so Synology with AI,photo recognition, just certain
apps and services that theyreach a certain point and then
they get downgraded or removedfrom the package center so you

(01:19:26):
can no longer get them anymore,and some of these sometimes are
the reasons people bought thesesystems, or at least major part.
So, at least in the case of NAS, that's their reasoning.
Because they don't like change.
They're worried about certainfeatures no longer being
available or not functioningright.
So that's the main point ofview.
Although I staunchly tellpeople to update, it makes me a

(01:19:48):
big old hypocrite when I knowthat that Windows 11 pop up is
getting ignored massively.
Or I'm using my web browser andChrome has suddenly turned
around and said you know,there's an update and I'm like
I've got 18 tabs on the go.

Ed (01:20:02):
I don't want to refresh these.
I think all of us we're guiltyof.
We say do this.
Like I always tell you, youmust do a backup.
I always do a backup.
And then maybe kind of threemonths ago, I lost all of my app
data and I didn't have a backupbecause it well, I had a backup
, but it was months old, becauseI was thinking, okay, this is

(01:20:26):
only temporary, it's like this,I'll do it later.
And I was a big hypocrite and Ipaid the price.

Robbie (01:20:34):
But I think it's that idea about I'm not going to do
the whole, do as I say, not whatI do, but we still need to
acknowledge these are the rightthings to do with the updates.
But sometimes life gets in theway, I think.
But it's when people make achoice to actively ignore these
things or to actively decide notto pursue them.
Now, that isn't to say thatwhat happened?

Ed (01:20:55):
I think as well, robbie, a lot of people you know, and I
can imagine it would be more sowith people who buy.
Maybe a turnkey now is like Iwas speaking to my brother the
other day and he wanted to setsomething up for CCTV and he was
just saying, ed, I just want tobe able to set it up and forget
about it.
I haven't got time to worryabout doing this, that and the

(01:21:17):
other.
And I think a lot of peoplelike that, they want to just put
it in the cupboard and theyforget about it.
And it might be months andmonths.
And I think to those people thatare listening to do that get
your calendar, put a date sixmonths in the future, on a
Saturday morning or something.
Say, I'm going to go to my NAS,I'm going to check if there's
any updates.
I'm going to do this just likeyou would when you schedule a

(01:21:39):
service for your car.
Now, you wouldn't.
You wouldn't keep driving yourcar, you wouldn't buy a nice
brand new top of the rangeMercedes or something like you
buy a top of the range NAS oryou build yourself a top of the
range server.
Well, if you bought a top ofthe range car and you just drove
it till it broke.
It would kind of be your fault.
You think, well, I couldn't bebothered to change the oil, I

(01:22:01):
couldn't be bothered to checkthe tyre pressure.

Robbie (01:22:04):
I feel I feel like I'm dog piling on users a little bit
.
I think it's important torecognise that one when people
do update.
It's only recently that we'veseen nuance in the way updates
are delivered.
So it used to be that it wasall updates, some updates, no
updates.
Now you can go for securityupdates or feature updates.

(01:22:26):
You can actually be moreselective.
And also in the case ofdeadbolt, for example, it was
recognised online.
I never really got it 100%confirmed that the vulnerability
with at least two platformsthat were impacted by it were
vulnerabilities within specificapplications which were found
because they obviously built onLinux and Linux can only ever

(01:22:47):
stay one step ahead of all thevulnerabilities anyway.
But in those scenarios, yes, ifpeople had updated, they would
have patched that vulnerabilityas soon as it was recognised.
But then there is a window ofthe vulnerability that was in
the back end of certain apps andthe attack vector having the
opportunity and in the case ofsome brands, during the deadbolt
ransomware impact, whichimpacted several brands.

(01:23:08):
In some cases that time wasweeks, if not months, between
the vulnerability and it beingclosed and in that area that
person could have had all theupdates on, but they didn't
matter.
So in some cases I don't wantto dogpull on users too much,
because there is an element ofbrand's responsibility and

(01:23:28):
user's responsibility withregards to ransomware and in
that case the brand did letpeople down and I think now it's
going to be quite a few yearsbefore people give QNAP, you
know, let them wash their handsof everything that happened.
I think it will be a good six,seven, eight years minimum.

Ed (01:23:46):
Can you give just a brief overview of what the deadbolt
ransomware is?
I'm assuming deadbolt is likethe name of the group that did
the attack.

Robbie (01:23:54):
Absolutely.

Ed (01:23:55):
Yeah.

Robbie (01:23:56):
It's all it did and a lot of people when it impacted
them, their first instinct wasto pull the cable out the back
of the NAS from their router orswitch.
But all it was was an injectedline of code to via SSH I
believe I have to double checkthat, depending on the
application and its range ofservices that was allowed to

(01:24:18):
have access to it just injectedand said basically, create a zip
of all of these files, of everysingle file that might be a
three meg photo, turned into athree meg zip, that was a Z zip
as a Z zip, seven zip perhaps,and then from there the actual
encryption key was created andthen purged from the system and

(01:24:39):
then create a word document, atext dot, txt or a rich text for
format, and then just createthis ransomware.
And that was it, and it was allin one injected command.
It wasn't installing a malwarethat sticks 8000 yahoo toolbars
on your web browser and yousmash your laptop into the sea.
It was one line of injectedcode.

Ed (01:25:00):
And that will still keep running even without the
internet plugged in, because itwill just chug along and encrypt
all of your files.

Robbie (01:25:06):
And the worst thing was that when it happened what they
did to get people's attentionthe last part of the ransomware
script replaced the usual loginscreen for your NAS.
So it was replaced by thedeadbolt screen that went your
system's been hit, you can't getin.
If you want to get in, go tothis key or go to this URL, make

(01:25:28):
us a payment.
I think it was like 1.01 and1.005 Bitcoin.
It's like 500 Nica to a grandor something.
Yeah, that doesn't look cool.

Ed (01:25:37):
Didn't they?
In the ransomware, they gavethe use, the end user the
opportunity to use the end userto pay to unlock their files.
But also they tried to getmoney from QNAP themselves
saying if you give us 5 Bitcoin,which must have 100,000.

Robbie (01:25:53):
It was 15.
I think I might be wrong.

Ed (01:25:55):
I think it was 5.4.
We'll tell you about thevulnerability, and it was 50,
and we'll give you thedecryption keys for everyone.
So QNAP could have paid whatthat would have been about a
million and they could have hadthe decryption keys and then
decrypted everyone's files forthem.
But obviously QNAP didn't wantto do that and I kind of really

(01:26:16):
agree with that.
You know, a lot of people maynot agree with me, because I
think when a company startsdoing that, it's going to
encourage attacks.
When one company's paid, thenit's going to encourage more of
that type of attack attackingthe end users to hold the
company to ransom, which itseems like they were trying to
do as well.

Robbie (01:26:35):
Also, there is the question mark of would they have
handed over the key?
Exactly it's a simple.
I mean there is no guarantee ofthat anyway, of them handing
over the key afterwards.
But a lot of users during thedeadbolt ransomware attacks,
their instinct obviously quickturn off the machine and the
problem was for some user casesit completely removed their

(01:26:58):
ability to interface with theNAS because during the
transition, during theencrypting of those files, to
produce the GUI to log in thenew login screen it wouldn't
generate, so lots of userscouldn't log into their system.
So the first thing you sawduring the deadbolt attacks for
the first month, month and a bit, depending on the brand that

(01:27:20):
was impacted, the brand actuallyhad to roll out and show people
how to get in to the GUI oftheir NAS because lots of people
would just turn the system offor when it would turn back on,
in some cases the encryptionwould just then continue because
it wasn't an active installedlike API or anything, it was
just a command that had beendelivered to the system.

(01:27:40):
So yeah, I mean the impact ofit is still being felt now
because there's lots of usersobviously that still hold QNAP
in disdain for that and againthey're going to have to spend a
still quite a few more years towash that off their hands.

Ed (01:27:54):
I heard Robbie as well, with the QNAP situation, that their
malware scanning software wouldactually remove the text file
telling people how they canactually pay.
What are your thoughts aboutthat?
Do you think they should havedone that?
Do you think that was takingaway the opportunity?
If someone wanted to try andpay to get their photos back?

(01:28:15):
They were literally taking awaythat opportunity for them.
And I don't mean just photos,they're encrypted files.
What's your thoughts on that?

Robbie (01:28:26):
I think the problem is, without knowing QNAP's intent,
whether that was an actualintentional action or a
byproduct of the malwareremoving tool finding that file.
Now, if it did just flat outdelete it, I think that's
terrible.
If it just quarantined the file, like a lot of antivirus or

(01:28:50):
malware scanners would do, togive you the option to go we
found this file, what do youwant to do?
That would be different, butbecause I don't know, I'd be
reluctant to comment on that.
What I would say, though, is ifit did just delete that file
and at least take away people'soption to do it, that's pretty
disgusting.

Ed (01:29:06):
I believe there's actually some type of file where, if you
have deleted your instructionsof where to pay that's actually
provided by the ransomwarepeople you can put one of your
files, you can have it scan oneof your files and it will retell
you where you have to pay theransomware to.

Robbie (01:29:26):
I didn't know that.
No, I know there was one.
To identify which ransomwareyou've been hit by, which, to me
, is one of the most depressingthings I've ever heard.
If you've been hit, the firstyou need to know is which one it
is, so you upload the file tothe website that goes sorry,
you've been hit by skull thunder.
It's terrible, but you know,there you are.
I mean, what a rubbish bit ofnews.

(01:29:47):
When deadbolt hit Acer storenazis and again, that was 2021,
I believe there was a guy thatworked in an office near me who
was a photography, was running alot of marketing agency stuff,
and he lost six years of hisportfolio to this Now, on the
one hand, I would love to say tohim well, where's your backup,

(01:30:09):
mate?
And again, no one wants to hearthat.
But at the same time, he'd setthis device up, he'd updated the
firmware pretty frequently.
It wasn't even that old afirmware he'd been hit by, he'd
been pretty Johnny on the spotabout it, but he'd still lost
all of that data.
Who's to blame in that scenario?
Is it Acer store for notrigorously tightening their
software for vulnerabilitiesthat they're not aware with

(01:30:31):
which are built on?
You know Linux, which you know,reports vulnerabilities pretty
frequently that everyone elsethen has to build on top of.
Is.
Or is it his fault for nothaving a sufficient backup in
place?
You know, we're all there.
There's poo on everyone's shoes.

Ed (01:30:46):
There are.
We all know everyone shouldhave a backup.
And you know I say like it'sthe old saying, raid is not a
backup.
But if you've got a file and ifyou delete that file you can
never get it back, you haven'tgot a backup.
So you know, on an unraidedserver, for instance, if I've
got 20 terabytes of files and Idelete all of them, if I can't

(01:31:08):
get them back from anywhere else, I don't have a backup.
In my opinion, a greatprotection against ransomware is
snapshots.
If you're running ZFS, ifyou've got a snapshot, you can
click a button and you can justbe back to where you were
straight away.
And now we've got ZFSofficially in on-raid.

(01:31:29):
I would say to anyone for yourreally important data, make a
ZFS pull, put your reallyimportant data there, make sure
it's snapshotted.
So if you ever did have aransomware attack, you can at
least roll it back to before thetime that the ransomware attack
happened and you would be fullyprotected.
But again, snapshots aren't abackup either, but it does allow

(01:31:53):
you to roll back to how it wasbefore the attack happened.
It's very powerful.
So anyway, I think that'senough talk about ransomware and
depressing things.

Robbie (01:32:10):
I think let's let's lighten up a bit.

Ed (01:32:15):
Let's talk about nuclear war .
No, let's, let's.

Robbie (01:32:20):
I'm glad this isn't my channel, volsgan, and let's talk
about a little bit of nuclearfallout.
You, karen, what's your jacketmade of?
Asbestos, great.

Ed (01:32:33):
So we've got some audience questions for you, and the first
question is for you, robbie, iswhat is the best NAS right now
for Plex in your opinion?
So the best media server NAS.

Robbie (01:32:46):
Well, price is going to be a big deciding factor.
So if I break it down intopricing, that NAS, I mentioned
earlier that flash store for 499, that's a good.
At the moment I'm sort ofplugging that for a good middle
ground, not with standard.
It's got a media mode on, itsupports Plex and I've got it
transcoding 8K.
It means that even if you don'town 8K media now, or if you own

(01:33:10):
8K media but your TV and allyour other devices don't support
8K, it can play and transcodeeverything down to, you know, 4k
and 1080p and stuff.
I mean, if your budget is 2 to300.
Nica the Synology DS223 is nottoo shabby.
We're running Plex on that in acontainer.
But again, that's a very lowend entry level.

(01:33:31):
The problem with saying thebest is every user case is
different.
There could be someone that'srunning.

Ed (01:33:36):
It depends how much media they've got and how many people.
If it's just them accessing itin their house, like them and
their wife watching together,that's going to be very
different than if you've got 20grandchildren, three aunts and
100 terabytes worth of media.
You're going to need somethingvery different.

Robbie (01:33:56):
And then when you factor in file formats like HEVC not
being supported on certaindevices, then you've got your
audio files, the ones thatsimply, simply simply have to
listen to Led Zeppelin in thatoriginal file format.
That's a flak rip.
That's absolutely insane.
Like for those users, you'regoing to have to go minimum

(01:34:16):
Pentium, maybe an i3, becauseyou need that integrated
graphics.
You can roll in a graphics card, but it's for Plex, it's not as
efficient in that regard theamount of power output you're
putting in and the noise.
I mean for those users.
Basically, the entry point Iwould say a Synology would be a
DS920, then if you want to go abit better, you go for something

(01:34:37):
like a QNAP74 series and theygo from Pentium all the way up
to an i9 12th Gen.
All of these are systems whereyou're trading off the power of
the device versus the cost ofthe device, versus the noise and
the you know TDP and generalpower consumption of the device.

(01:34:57):
A lot of those users that havelarge media collections, those
users are going to be very awareof a server that's on 24 seven,
making a lot of noise andhitting the old electricity bill
, particularly now midwaythrough 2023 and energy crises.
So, yeah, a 920 entry pointfrom Synology and the 7.4 series

(01:35:18):
.
I would say for anyone thatwants to take it seriously, or
if you've got a couple of M2NVMe's from an old laptop
knocking around that you want toreuse, look up that flash door,
because that flash door again,you can get the six bay model,
stick one M2 NVMe and startracking them up and you can get
M2 NVMe's even with QLC NAND nowthat up to 8 TB in scale, and

(01:35:40):
then you add to that.
For me that kind of is themiddle ground and it's
potentially going to be myrecommended Plexna server of the
year when we do our best ofs inDecember.

Ed (01:35:51):
Looping back to when we're talking about SSDs.
How do you find the actualdurability of running an SSD
server compared to a mechanicaldrive server?
Are they more reliable or lessreliable?
Would you imagine you would see, say, a SATA SSD that's nine
years old sitting in your NASbox behind you in the same way

(01:36:11):
from a mechanical hard drive?

Robbie (01:36:12):
I can see it now because production in NAND I mean again
.
I'll move slightly away fromNVMe, but I will circle back to
that in a second.
But when it comes to SATA, Imean again, we're seeing a lot
more SATA SSDs arrive, where theNAND production is being built

(01:36:32):
to NVMe level utilization butbeing used in a SATA SSD.
So you've got a SATA SSD that,guns blazing on its day, can hit
500 megs if you're lucky, butthey're using NAND that's being
developed at production for SSDsthat can crank out 7 to 8000

(01:36:53):
megs.
So, consequently, the NAND isbuilt to a standard, or at least
to a production qualitystandard, to withstand so much
more than it did before.
Nand is considerably moredurable than it has been for a
long time and we're seeing a lotmore SSDs arrive with a drive
rights per day DWPD of exceeding1.0, which means you could get

(01:37:16):
a terabyte SSD completely, fillit, delete and rewrite one
terabyte on this one terabytedrive every day, and it would
still last five years, accordingto the manufacturer.
So yes, it's substantially moredurable than they've ever been.

Ed (01:37:30):
Moving on to the next question for you, robbie, is
will you stop recommendingWestern digital disks after the
WDDA and the 26 to 80 hourswarning?

Robbie (01:37:42):
Well, so that's what we discussed earlier on.

Ed (01:37:44):
It is.

Robbie (01:37:45):
I will.
I still continue to recommendWD drives, but nowhere near as
much as I did.
I've got to.
When I talk about WD red drivesnow, I talk about them with a
caveat.
In the minute you're talkingabout something with a caveat,
you're doing it because you wantpeople to be fully aware of
everything.
But there's no denying that I'mgoing to add the caveat for a

(01:38:06):
long time that, oh, you're goingfor a WD red drive?
Oh yeah, actually they're agood drive, as long as you don't
go for the SMR ones.
Or if you're going to go for aWD drive in your NAS, great,
just make sure WDDA is disabledby default.
Or if you can't disable it,just know that at three years it
may spit a warning at you thatyou can largely disregard, and I

(01:38:26):
hate adding caveats like that,but it's sort of their own fault
.

Ed (01:38:32):
And the next question, which is kind of in a similar vein to
the previous question, isthey're saying why do you
continue to recommend QNAP NASafter Deadbolt?

Robbie (01:38:41):
With QNAP and Deadbolt.
I think I recommend quite a lotof QNAPs to be kept offline.
I also think that the brand hasintroduced a lot of network
security precautions and changesto default structure within
their OS.
I still maintain in every QNAPreview towards the conclusion,
you know, a highlight Deadbolt.
If you look at any of my beforeyou buys, generally you'll find

(01:39:04):
that the third or fifth pointis always they were hit by
Deadbolt and I think I'm goingto be doing this probably till
the end of the decade becausethey need to have an unbroken
record.
Now, time was when I talkedabout Synology a long time ago.
I would highlight Synolockerbecause they I can't remember it
was 2012 or 2014, back in thevery early days of me really

(01:39:26):
getting into NAS, they were hitby our malware.
They were hit by Synolockerthat was locking up the system
to try to make money out ofpeople.
Now, synology they dealt withit I would argue, slightly
better than QNAP dealt withDeadbolt, but they were still
hit by it.
But now you look online andthere's not much information
about it, and that's becausethey learnt their lesson.

(01:39:46):
They carried on, moved forwardreally rigorously, changed their
security protocol and, althoughsome of the decisions they've
made, such as streamlining USBcompatibility, changing a lot of
the applications you can use ontheir system, all they say in
the aid of security, and keepingtheir system as robust and as

(01:40:07):
protected as possible.
We can't deny that in a decadesince Synolocker might be give
or take that there hasn't beenanything big like that happen
again.
There's been the odd thing popup on their security advisory,
but that's really it so for them.
That's why I don't bring it upanymore and I will continue to
recommend QNAP, but much like WD, with caveats, and I'll tell

(01:40:27):
people to use a QNAP withvigorous security protocol.
I will, you know, I will applaudthem for them changing a lot of
their QTS platform, disablingSSH by default, disabling the
admin account, you know, notallowing users to make insecure
choices without fullyacknowledging the.
You know the risks.

(01:40:48):
And they, you know, included alot more third party VPN support
on their system and again Imentioned it earlier on, with
integrating Towskow.
They've done a lot of work tointegrate those things in.
They even rolled out their ownbounty program, like Synolocker
did.
But they've still got blood ontheir hands as far as I'm
concerned, and they will alwaysbe presented with that caveat.
But if you're looking at a NASsystem that you want to get a

(01:41:12):
turnkey solution to put on rateone or Thunderbolt NAS or any
number of hardware solutionswhich are just simply not
available to be purchased offshelf, qnap's a great option for
those, but as long as peoplestay informed about it, as long
as they know the past andthey're aware of the risks with
regards to their own choices andsometimes in the way these

(01:41:33):
devices are set up.
That's why I still continue torecommend them.
Start with caveats.

Ed (01:41:39):
And the next question on our list is at what point should I
start considering a NAS insteadof Unraid, or vice versa?

Robbie (01:41:48):
Right.
Well, I think we've coveredmost of this, but I would say,
because NAS is our much moreuser friendly and Unraid is by
far the most user friendly, Ireckon there's a lot of users
that can use Unraid and neverhave to touch command line
unless they really want to thesedays.
I think there's been a lot ofreally great stuff done with
Unraid there, but it still isn'tquite as user friendly as

(01:42:09):
turnkey NAS solutions and that'sdone by design because they
want to target you know againthat money, rich, time, poor
user.
The levels of support affordedby NAS brands is probably better
.
So if you need direct supportfrom the brand, you generally
get.
It's easier to talk to someonewho's a member of that brand

(01:42:32):
more quickly.
I think, on that larger scalelevel, the range of client
applications as well.
So the range of applicationsfor your iPhone, your Android,
your desktop there is justenormous number of client
applications that once you gointo open source platforms and
again, I know Unraid isn'tstrictly open source, but it
still lives within that bubble,I believe the number of client

(01:42:56):
applications to keep things easyfor you to interact with it.
There's not as many diverseoptions out there.
So it's the level of hardwarefor what you're buying.
The amount you put in to buyinga create your NDRI server can
be intimidating, and if youdon't know what you're doing,
building it, you don't get, youknow, a bit of silicon under
your fingernails.
You don't want to, you know,have to deal with a bunch of

(01:43:17):
warranties.
And NAS is a good choice, andthat's why I think for those
that want an easy first stepinto private server ownership
I'm back to the first questionyou asked about going cloud to
NAS, and NAS is a great choice.
However, once you know moreabout NAS and you become a lot
more versed, you don't have toget to enthusiast level.

(01:43:38):
You just have to use one for awhile to suddenly find
limitations in a NAS which arethere by design.
They are there to stop youbreaking the device, stop you
putting it in a state wheresupport would be required, and
for those users that suddenlystart feeling the glass ceiling
of a NAS, that's when thingslike unraid are unparalleled.

(01:43:59):
When it comes to unraid, itgives you the probably the best
middle ground there in terms ofadaptability and customization,
but also in terms of low costand access as well, and, I think
, for users that have owned aNAS for a while as a hobbyist or
owned a NAS and then suddenlystarted feeling the glass
ceiling and the limitations andyou feel like you're getting

(01:44:21):
treated a bit like a baby.
For those users that's when thetransition to an unraid server
makes a lot of sense, becausethose are users that have now
become skilled enough thatthey're feeling the limitations
of some NAS systems.
If you're an enterprise userit's slightly different, because
that's when you need somethingthat's certified for your needs

(01:44:43):
and maybe your HR not your HR,but I think in your accounts
team needs you to sign off onthis product and you need it to
be integral in that scenario.
But you don't want to spend alot of time reconnecting starter
cables and stuff where DIY isnot an option.
Until I ever see an unraidturnkey solution, those users I

(01:45:05):
would probably direct towards aNAS, a turnkey NAS, unless
they've got a good intelligencesystem.
Admin on site.

Ed (01:45:12):
And moving on to the next question.
Now I just want to point outthese aren't my own questions.
They're all from the website.

Robbie (01:45:19):
So, oh, I love this.
This is going to be brilliant.
What hatching job is this manalive?

Ed (01:45:26):
So the next question is do Synology and QNAP pay you to
make videos for them?

Robbie (01:45:35):
No, to date and I think I've talked about this before I
was going to do an Ask MeAnything video on this.
I've been paid by three brandsever and I'll get to them in a
bit but Synology and QNAP havenever paid me.
They send hardware upon myrequest.
In most cases I never get tokeep it.
So I'll sign a kind of anagreement where I keep hold of

(01:45:58):
the hardware for a loan period,depending if it's a particularly
sophisticated piece ofequipment, for example the
TS855X.
I'm doing a review on that.
I've got that for maybe ninedays.
It's been taken away from me onFriday.
I've done a bunch of videos onit.
I've done some reviews.
They never impact my reviews.
They never tell me oh, don'tsay that.
And indeed, even when I didthose videos where I was

(01:46:20):
installing TrueNAS, unrayed andProxmox and more on QNAP systems
and you know, mucking aroundand putting unofficial memory
and Synology systems, not oncehave they gone.
We're not sending you anythingagain.
Get out, get out.
They never did that.
They've never paid me.
What I get from them is accessto the hardware and independence
.
Now, if a device I believe I'mgoing to be using for a long

(01:46:43):
time, you can probably normallysee them on the left hand side
of the screen when I'm making avideo, those ones I've got on
long term loan or I get to keepthem.
So some devices the brand willsay we're sending it to you and
you can keep it, but I've neverbeen paid by them.
Generally that's called an MDF,a marketing development fund by
the brands.
I know because I have beenoffered by different brands

(01:47:04):
money in the past and that'spaid for from their marketing
budget to promote their product.
But once you do that, it's a.
You get told what to do and Iknow some platforms do that and
do engage in that and it can bevery profitable.
But we generally go with adsand we go with affiliated
marketing hits.
Why, at the end of every videoI talk about links in the
description that only use themif you've actually liked the

(01:47:27):
video, and I find that to be thegood middle ground.
And the three occasions we havetaken payment, we took one
payment from Buffalo.
They had just released theirWindows storage server devices
and they wanted us to do a fewreviews for it and we flat out
said no because we didn't havetime.
And they were adamant.

(01:47:48):
They wanted us to make somevideos on it.
And we were like we simplydon't have time.
The only way we'll do that isif you make it financially
viable for us.
And it was very in the early,very much in the early days.
But we went independent.
So we did a three part serieson WSS, buffalo, nas and they
never got to tell us what wecould tell and I think, the
third video.
We were quite scathing.
The second payment was withTerra Master, to have their name

(01:48:11):
at the top of our site alongthe top bar, and we said we
can't put you up there becauseno one really wants to click on
your button and we're running awebsite here.
And they were like well, can wepay a retainer to have our
little menu there at the top?
And we're like yeah, sure,knock yourself out.
So that was the second one.
And the third one was one fromthe end of last year.

(01:48:34):
I contacted WD and said oh, youknow your WD Red Drives.
They're going to be 10 yearsold.
I'd love to make a video on it.
And they said that'd be great,we'll pay for it.
And I was like well, no, I wasgoing to make it anyway.
And they were like no, no,we'll pay for it We'll pay.
Well, you know, we'llfacilitate it to be in this way.
And I foolishly said, yep, sure, and you can watch the video in

(01:48:54):
the article.
It's covered in the wordsponsored and they wanted it
done a certain way and I wasn'toverly pleased with the result.
If I'm honest.
I like the video, it's quitesnappy, but I would have gone
into a lot more depth about thewhole thing.
And those are the three timesI've ever taken payments from
brands, and that's.
I thought every time we did itto upgrade the mail server, to

(01:49:14):
upgrade the website, to covermore traffic because it was
starting to fall over a lot, andthat's really it.
And you know I've not been paidreally by any brand, that's a
long answer wouldn't it.

Ed (01:49:26):
And second to last question why are NAS drives so expensive?

Robbie (01:49:32):
One word software.
It's the NAS.
Software is what makes themexpensive.
If you broke it down to thehardware, if you tried to build
the majority of NASs barring afew exceptions take for example
that Plex NAS I mentionedearlier, the QNAP 74 series If
you were to build that NAS,which cost two to two and a half
grand to buy, you couldprobably put it together for

(01:49:53):
about 12, maybe 13, 30 and ahalf Nica.
But you would that's if you'regoing for a specific server case
, that's, a six or eight by notjust repurposing an old PC tower
and getting like caddies.
The reason you're paying forthat software is for that access
for all of the utilities forthe long term, firmware updates
for that software.

(01:50:14):
And that's why they'reexpensive, comparative because,
unlike building your own NAS orgoing again for one of those top
10 type solutions or that storeacts a kickstart that's still
in progress, all of those onlypaying for the hardware and a
bit of labor lasered on top.
Whereas a NAS there's a lotmore software to it.
So it has to be a combinedhardware software solution.

(01:50:35):
None of the brandsrealistically sell their
software on its own, althoughyou can buy virtual VM versions
of that software that run ontheir platform.
But all of them have to be ahardware software combined
solution and that goes all theway down to like one by arm, 512
meg that's right Meg ramsolutions for about 100 Nica

(01:50:58):
Even that 100 quid or so, notincluding tax, the 20 Nica on
top.
Some of that is for thesoftware, so it's always the
software that's making thoseprice tags higher.

Ed (01:51:08):
I'm not sure if we should interpret the question as well,
Robbie, from the person whoasked it.
I'm not sure if he's also maybeasking about the actual NAS
hard drives, why they're moreexpensive than regular ones.

Robbie (01:51:22):
Oh, that's hard.
Oh, we will have it's NAS harddrive.

Ed (01:51:25):
You know NAS drives.
I don't know if he meant NASdrivers in a in a NAS all in one
, or he's meaning a NAS driversand hard drive, so maybe we
should answer it both ways.

Robbie (01:51:36):
When it comes to NAS hard drives, I would argue again
it comes down to NAS harddrives and they're not that much
expensive, to be fair, perterabyte If you compare them.
If we're talking about thephysical NAS media, the price
between a 10 TB to get barracudaand a 10 TB iron wolf NAS hard
drive, the price difference perterabyte isn't actually that

(01:51:56):
much.
But that is a drive that's beendesigned to spin up, spin down
in a more sporadic environmentand therefore its durability has
to be scaled to thataccordingly as well.
You always find, I know,meantime between failure is a
bit of a I nearly said anegative word there not a great
way to measure drive lifeoutside of multi drive
environments, but it's certainlyhigher on NAS storage media.

(01:52:20):
And then NAS storage mediabeing in raid environments again
adds to the higher vibrationsensors being required there,
largely heat, more heat beinggenerated there and generally
more cash as well on a lot ofthose drives to deal with a lot
of the overhead and theinstructions that are handed to
and from the drive controllers.
I think that's mainly why NASstorage media costs more,

(01:52:41):
because it's just it's gearedtowards another utilization and
again not to use that incrediblytrite and very basic comparison
.
But if we look at cutlery in adraw, I think a butter knife
costs less than a proper steakknife because it's about it
being durable to the task andagain, that's an incredibly

(01:53:01):
basic simile.

Ed (01:53:03):
And also maybe Robbie as well.
Nas drives often have longerwarranty, so that has to be
factored into the price.

Robbie (01:53:10):
Absolutely yeah to a down fine point yeah, two to
three years to five years yeah.

Ed (01:53:14):
Right, our last question.
I probably shouldn't really askit because we've kind of
answered it anyway, but I don'twant to miss anyone out.
So is now the time to move overto SSD and NAS servers, and are
they durable enough?
How about just a yes or noanswer for that one, Robbie?

Robbie (01:53:32):
I can't just say yes or no.
I'm broken.
I'm broken as a human being.
You've exposed me.
No, mostly.
Yes, it's just it's still goingto cost you more.
And if you're not using asystem that can leverage at
least some of the performance,then no.
So if you go for some basic arm, one or two bay NAS, you're
never going to be able to fullyutilize those SSDs, even with or

(01:53:53):
without PCI lane restrictions.
So, unless if you're running anx86, AMD or Intel based, then
yeah, or you definitely arerunning better than one gig
network out, Because even thoughyou'll never really get to
fully realize the SSD speed andyou're going to oversaturate
even 2.5 gig connections, even10 gig depending on your rate

(01:54:13):
structure, you can at least makethe most of it and internal
with databases and moving a lotof the hard work away from the
CPU and memory.
This is the longest yes or noanswer you've ever received.
But yeah, I'm incapable,absolutely incapable.
But yeah, yes for most users,depending on the hardware
architecture of the NAS inquestion.

Ed (01:54:34):
Okay, well, thank you very much for answering those
questions, robbie.
That's the last of them, so Ithink it's a good place to wrap
up.
We've been speaking for quite along time.
I've really enjoyed talking toyou today, and I'm sure our
listeners are going to find itreally interesting hearing your
expertise.
For people that haven't seenyour website and haven't seen

(01:54:55):
your YouTube channel, where canpeople get hold of you and see
your work?

Robbie (01:55:01):
They can go to the South Coast of England and they'll
generally bump into me screamingoff the end of the pier that
RAID is not a backup.
Or they can Google NAScompares,because no one types URLs these
days.
Or they can find me online justunder the name NAScompares or
NASweirdo.
That'll do it, yeah.

Ed (01:55:20):
Also, guys, you can obviously see all the links in
the show notes below.

Robbie (01:55:23):
Perfect.
Thanks so much for having me on.

Ed (01:55:26):
Anyway, thank you very much for your time.
I hope you have a great rest ofyour day and speak to you soon.
Cheers.
Advertise With Us

Popular Podcasts

Dateline NBC

Dateline NBC

Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com

24/7 News: The Latest

24/7 News: The Latest

The latest news in 4 minutes updated every hour, every day.

Therapy Gecko

Therapy Gecko

An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.

Music, radio and podcasts, all free. Listen online or download the iHeart App.

Connect

© 2025 iHeartMedia, Inc.