March 10, 2025 • 43 mins
Tape storage is making a comeback. Hear why as SNIA experts highlight the evolution of tape storage and the importance of data formats and standardization in driving innovation within the storage industry. This discussion dives into the significance of security in data management, evolving roles of technologies like Linear Tape File System (LTFS) and the SNIA Cloud Data Management Interface (CDMI), and how collaborative efforts driven by SNIA shape the future of interoperability. You'll also learn how exciting new initiatives like the SNIA Cloud Object Storage Compatibility Plugfest pave the way for enhanced interoperability. Join this conversation on:
• The resurgence of tape storage in the AI era
• Standards to address cybersecurity challenges in the cloud
• Emerging data structures like directed graphs reshaping data access
• Security protocols becoming integral within data management frameworks
• How LTFS enables interoperability between different vendor solutions
• Call for new members to join SNIA and contribute to future standards
SNIA is an industry organization that develops global standards and delivers vendor-neutral education on technologies related to data. In these interviews, SNIA experts on data cover a wide range of topics on both established and emerging technologies.
About SNIA:
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:01):
All right.
Thank you everybody for joiningtoday.
My name is Eric Wright.
I am the host of the SNEAExperts on Data podcast and
co-founder of GTM Delta Superexciting, because we've got
three amazing humans that I'mexcited to chat with.
We're going to dive into sometech to some people and we're
going to talk about the idea offormat.
Now, format is a funny thing.
(00:23):
In fact, this is probably thefirst thing we want to start
with.
Is what do we say when we talkabout format?
Under the auspices of an amazinggroup of technologists of SNEA
coming together buildingstandards, we've also focused a
lot on how do we make sure thatour practice areas align with
what we call data focus areas.
So if you go to the SNEAwebsite, you're going to see a
(00:44):
lot of really, really neat stuffthat we're doing around making
sure that we're bringing downsort of columns and pillars of
where practice areas align, butwe're actually seeing a lot of
cross pillar alignment.
We'll get into that.
More than anything, let's getinto the amazing folks I'm
sharing the call with today.
So if you don't mind doing aquick introduction, let's start
with Curtis and then I'll workour way around.
Speaker 2 (01:06):
Well thanks, Eric.
I'm Curtis Ballard with HewlettPackard Enterprise.
I'm a storage strategytechnologist with Hewlett
Packard Enterprise, working onfuture storage technologies with
the SNEA organization.
I'm a member of the technicalcouncil where I help the SNEA
organization organize, run,oversee their technical efforts.
Speaker 1 (01:32):
You are one of the cat herders.
Welcome to the joy, george.
Speaker 3 (01:39):
Hey guys, my name is George Kambouris and, yes, it's
a Greek last name.
German accent People get soconfused by this Industry
veteran.
For more than 20 years workedfor EMC, dell, hitachi.
Right now.
I'm in that up right now and mymain focus is product
management engineering.
So my job is to define what andwhy and what is really a house
(02:02):
on this one.
I did work with Snea in 2005 onthe NDMP Twig back then long
time ago, and right now I'mworld director at Sneha driving
new standards with the wholeteam.
Speaker 1 (02:15):
Fantastic, and I love the potpourri of both accent
and background.
This is such a we've got a good, diverse representation.
And last but very certainly notleast, michael, if you want to
do a quick intro and then we'regoing to jump in.
Speaker 4 (02:29):
Awesome.
Hi, this is Michael Horde, andI've been in the industry for 30
years working with McDonnellDouglas, Boeing, Sequent
Computer, IBM, Seagate and Intel, and so I'm the chair for the
Cloud Storage TechnologyInitiative, and we have two
(02:51):
subgroups or they're differenttwigs and activities that are
going on.
One is the Cloud Storage Twigthat's working on CDMI, the
cloud data management interface,and then there's another group
of folks that we've just puttogether over the last year and
that's the Cloud Object StorageCompatibility Plugfest, and
(03:15):
we'll talk more about that later.
Speaker 1 (03:18):
Yes, remind me, definitely let's cover Plugfest,
because that's actually such afantastic format.
Let's cover Plugfest becausethat's actually such a fantastic
format, and I've participatedin some myself back in the day
before I was, when I was avendor and contributing on some
of the working groups, so very,very cool.
But, like as as James Liptonsays, we we start, as always,
(03:42):
from the beginning, which shouldbe.
What does it mean when we talkabout format data?
George, do you want to walk usthrough what we're talking about
in this?
Speaker 3 (03:49):
So to remind everybody about SNEA's main
mission, right, snea is, peoplesay, a safe space where
competitors come together tocollaborate in the favor and the
benefit of the end customerright.
We don't want proprietarycompanies which is great for
them to have something uniquefor them, but the customers are
asking for interoperability towork together on common
(04:09):
standards.
So SNEA is there to drive thisinnovation across competitors in
a safe space.
Just a reminder about SNEA inthe overall context.
So if you go, if you haven'tbeen there, you should go there.
On the SNEA org website there'sactually a section we call
Completed Standards and this ismentioned for a bunch of areas
innovation-wise, starting withDDF, the Common Rate Disk Format
(04:33):
, to LTFS right, for example,the Neurotank Pulse System.
We have a bunch of standardswe've done over the years.
The key is how we drivealignment between these
standards develop.
The key is how we drivealignment between these
standards, develop them together, roll them out.
And of course, there'sinnovation happening in the
market with new technologiescoming out over time and we have
to adapt and change with thesethings.
So format is one of the crucialareas when I work in SNEA to
(05:02):
standardize and work together.
So we categorize now under thecomplete standards, a new format
area to drive innovation andbring new science out of the
market.
Let's just mention CDMI as anexample.
We can deep dive into this aswell, but let's start with the
beginning first.
The standards are there todrive innovation across
competitors in the favor of theend customer.
Speaker 1 (05:21):
Well, that is the thing that I so adore about this
.
I've been lucky.
I've worked with a lot of, I'llsay, vendor agnostic
communities and even vendorassociated communities, where we
make a point of separatingbrand from purpose, and it is
such a great way that this comestogether, because we literally
are competitors by day sometimes, but we understand that
(05:46):
together we can innovate if wecollectively define standards
and the entire industry sees thebenefit of it.
And it's such a beautiful setof people that I've been able to
learn from and become friendswith over the years.
This room alone, you know,alone is proof of that.
So maybe let's talk about someof the innovations that have
(06:10):
happened.
Because you talked about stufflike LTFS, people will say, ooh,
I'm looking forward to LTO6.
May not sound exciting, but notrealizing that there is a lot
of innovation even in stuff thatis tried and true, standards
that are evolving, along withphysical, you know, and software
capabilities.
(06:30):
I'll let you guys go.
Maybe, curtis, I'd love to getyour thoughts on.
You know what excites you aboutwhat's being innovated on?
Speaker 2 (06:41):
Yeah, so you mentioned LTFS and people being
excited about LTO6.
I chuckled to myself when yousaid that, eric, people are
currently excited, I hope, aboutLTO10.
Speaker 1 (06:53):
We're getting way ahead.
I'm so far behind.
Speaker 2 (06:56):
Yeah, technology changes so fast you just can't
keep up with it.
So, for those that don'trecognize the acronym, lto is
linear tape open and it's a tapetechnology which a lot of
people think is gone.
I've worked for Hewlett PackardEnterprise for over 30 years
(07:17):
now, and when I joined thecompany, I actually worked for
the team that was directlyconnected to the group that did
tape products and people told medon't ever get put on one of
the tape projects.
Tape is dead.
That was over 30 years ago andjust in the last year I can't
(07:38):
count the number of times I'verun into people that I've told I
work for Hewlett PackardEnterprise and they've said oh,
hewlett Packard Enterprise, youguys have tape products, don't
you?
We've got AI solutions wherewe're creating lots of data and
we're looking for ways to saveit cheaply.
So tape is still very muchalive and is actually getting a
(08:01):
bit of a resurgence with all ofthe AI training and needs to
have cheap ways to archive datathat was used in AI and archive
data that's generated by AI ordata that you might want to use
for AI.
And LTFS is a standard thatSNEA worked on a few years ago.
It's not currently under activedevelopment because it's
(08:23):
successfully being used in a lotof industries today.
Ltfs is the Linear Tape FileSystem.
It's a format that Snykstandardized for writing file
system data out to any streamingdevice, generally a tape device
, any streaming device,generally a tape device in a
(08:44):
defined format that isself-describing and is
standalone in a single tapecartridge so that it can be used
across multiple vendors,creating data on tapes or other
streaming devices that can beused for archival purposes and
long after the vendor that wrotethe data is gone.
(09:09):
The standard developed by SNEAand published as an ISO standard
will have the specificationsand their open source software
developed around it that willallow customers to always read
their data.
I do not currently workdirectly with tape products so,
out of curiosity, as we werepreparing for this call, I went
out to the GitHub repository forthe LTFS software and thought,
(09:34):
hmm, what's going on out here?
It's really stable.
So there haven't been a lot ofrecent changes.
But the most recent change wasin a release published three
days ago when I went out andchecked it, so actively being
used every day, and SNEA's stillgot the expertise and is ready,
(09:58):
if there's ever a reason, toupdate that specification.
Speaker 1 (10:03):
As my favorite phrase is you call it legacy, I call
it production.
Speaker 3 (10:07):
Yes, yes, Sorry.
Speaker 1 (10:09):
George, I stepped on you there.
Speaker 3 (10:11):
I had a discussion with a customer BigBank I won't
say the name for a second, buttalk about how much tape they
actually have, right, and wetalk about exabytes on this
thing.
And he goes we have over azettabyte of tapes.
And I go, wow, right.
And they say, oh, we use theWorld War II silos, the bunkers,
(10:31):
they store these tapes.
The temperature is perfect forthese guys and they have no
intention to change this back toa point because they keep them
around.
But they have to be able to goback and read data from 10, 20
years ago.
Speaker 1 (10:48):
So it is a standard out there which is heavily used
and will continue to be used inthe future.
Right, Well, I'll tell youthat's one thing, that standards
are not just about what we'redoing now, but how we will
cleanly deprecate thesetechnologies and methods as we
go forward.
It's as important to thinkabout the transition between
formats and styles andtechnologies, because I mean,
mean and look, it doesn't evennecessarily require a, an actual
(11:09):
linear tape.
You know we've, there are tonsof products that build in this
virtual you know, front-endlayers so that we can emulate
tape, and that is again anotherway where we're going to see,
you know, know, cloud storagethat appears like tape but
doesn't face the penalties of,you know, archive.
You know, you make a requestand in a couple of days you may
(11:31):
or may not see that request comeback.
It's, there's lots of wayswhere we can use that interface,
but even higher speedtechnologies we may find that
you know some really fast.
We may find that some reallyfast.
Apparently, tapes are beingused in production for stuff
like data training and much more.
Speaker 4 (11:54):
I'd like to add a little bit too.
So, in terms of excitinginnovation, of course, the AI
and with regard to data format,of course, a lot of the new data
structures, and that's with thedirected graph or the knowledge
graph or vector databases.
So there's a lot of work inthat area.
(12:15):
A lot of it's driven out ofdata science itself, but when
you look at CDMI itself, it isstructured as a directed graph,
so it's already set up in termsof providing access, control and
then the set of functions thatyou can essentially build,
(12:38):
operations that you can buildout, the structure for directed
graphs, and that's thecontainers, the data, the
metadata and the operations tomanage all of that.
That's what's in CDMI today, so2.0, that was standardized in
(12:59):
September 2020.
And right now, david Slick isthe chair for the technical
working group for this cloudstorage twig, and so I talked
with him just a couple of daysago, and so the intention is to
bring to the table new datastructures.
(13:22):
So that's data format and thenthe potential for different
semantics with regard tolanguage support, protocol
support.
There's several NFS, smb.
Those are very well established, but there's others that we
need to look at.
So part of the effort is toassess the industry what are the
(13:43):
needs?
And, just like what was sharedearlier I think, george, you
mentioned SNEA is entirelymember-driven.
So everything we do comes frommember requests and when we see
a pattern where gosh, there's alot of folks coalescing, and
(14:05):
then we have that discussion asan industry group to then get
consensus and then start drivingcollaboration.
Snea is a third-party vendorneutral.
We put together forums,roadmaps and then the technical
initiatives and then finallystandards efforts where that's
(14:26):
appropriate.
So that's exciting.
The other aspect, just briefly,is on the data formats with
regard to multi-protocolgateways.
So when you're looking atactual on-prem to cloud and
multi-cloud, at actual on-premto cloud and multi-cloud, all of
this is so dynamic in terms ofshifting data.
(14:49):
So the data format integrity isto have a unified management,
but also the capability to haveconsistency and integrity across
all the multi-protocols.
And, of course, the existingcloud has a number of very large
entities, but those are alldifferent.
(15:12):
There's a lot of protocoltranslation that happens under
the covers and that's preciselywhat we're working on for the
Plugfest.
So I just wanted to give thatlittle point on that.
Speaker 1 (15:23):
Go ahead.
So I just wanted to give thatlittle point on that part Go
ahead, yeah, the plug fest andthis, the human cross.
Collaboration is what's amazing, and in fact I've seen a lot of
vendors who made the choice tonot run to the market with a
product but run to the industrywith a standard and then,
through developing thosestandards, then develop a
(15:45):
product that has, you know, away to interact by that standard
, because that ensures that theyhave partners and innovation is
shared.
Because you know, they alwayssay we, we don't want to
reinvent the wheel.
Well, we've actually reinventedthe wheel quite a bit, because
previous wheels used to used tonot be good like, or they were
fine at the time.
Speaker 4 (16:03):
but wheels have evolved, not necessarily at an
incredible pace, but it needs tohappen, or they become a
reference architecture and thenthe new foundation is built
thereon.
Speaker 1 (16:15):
Right, yeah, we forget about how much happens
that way.
Speaker 3 (16:21):
You just mentioned how people innovate and the
environment is changing.
Let me go back with Michael totalk about CDMI.
Yes, september was 2020 lastupdate, but the version 1.1 was
in 2011.
So why was there a need let'stouch this for a second for
changes?
Here's one example 18 monthsago, customers were saying how
(16:43):
do I prevent hackers coming in?
That was 18 months ago.
The last six months, all theCIOs, ceos in our customer
briefings are now saying what doI do after I get hacked?
How can I restore?
So the cyber attacks happeningleft and right are rapidly
increasing.
So the market shifted now fromhow do I stop hackers to how do
(17:07):
I react?
How can I recover fast?
Ransomware attacks is anexample right, huge out there
right now.
So the market has shifted.
As a result, we have to adjustthe standards as well.
So we went beyond than justsaying how do I store data?
How do I access protocols todata?
But how do I secure it?
How do I do key management?
How do I access acrossprotocols to ensure that
(17:29):
integrity?
Who has authentication,authorization, different methods
, right.
So the CDMI style has evolvedbig time over the last, I would
say, five years to adoptinnovation and security change
over time, security is becominga key point going forward.
Speaker 1 (17:46):
And security is probably one of the greatest
reasons we always say that.
You know, when DevOps became athing that dripped off the
tongues of all these you knowexcited pundits and CIOs, and
then along the way, about a yearand a half later, suddenly
comes DevSecOps and someone said, whoa, why do you have to like
explicitly put sec in the middle?
(18:06):
It was implied.
I'm like, well, it may havebeen implied, but they weren't
actually there.
So, yeah, we've had to go outof our way.
But in the SNEO, in the TWIGs,so the technical working groups
and a lot of the other parts ofthe community, the fact that
you're in the room with peoplethat are security experts and
they have software developers,so all of these people that in
(18:28):
production environmentscollaborate, but they never
collaborate in the planning room.
We're doing that now.
It's actually the successes insoftware optimization for things
like what we do with LTFS andCDMI.
Those software innovationshappen because they're sitting
in the room, with peopleunderstanding where the
standards came from and howother people are using it as
(18:50):
well, which that you know,george, you we've talked before
about this like the use case iswhat matters.
I'm a nerd, I love the bits, Ilove the speeds and feeds.
But in the end I really lovewhen I see like medical benefits
and human benefits andscientific benefits that we get
from doing stuff with thesetechnologies and these standards
(19:13):
.
Speaker 3 (19:13):
You just mentioned personas, right?
Let's touch on this for asecond.
We used to have the chiefsecurity officer, but now they
have a new title of chief dataofficer, for example, and these
guys are talking to senior aswell, saying look, my challenges
are 58 applications, I have toshare it across them.
How do I do it securely and nothaving to reinvent the wheel?
(19:33):
Back to your previous commentright, these guys are talking to
us about standards and CDMI.
One example is awesome forthese guys.
Let's talk about one use case.
Let's take healthcare as anexample.
Right, biggest challenge If youare not up to date and you take
your cell phone, you probablyhave a healthcare app on your
cell phone Most likely I have onthat one and I can talk to my
(19:54):
doctor in a secure chat and theycan make prescriptions in the
app online with my CVS pharmacy.
Whatever All this, automationhas to be standardized,
regulated, secure, gdprcompliant, hipaa compliant.
Automation has to bestandardized, regulated, secure,
gdpr compliant, tippercompliant.
The list goes on and on right.
To get to this point, it tookyears of innovation and cdmi is
(20:15):
the back end of this foundation.
So it's just one example of usecase we can say a start is
driving innovation to theconsumer, in this case, to make
our lives much, much easier.
I love this app.
So easy before call the doctor,make an appointment, go inside,
inside, get the prescription,try to see me.
Oh my goodness, right Now it'sall online.
My kids now are expecting thisto be a standard going forward.
(20:36):
Ten years ago it didn't exist.
So major innovation happening.
Speaker 1 (20:41):
Yeah, I would spend five hours to avoid a 15-minute
doctor's appointment any day andthe fact that work has happened
in the industry using standardsyou know, we always say it's
the standards well-defined meansthat the artifacts can change.
It's the interplay betweenthose artifacts that's the most
important thing, and it's noteven at the moment in time, but
(21:04):
the thing that we thought wasnot useful.
Like archive data, people arelike, ah great, what are we
going to do with this?
We store it for the seven years, which is actually not even a
real thing.
The seven years is the same asdon't yell fire in a movie
theater.
It's not actually written down,but we use it as a standard,
supposedly.
But we took all this data whichwe just said we'll put it over
in the corner just in case weneed it.
(21:25):
Well then AI came along and weneeded to train on company and
customer data safely.
Well, where do you think thatdata has been for a long time,
right?
So all of a sudden, the archivebecame primary storage and it
flipped on its end because theuse case changed.
Speaker 2 (21:45):
Yeah, and the topic we're talking about today became
way more important the formats,you know, the formats that
organizations like SNEA helpdefine.
Snea is not the only one, butSNEA focuses on formats of data
in storage ecosystems, and it'scritical that we have these
(22:25):
well-defined standards for howyour information is formatted.
Information exchange thatenables the technology that so
many people count on today, andSNE is here to help define those
formats.
And one of the things we'retrying to do today is look at
what's the next gap gap.
(22:46):
Where is the next format thatwe need to be defining to enable
that next important informationexchange opportunity?
Speaker 3 (22:55):
I have a comment, if I may finish the previous
thought on this one, which was agood example.
So back to my healthcareexample.
Let's first a little bit showyou how these things work
together.
Back to format data from Ericin the beginning.
Right, I just touched a bit onCDMI.
However, if you swipe your IDat the doctor, they have a
(23:18):
system in place with CDMI to goback to the tape device before
and pull up your x-rays.
So now we have CDMI workingwith LTFS together.
Number one this is someone'ssort of a system probably a NAS
device or something else storagesystem using DDF formats.
So now we have three formats inone healthcare use case working
(23:39):
together from SNEA.
So it's not just one example ofa standard.
These work all hand-in-handtogether.
Just to add as well,interoperability between systems
is very important If work allhand in hand together.
Just to have this as well,interoperability between is very
important.
If you haven't seen or haven'tbeen to our conference, by the
way, the State of Web Conferencewas in September in Santa Clara
and we had, for example, davidSilk give an awesome
presentation on CDMI withexample use cases.
(24:01):
If you haven't been there, nextone is coming up, one in
Colorado early next year andthen of October next year in
Santa Clara Again.
Please come there for that one.
But you just mentioned to usfutures.
Just to give you a quickglimpse of another example of
formats, I'm driving now a newinitiative within SNIA as well,
how we can start across NFS, smband S3.
(24:22):
Just an example it doesn'texist yet.
It's a brand new program.
It might not complete anytimesoon, but customers are asking
us.
I have compliance issues, Ihave auditing issues, I have
share of custody issues.
I'm working nowcross-functional SNEA across all
our competitors together tocome to the end customer view
and say how can we make theirlives easier to have one start
(24:46):
across these things?
It might take us six months,two years, it doesn't matter.
The point is the customers havea challenge.
We recognize the SME as a teamand we work on new stars going
forward, so we don't stop today.
We have new stuff coming allthe time.
Speaker 4 (24:59):
Right.
And then going back to the dataformat, that it's not just
simply the data itself, but it'sthe metadata that really should
be stored with the data, andthose should be very closely
coupled and go together asthings progress, wherever that
data is migrated or used acrossthe multi-cloud.
(25:23):
And that's actually what CDMIdoes.
And so when you look at whatGeorge was talking about the
healthcare example, what they'redoing as far as data science or
the data chief data officer, etcetera, they're looking at the
overall structure of all dataand that's the ontology, and so
(25:46):
it's literally at the highestlevel.
And then, in order to make thatsearchable, then they're using
knowledge graphs and algorithmsto then get to more precise or
exact results.
Instead of like with a vectordatabase, it's more of a
similarity approach algorithmand that is not exact.
(26:08):
So that's kind of the bleedingedge of using vast amounts of
data and with the tape that wasdescribed with zettabytes.
But it's really, how do youpackage that into a way that's
consumable and useful in ourlifetime instead of calculating
(26:29):
over decades?
It needs to be responsive, andthat's where the technology is
at right now, today, and I'mexcited about Sneha's play in
this.
Speaker 3 (26:40):
I love your example about AI Michael.
Do you want to touch base onsecurity for a second, because
we have seen the news?
Oh, chep CPT index, blah, blahblah without my authorization.
Can you touch on this onesecurity for a second?
Because we have seen the news?
Oh, chat, gpt index, blah blah.
Without my authorization.
Can we touch on this one, too,for a second?
Speaker 1 (26:53):
Yeah.
Speaker 3 (26:55):
So I'll start off and then, michael, maybe we could
complement this one.
You have seen in the news aboutall these ILM models out there
learning on data sets and youcan see all these actors.
You and companies now said youuse my data without my
permission.
You just touched base with this, michael, about security.
Right, the databases Want toelaborate maybe on the security
(27:17):
aspect of this one?
Speaker 4 (27:17):
Yeah, should we ask them this one?
Yeah, so with the and I'll justsay the example is CDMI.
It has all of the accesscontrols.
It has delegated access controlso you can use the proprietary
or the system of your choiceon-prem.
And so there's all the portsand exports in and out of going
(27:42):
down to ACES, which is accesscontrol entries.
So it's very fine, tuned downto the individual object level
data, they're able to look atthose aspects that George just
talked about.
As far as the heritage orlineage of that data, where did
(28:06):
it come from?
What are the protection methods?
Need to know who should beinvolved.
So all of that is somethingthat exists today.
So the next step is reallyapplying that into real systems.
Speaker 1 (28:21):
So go ahead, and this is what you really highlighted.
This thing of that definitionof the sureness of the format
means that everything else we docan always rely on that
consistency across platforms,across software, across hardware
platforms, across software,across hardware.
(28:44):
Even RAID is well.
Again, not a lot of innovationshappening in RAID, but yet the
hardware and softwareinnovations that allow us to do
more with understanding how tooptimize and extend the life of
hardware and protect anothergroup that we have as a
data-focused area.
I had a little mini server backI'm an older fellow so I've been
I've had used to have thoselike tiny servers started in
distributed computing and we hadextra ones that were left over
(29:08):
because they've beendecommissioned and I thought, oh
, you know I'm a nerd, I want todig in and do some new nt
windows, nt goodness.
So I found an old server andlike, oh, we had a stack of the
old drives and so I took fivedrives and I put them into the
tower and I powered on theserver to get ready to put in
the cd and it came up to theoperating system screen and I
thought, oh, and I logged inwith my regular production id
(29:32):
and it was like wait a second,because, just by sheer
coincidence, I picked up fivedrives that were originally from
it, stuck them back in and itbecame the server.
It was that six months earlierhad been unplugged and well,
security problem because wedidn't really think about how to
clean that data.
The fact that five disparatedrives had enough information to
(29:53):
be able to tell the controllerto rebuild the operating system.
In this it went back in placeand again it's like what do we
store with metadata?
Where do we store index indices?
Where do we do all these things?
Those common standards, then,mean that if we want to do
things like emulation andinnovation, it has to be done on
(30:14):
the basis of a foundation.
Cinder blocks are always thesame shape for a bloody reason,
so that we can build houses ontop of them.
Speaker 3 (30:22):
Exactly.
Speaker 1 (30:23):
That's boring.
Speaker 3 (30:25):
Yeah, that's right, you mentioned a lot of
innovation happening.
Let's touch on this for asecond because, as we might have
seen the press now Kiyokya, wd,samsung, you name it right they
just announced 120 terabytemedia on Flash 256 come running.
(30:46):
After this, we have seen thefirst pre-announcement of one
petabyte drives Media drivescoming right.
So, while RAID's not evolving alot, let's discuss, because
rebuild times is one of the mainthings people are worried about
.
If you have too long rebuildtimes, you're exposed.
Yes, there's erasure coding.
Yes, there's rate DP and tripletech.
(31:07):
These are all out there, but wehave to look at the market and
always evolve and adjust.
So you might see moreinnovation coming in the rate
group division about ratestandards going on as these
capacity evolve, so there mightbe another version coming out.
We never stop, we alwaysinnovate.
Just to make the point here,it's not dead and done.
We're done with it.
We will innovate going forwardbecause the technology evolves
(31:31):
over time.
It doesn't stop.
Speaker 1 (31:34):
It's.
I'll say SNEA is not unlike themoon mission that people think
it was a trip to the moon.
But what they don't realize isit was the beginning of decades
of innovation because ofresearch done on that trip to
the moon.
That's, that's why thesecommunities have to exist, and
even we talked about differentstandards bodies, iso, ieee.
We have cross, you know,organization collaboration.
(31:58):
That's going on.
We've had folks on the podcastalready that are doing work
within and across the standardsorgs, and this can be done
because built on standards meansinnovation happens faster.
I just want to hammer that likeI should put that on my
forehead.
Standards drive innovation.
It may seem boring to slow downand define it right at the
(32:21):
start, but boy oh boy does theflywheel kick in because of the
work that's being done by thisgroup.
Speaker 3 (32:29):
That reminds me, eric , a great suggestion to remind
our audience here.
We do want more members.
Okay, there are a lot ofcompanies out there, especially
on AI, for example, startcompanies popping up left and
right.
We live and die by our members.
As Michael mentioned before, wehave a great set of companies
(32:50):
working together already, but wewant more.
We need innovation.
We need people coming in todrive standards.
If you haven't been looking atSNEA yet, please look at our
website.
Look at our membership.
We just changed our structureMuch more easy to join SNEA now
and participate in certain workgroups at every level of this
one.
If you have suggestions on newstandards, don't think you're on
(33:12):
your own.
Come to us.
We're happy to work with you.
Bring you in.
We are open for new ideasuggestions.
So don't think we're closed off.
We want more members.
Just to point this out to theaudience here.
Speaker 1 (33:25):
The beauty of how we can safely share how we're doing
things that still allow us to,in our individual companies,
continue to work as the way wedo but the fact that you can
probably save a ton of time Justbeing in that room and seeing,
like you said, the talks thathappened at SDC and they're on
(33:45):
the YouTube channel as well.
There's quite a few that arethere, so folks should go check
those out.
You get to interact withengineers who are building these
tools and seeing real customerexperiences.
There's very rare opportunitiesto get in front of that kind of
community.
So, oh boy, I could spend allday with you guys.
(34:06):
But I definitely want to say,as a wrap for what we, what you
see ahead and you know,specifically around format, but,
sneha, in general, with 2025,curtis, I'll let you lead us out
what's got you fired up about,what's new and coming.
Speaker 2 (34:22):
What's got you fired up about what's new and coming?
Oh well, we've mentioned itseveral times already, but AI
(34:45):
has just introduced so manychallenges.
With every AI tool that theyuse, stores, provenance
information in a differentformat, and there's
opportunities for identifyinghow we can take some of these
tools that the world is buildingon for AI and start creating a
little bit more structure aroundhow those things can work
(35:09):
better together.
Especially, you know, there'sso many things throughout the
storage infrastructure behind AIthat, as everybody is focused
on the GPUs and the compute andthe processing, there's this
entire storage layer thatunderlies that.
You've got to get that datainto your AI processing and
(35:31):
there is so much opportunity forbuilding out that storage layer
for accelerating your AI worldaccelerating your AI world.
Speaker 1 (35:47):
And, michael, what's got you excited to wake up on
Christmas morning knowing thatyou've got a year ahead and
presents?
Speaker 4 (35:51):
to open.
Well, I'd have to say thatevery day is pretty much like
that, opening a brand newpresence every day through the
year, and the reason I'm sayingthat is this last year we had a
very remarkable activity withthe Plugfest.
There was a whole community puttogether and many of the
(36:12):
contributors are not traditionalSNEA members.
So we're reaching out intoareas for cloud storage.
So we're reaching out intoareas for cloud storage that and
(36:33):
it's not just the storage, butit's the gateways and other
infrastructure that goes on-premand various software.
So, and to get very specific,with the Plugfest, we conducted
that in September.
We had a November 21st webinar.
So I'd love to like, in yourcomments, we can publish that
link for that.
Speaker 1 (36:47):
Yeah.
Speaker 4 (36:48):
That was a report to the industry on the activity
there and that gives a lot ofthe detailsFest and we're going
(37:08):
through the review processwithin SNEA to create a new
organization for that, and sothat was just recent.
We've got our PlugFest callscoming up in January as well as
the open source test suite, sowe have separate calls for that.
And then just yesterday I had awonderful conversation with one
(37:32):
of the candidates we're lookingat for leading the open source
test suite development, and whyI point that out is what excites
me is to have thoseconversations where these are
time commitments and it's noteasy.
People are extremely busy, butwhen you have something vitally
(37:56):
important that is exciting andthen having that discussion
about okay, where is theindustry going, how can we
influence that and provide asignificant impact.
That's the kind of discussionjust individually we have, but
also with our forum we had abirds of a feather session
(38:19):
looking at strategies and thenthe final thing is I'm doing
some outreach to open sourcetest people.
So it's a very long list of newpeople.
So my goal personally is todouble the size of our test
matrix by April for the nextPlugfest.
Speaker 1 (38:39):
Fantastic, and that's where real testing happens.
I always say there's no greaterlie than things like I have
read and accept the terms andconditions and all the tests are
in the PR.
Thank you, yeah, nice,excellent, well, definitely,
we've got a lot ahead and I wantto make sure that we bring it
(39:02):
back.
So, george, my German and mybeautiful cornucopia background
friend, what is ahead for us andwhat has you excited about 2025
?
Speaker 3 (39:18):
So being the board of directors for SNIA is an
amazing role.
The collaboration is awesomeacross the companies.
Also, being part of themembership committee, with our
new structure we just rolled outand we haven't seen it.
Please check it out.
We now made it easier for youto join, and Michael just
mentioned virtual feather event.
(39:38):
I hosted my first one at the HTCin September and I was just so
happy to see that evencompetitors in a room can talk
about customer outcomes and howwe can benefit, and the concern
about companies saying, oh, Iwould share my confidential IP.
You don't.
(39:59):
You don't have to share any IPwhatsoever.
This is about the customeroutcome.
It's not about we want your IPto publish your IP and you no
longer differentiate in themarket.
No, we talk with them and sayyou can be first to market.
If you have some innovation,bring it up.
We want to work with you onthese things.
What excites me right now is2025 coming up on this one
(40:20):
Working with these new parts ofweather events.
They will drive new standardsacross all these competitors
working as one team.
There's no A B companies.
What's the point of the endcustomer?
How do we get there on thesethings, and I'm looking forward
to the Colorado event coming upnext year, the STC next year
October coming up for even moreinnovation and driving new
starts in the future.
Speaker 1 (40:42):
That's what excites me about the future.
We almost need a podcast justcalled the humans of SNEA,
because that is the reason why Ilove being there and, as you
said, we know how to safelyexchange ideas without breaking
the boundaries of of IP.
You know, we, we're allprofessionals and, in fact,
they're the people that are theclosest to the metal but also
(41:04):
the closest to the customer, andbeing able to bring those
together is such a fantasticthing.
So, yeah, definitely, folkslisten to to these, this amazing
crew.
So, michael George and Curtis,thank you for sharing time with
us today.
And if people want to reach out, of course, connect with all of
us through SNEA events.
There's great virtual events aswell as in-person events.
(41:24):
Head to SNEAorg and see whereall the action is.
There's lots of stuff that'scoming up, and quickly.
So for folks, if they did wantto connect with you directly,
what's the best way, curtis?
If we want to spend more Curtistime, how do we do so?
We want to spend more Curtistime.
Speaker 2 (41:41):
How do we do so?
If they want to connect with medirectly, my email address is
widely published through anumber of industry organizations
and they're free to look me upthrough any of the multitudes of
documents I've contributed tosend me an email.
Of course, there's always acommunication mechanisms like
LinkedIn and other social mediaplatforms.
Speaker 1 (42:04):
Fantastic George.
Speaker 3 (42:06):
Same here.
I've been in this for 20 plusyears, many companies but there
are open websites like LinkedIn.
You mentioned other ones.
Our published names are inthere.
If not, surely they can reachout to every three of you to
reach out to us and, by the way,there are no dumb questions.
Anything you might have, reachout to us.
Speaker 1 (42:22):
Happy to have worked with you, bingo.
I love that.
That's exactly it, michael.
Speaker 4 (42:27):
Yeah, so easiest way is cstichair at sneaorg Very,
very quick.
Speaker 1 (42:37):
And definitely anybody who is at sneaorg is a
friend of mine.
So, folks, thank you forsharing an amazing year together
.
We saw a lot of cool changes in2024, much more coming ahead,
much more great conversationslike this.
So, even if you're seeing thisafter the new hero is already
launched, there's many morecoming.
So head to the website.
The podcast, as well, isavailable for audio subscription
(43:00):
, and we're also on YouTube.
The podcast, as well, isavailable for audio subscription
, and we're also on YouTube, sothese beautiful smiles can be
seen ad infinitum, which is alsothe reason why none of us
really need to give out an emailaddress.
We're everywhere.
But come to come to SDC.
It is truly one of the best,one of the best groups of folks
and a great.
It's a nerd fest in thegreatest possible way, and with
that, thank you very much, team.
(43:22):
Enjoy the experts on data.
Thank you, yeah, this is.
This is a lot of fun.
I love my job, I love all thatwe do, and we'll see you all at
the next snea event thank you.
All right.
Thank you everybody for joiningtoday.
My name is Eric Wright.
I am the host of the SNEAExperts on Data podcast and
co-founder of GTM Delta Superexciting, because we've got
three amazing humans that I'mexcited to chat with.
We're going to dive into sometech to some people and we're
going to talk about the idea offormat.
Now, format is a funny thing.
(00:23):
In fact, this is probably thefirst thing we want to start
with.
Is what do we say when we talkabout format?
Under the auspices of an amazinggroup of technologists of SNEA
coming together buildingstandards, we've also focused a
lot on how do we make sure thatour practice areas align with
what we call data focus areas.
So if you go to the SNEAwebsite, you're going to see a
(00:44):
lot of really, really neat stuffthat we're doing around making
sure that we're bringing downsort of columns and pillars of
where practice areas align, butwe're actually seeing a lot of
cross pillar alignment.
We'll get into that.
More than anything, let's getinto the amazing folks I'm
sharing the call with today.
So if you don't mind doing aquick introduction, let's start
with Curtis and then I'll workour way around.
Speaker 2 (01:06):
Well thanks, Eric.
I'm Curtis Ballard with HewlettPackard Enterprise.
I'm a storage strategytechnologist with Hewlett
Packard Enterprise, working onfuture storage technologies with
the SNEA organization.
I'm a member of the technicalcouncil where I help the SNEA
organization organize, run,oversee their technical efforts.
Speaker 1 (01:32):
You are one of the cat herders.
Welcome to the joy, george.
Speaker 3 (01:39):
Hey guys, my name is George Kambouris and, yes, it's
a Greek last name.
German accent People get soconfused by this Industry
veteran.
For more than 20 years workedfor EMC, dell, hitachi.
Right now.
I'm in that up right now and mymain focus is product
management engineering.
So my job is to define what andwhy and what is really a house
(02:02):
on this one.
I did work with Snea in 2005 onthe NDMP Twig back then long
time ago, and right now I'mworld director at Sneha driving
new standards with the wholeteam.
Speaker 1 (02:15):
Fantastic, and I love the potpourri of both accent
and background.
This is such a we've got a good, diverse representation.
And last but very certainly notleast, michael, if you want to
do a quick intro and then we'regoing to jump in.
Speaker 4 (02:29):
Awesome.
Hi, this is Michael Horde, andI've been in the industry for 30
years working with McDonnellDouglas, Boeing, Sequent
Computer, IBM, Seagate and Intel, and so I'm the chair for the
Cloud Storage TechnologyInitiative, and we have two
(02:51):
subgroups or they're differenttwigs and activities that are
going on.
One is the Cloud Storage Twigthat's working on CDMI, the
cloud data management interface,and then there's another group
of folks that we've just puttogether over the last year and
that's the Cloud Object StorageCompatibility Plugfest, and
(03:15):
we'll talk more about that later.
Speaker 1 (03:18):
Yes, remind me, definitely let's cover Plugfest,
because that's actually such afantastic format.
Let's cover Plugfest becausethat's actually such a fantastic
format, and I've participatedin some myself back in the day
before I was, when I was avendor and contributing on some
of the working groups, so very,very cool.
But, like as as James Liptonsays, we we start, as always,
(03:42):
from the beginning, which shouldbe.
What does it mean when we talkabout format data?
George, do you want to walk usthrough what we're talking about
in this?
Speaker 3 (03:49):
So to remind everybody about SNEA's main
mission, right, snea is, peoplesay, a safe space where
competitors come together tocollaborate in the favor and the
benefit of the end customerright.
We don't want proprietarycompanies which is great for
them to have something uniquefor them, but the customers are
asking for interoperability towork together on common
(04:09):
standards.
So SNEA is there to drive thisinnovation across competitors in
a safe space.
Just a reminder about SNEA inthe overall context.
So if you go, if you haven'tbeen there, you should go there.
On the SNEA org website there'sactually a section we call
Completed Standards and this ismentioned for a bunch of areas
innovation-wise, starting withDDF, the Common Rate Disk Format
(04:33):
, to LTFS right, for example,the Neurotank Pulse System.
We have a bunch of standardswe've done over the years.
The key is how we drivealignment between these
standards develop.
The key is how we drivealignment between these
standards, develop them together, roll them out.
And of course, there'sinnovation happening in the
market with new technologiescoming out over time and we have
to adapt and change with thesethings.
So format is one of the crucialareas when I work in SNEA to
(05:02):
standardize and work together.
So we categorize now under thecomplete standards, a new format
area to drive innovation andbring new science out of the
market.
Let's just mention CDMI as anexample.
We can deep dive into this aswell, but let's start with the
beginning first.
The standards are there todrive innovation across
competitors in the favor of theend customer.
Speaker 1 (05:21):
Well, that is the thing that I so adore about this
.
I've been lucky.
I've worked with a lot of, I'llsay, vendor agnostic
communities and even vendorassociated communities, where we
make a point of separatingbrand from purpose, and it is
such a great way that this comestogether, because we literally
are competitors by day sometimes, but we understand that
(05:46):
together we can innovate if wecollectively define standards
and the entire industry sees thebenefit of it.
And it's such a beautiful setof people that I've been able to
learn from and become friendswith over the years.
This room alone, you know,alone is proof of that.
So maybe let's talk about someof the innovations that have
(06:10):
happened.
Because you talked about stufflike LTFS, people will say, ooh,
I'm looking forward to LTO6.
May not sound exciting, but notrealizing that there is a lot
of innovation even in stuff thatis tried and true, standards
that are evolving, along withphysical, you know, and software
capabilities.
(06:30):
I'll let you guys go.
Maybe, curtis, I'd love to getyour thoughts on.
You know what excites you aboutwhat's being innovated on?
Speaker 2 (06:41):
Yeah, so you mentioned LTFS and people being
excited about LTO6.
I chuckled to myself when yousaid that, eric, people are
currently excited, I hope, aboutLTO10.
Speaker 1 (06:53):
We're getting way ahead.
I'm so far behind.
Speaker 2 (06:56):
Yeah, technology changes so fast you just can't
keep up with it.
So, for those that don'trecognize the acronym, lto is
linear tape open and it's a tapetechnology which a lot of
people think is gone.
I've worked for Hewlett PackardEnterprise for over 30 years
(07:17):
now, and when I joined thecompany, I actually worked for
the team that was directlyconnected to the group that did
tape products and people told medon't ever get put on one of
the tape projects.
Tape is dead.
That was over 30 years ago andjust in the last year I can't
(07:38):
count the number of times I'verun into people that I've told I
work for Hewlett PackardEnterprise and they've said oh,
hewlett Packard Enterprise, youguys have tape products, don't
you?
We've got AI solutions wherewe're creating lots of data and
we're looking for ways to saveit cheaply.
So tape is still very muchalive and is actually getting a
(08:01):
bit of a resurgence with all ofthe AI training and needs to
have cheap ways to archive datathat was used in AI and archive
data that's generated by AI ordata that you might want to use
for AI.
And LTFS is a standard thatSNEA worked on a few years ago.
It's not currently under activedevelopment because it's
(08:23):
successfully being used in a lotof industries today.
Ltfs is the Linear Tape FileSystem.
It's a format that Snykstandardized for writing file
system data out to any streamingdevice, generally a tape device
, any streaming device,generally a tape device in a
(08:44):
defined format that isself-describing and is
standalone in a single tapecartridge so that it can be used
across multiple vendors,creating data on tapes or other
streaming devices that can beused for archival purposes and
long after the vendor that wrotethe data is gone.
(09:09):
The standard developed by SNEAand published as an ISO standard
will have the specificationsand their open source software
developed around it that willallow customers to always read
their data.
I do not currently workdirectly with tape products so,
out of curiosity, as we werepreparing for this call, I went
out to the GitHub repository forthe LTFS software and thought,
(09:34):
hmm, what's going on out here?
It's really stable.
So there haven't been a lot ofrecent changes.
But the most recent change wasin a release published three
days ago when I went out andchecked it, so actively being
used every day, and SNEA's stillgot the expertise and is ready,
(09:58):
if there's ever a reason, toupdate that specification.
Speaker 1 (10:03):
As my favorite phrase is you call it legacy, I call
it production.
Speaker 3 (10:07):
Yes, yes, Sorry.
Speaker 1 (10:09):
George, I stepped on you there.
Speaker 3 (10:11):
I had a discussion with a customer BigBank I won't
say the name for a second, buttalk about how much tape they
actually have, right, and wetalk about exabytes on this
thing.
And he goes we have over azettabyte of tapes.
And I go, wow, right.
And they say, oh, we use theWorld War II silos, the bunkers,
(10:31):
they store these tapes.
The temperature is perfect forthese guys and they have no
intention to change this back toa point because they keep them
around.
But they have to be able to goback and read data from 10, 20
years ago.
Speaker 1 (10:48):
So it is a standard out there which is heavily used
and will continue to be used inthe future.
Right, Well, I'll tell youthat's one thing, that standards
are not just about what we'redoing now, but how we will
cleanly deprecate thesetechnologies and methods as we
go forward.
It's as important to thinkabout the transition between
formats and styles andtechnologies, because I mean,
mean and look, it doesn't evennecessarily require a, an actual
(11:09):
linear tape.
You know we've, there are tonsof products that build in this
virtual you know, front-endlayers so that we can emulate
tape, and that is again anotherway where we're going to see,
you know, know, cloud storagethat appears like tape but
doesn't face the penalties of,you know, archive.
You know, you make a requestand in a couple of days you may
(11:31):
or may not see that request comeback.
It's, there's lots of wayswhere we can use that interface,
but even higher speedtechnologies we may find that
you know some really fast.
We may find that some reallyfast.
Apparently, tapes are beingused in production for stuff
like data training and much more.
Speaker 4 (11:54):
I'd like to add a little bit too.
So, in terms of excitinginnovation, of course, the AI
and with regard to data format,of course, a lot of the new data
structures, and that's with thedirected graph or the knowledge
graph or vector databases.
So there's a lot of work inthat area.
(12:15):
A lot of it's driven out ofdata science itself, but when
you look at CDMI itself, it isstructured as a directed graph,
so it's already set up in termsof providing access, control and
then the set of functions thatyou can essentially build,
(12:38):
operations that you can buildout, the structure for directed
graphs, and that's thecontainers, the data, the
metadata and the operations tomanage all of that.
That's what's in CDMI today, so2.0, that was standardized in
(12:59):
September 2020.
And right now, david Slick isthe chair for the technical
working group for this cloudstorage twig, and so I talked
with him just a couple of daysago, and so the intention is to
bring to the table new datastructures.
(13:22):
So that's data format and thenthe potential for different
semantics with regard tolanguage support, protocol
support.
There's several NFS, smb.
Those are very well established, but there's others that we
need to look at.
So part of the effort is toassess the industry what are the
(13:43):
needs?
And, just like what was sharedearlier I think, george, you
mentioned SNEA is entirelymember-driven.
So everything we do comes frommember requests and when we see
a pattern where gosh, there's alot of folks coalescing, and
(14:05):
then we have that discussion asan industry group to then get
consensus and then start drivingcollaboration.
Snea is a third-party vendorneutral.
We put together forums,roadmaps and then the technical
initiatives and then finallystandards efforts where that's
(14:26):
appropriate.
So that's exciting.
The other aspect, just briefly,is on the data formats with
regard to multi-protocolgateways.
So when you're looking atactual on-prem to cloud and
multi-cloud, at actual on-premto cloud and multi-cloud, all of
this is so dynamic in terms ofshifting data.
(14:49):
So the data format integrity isto have a unified management,
but also the capability to haveconsistency and integrity across
all the multi-protocols.
And, of course, the existingcloud has a number of very large
entities, but those are alldifferent.
(15:12):
There's a lot of protocoltranslation that happens under
the covers and that's preciselywhat we're working on for the
Plugfest.
So I just wanted to give thatlittle point on that.
Speaker 1 (15:23):
Go ahead.
So I just wanted to give thatlittle point on that part Go
ahead, yeah, the plug fest andthis, the human cross.
Collaboration is what's amazing, and in fact I've seen a lot of
vendors who made the choice tonot run to the market with a
product but run to the industrywith a standard and then,
through developing thosestandards, then develop a
(15:45):
product that has, you know, away to interact by that standard
, because that ensures that theyhave partners and innovation is
shared.
Because you know, they alwayssay we, we don't want to
reinvent the wheel.
Well, we've actually reinventedthe wheel quite a bit, because
previous wheels used to used tonot be good like, or they were
fine at the time.
Speaker 4 (16:03):
but wheels have evolved, not necessarily at an
incredible pace, but it needs tohappen, or they become a
reference architecture and thenthe new foundation is built
thereon.
Speaker 1 (16:15):
Right, yeah, we forget about how much happens
that way.
Speaker 3 (16:21):
You just mentioned how people innovate and the
environment is changing.
Let me go back with Michael totalk about CDMI.
Yes, september was 2020 lastupdate, but the version 1.1 was
in 2011.
So why was there a need let'stouch this for a second for
changes?
Here's one example 18 monthsago, customers were saying how
(16:43):
do I prevent hackers coming in?
That was 18 months ago.
The last six months, all theCIOs, ceos in our customer
briefings are now saying what doI do after I get hacked?
How can I restore?
So the cyber attacks happeningleft and right are rapidly
increasing.
So the market shifted now fromhow do I stop hackers to how do
(17:07):
I react?
How can I recover fast?
Ransomware attacks is anexample right, huge out there
right now.
So the market has shifted.
As a result, we have to adjustthe standards as well.
So we went beyond than justsaying how do I store data?
How do I access protocols todata?
But how do I secure it?
How do I do key management?
How do I access acrossprotocols to ensure that
(17:29):
integrity?
Who has authentication,authorization, different methods
, right.
So the CDMI style has evolvedbig time over the last, I would
say, five years to adoptinnovation and security change
over time, security is becominga key point going forward.
Speaker 1 (17:46):
And security is probably one of the greatest
reasons we always say that.
You know, when DevOps became athing that dripped off the
tongues of all these you knowexcited pundits and CIOs, and
then along the way, about a yearand a half later, suddenly
comes DevSecOps and someone said, whoa, why do you have to like
explicitly put sec in the middle?
(18:06):
It was implied.
I'm like, well, it may havebeen implied, but they weren't
actually there.
So, yeah, we've had to go outof our way.
But in the SNEO, in the TWIGs,so the technical working groups
and a lot of the other parts ofthe community, the fact that
you're in the room with peoplethat are security experts and
they have software developers,so all of these people that in
(18:28):
production environmentscollaborate, but they never
collaborate in the planning room.
We're doing that now.
It's actually the successes insoftware optimization for things
like what we do with LTFS andCDMI.
Those software innovationshappen because they're sitting
in the room, with peopleunderstanding where the
standards came from and howother people are using it as
(18:50):
well, which that you know,george, you we've talked before
about this like the use case iswhat matters.
I'm a nerd, I love the bits, Ilove the speeds and feeds.
But in the end I really lovewhen I see like medical benefits
and human benefits andscientific benefits that we get
from doing stuff with thesetechnologies and these standards
(19:13):
.
Speaker 3 (19:13):
You just mentioned personas, right?
Let's touch on this for asecond.
We used to have the chiefsecurity officer, but now they
have a new title of chief dataofficer, for example, and these
guys are talking to senior aswell, saying look, my challenges
are 58 applications, I have toshare it across them.
How do I do it securely and nothaving to reinvent the wheel?
(19:33):
Back to your previous commentright, these guys are talking to
us about standards and CDMI.
One example is awesome forthese guys.
Let's talk about one use case.
Let's take healthcare as anexample.
Right, biggest challenge If youare not up to date and you take
your cell phone, you probablyhave a healthcare app on your
cell phone Most likely I have onthat one and I can talk to my
(19:54):
doctor in a secure chat and theycan make prescriptions in the
app online with my CVS pharmacy.
Whatever All this, automationhas to be standardized,
regulated, secure, gdprcompliant, hipaa compliant.
Automation has to bestandardized, regulated, secure,
gdpr compliant, tippercompliant.
The list goes on and on right.
To get to this point, it tookyears of innovation and cdmi is
(20:15):
the back end of this foundation.
So it's just one example of usecase we can say a start is
driving innovation to theconsumer, in this case, to make
our lives much, much easier.
I love this app.
So easy before call the doctor,make an appointment, go inside,
inside, get the prescription,try to see me.
Oh my goodness, right Now it'sall online.
My kids now are expecting thisto be a standard going forward.
(20:36):
Ten years ago it didn't exist.
So major innovation happening.
Speaker 1 (20:41):
Yeah, I would spend five hours to avoid a 15-minute
doctor's appointment any day andthe fact that work has happened
in the industry using standardsyou know, we always say it's
the standards well-defined meansthat the artifacts can change.
It's the interplay betweenthose artifacts that's the most
important thing, and it's noteven at the moment in time, but
(21:04):
the thing that we thought wasnot useful.
Like archive data, people arelike, ah great, what are we
going to do with this?
We store it for the seven years, which is actually not even a
real thing.
The seven years is the same asdon't yell fire in a movie
theater.
It's not actually written down,but we use it as a standard,
supposedly.
But we took all this data whichwe just said we'll put it over
in the corner just in case weneed it.
(21:25):
Well then AI came along and weneeded to train on company and
customer data safely.
Well, where do you think thatdata has been for a long time,
right?
So all of a sudden, the archivebecame primary storage and it
flipped on its end because theuse case changed.
Speaker 2 (21:45):
Yeah, and the topic we're talking about today became
way more important the formats,you know, the formats that
organizations like SNEA helpdefine.
Snea is not the only one, butSNEA focuses on formats of data
in storage ecosystems, and it'scritical that we have these
(22:25):
well-defined standards for howyour information is formatted.
Information exchange thatenables the technology that so
many people count on today, andSNE is here to help define those
formats.
And one of the things we'retrying to do today is look at
what's the next gap gap.
(22:46):
Where is the next format thatwe need to be defining to enable
that next important informationexchange opportunity?
Speaker 3 (22:55):
I have a comment, if I may finish the previous
thought on this one, which was agood example.
So back to my healthcareexample.
Let's first a little bit showyou how these things work
together.
Back to format data from Ericin the beginning.
Right, I just touched a bit onCDMI.
However, if you swipe your IDat the doctor, they have a
(23:18):
system in place with CDMI to goback to the tape device before
and pull up your x-rays.
So now we have CDMI workingwith LTFS together.
Number one this is someone'ssort of a system probably a NAS
device or something else storagesystem using DDF formats.
So now we have three formats inone healthcare use case working
(23:39):
together from SNEA.
So it's not just one example ofa standard.
These work all hand-in-handtogether.
Just to add as well,interoperability between systems
is very important If work allhand in hand together.
Just to have this as well,interoperability between is very
important.
If you haven't seen or haven'tbeen to our conference, by the
way, the State of Web Conferencewas in September in Santa Clara
and we had, for example, davidSilk give an awesome
presentation on CDMI withexample use cases.
(24:01):
If you haven't been there, nextone is coming up, one in
Colorado early next year andthen of October next year in
Santa Clara Again.
Please come there for that one.
But you just mentioned to usfutures.
Just to give you a quickglimpse of another example of
formats, I'm driving now a newinitiative within SNIA as well,
how we can start across NFS, smband S3.
(24:22):
Just an example it doesn'texist yet.
It's a brand new program.
It might not complete anytimesoon, but customers are asking
us.
I have compliance issues, Ihave auditing issues, I have
share of custody issues.
I'm working nowcross-functional SNEA across all
our competitors together tocome to the end customer view
and say how can we make theirlives easier to have one start
(24:46):
across these things?
It might take us six months,two years, it doesn't matter.
The point is the customers havea challenge.
We recognize the SME as a teamand we work on new stars going
forward, so we don't stop today.
We have new stuff coming allthe time.
Speaker 4 (24:59):
Right.
And then going back to the dataformat, that it's not just
simply the data itself, but it'sthe metadata that really should
be stored with the data, andthose should be very closely
coupled and go together asthings progress, wherever that
data is migrated or used acrossthe multi-cloud.
(25:23):
And that's actually what CDMIdoes.
And so when you look at whatGeorge was talking about the
healthcare example, what they'redoing as far as data science or
the data chief data officer, etcetera, they're looking at the
overall structure of all dataand that's the ontology, and so
(25:46):
it's literally at the highestlevel.
And then, in order to make thatsearchable, then they're using
knowledge graphs and algorithmsto then get to more precise or
exact results.
Instead of like with a vectordatabase, it's more of a
similarity approach algorithmand that is not exact.
(26:08):
So that's kind of the bleedingedge of using vast amounts of
data and with the tape that wasdescribed with zettabytes.
But it's really, how do youpackage that into a way that's
consumable and useful in ourlifetime instead of calculating
(26:29):
over decades?
It needs to be responsive, andthat's where the technology is
at right now, today, and I'mexcited about Sneha's play in
this.
Speaker 3 (26:40):
I love your example about AI Michael.
Do you want to touch base onsecurity for a second, because
we have seen the news?
Oh, chep CPT index, blah, blahblah without my authorization.
Can you touch on this onesecurity for a second?
Because we have seen the news?
Oh, chat, gpt index, blah blah.
Without my authorization.
Can we touch on this one, too,for a second?
Speaker 1 (26:53):
Yeah.
Speaker 3 (26:55):
So I'll start off and then, michael, maybe we could
complement this one.
You have seen in the news aboutall these ILM models out there
learning on data sets and youcan see all these actors.
You and companies now said youuse my data without my
permission.
You just touched base with this, michael, about security.
Right, the databases Want toelaborate maybe on the security
(27:17):
aspect of this one?
Speaker 4 (27:17):
Yeah, should we ask them this one?
Yeah, so with the and I'll justsay the example is CDMI.
It has all of the accesscontrols.
It has delegated access controlso you can use the proprietary
or the system of your choiceon-prem.
And so there's all the portsand exports in and out of going
(27:42):
down to ACES, which is accesscontrol entries.
So it's very fine, tuned downto the individual object level
data, they're able to look atthose aspects that George just
talked about.
As far as the heritage orlineage of that data, where did
(28:06):
it come from?
What are the protection methods?
Need to know who should beinvolved.
So all of that is somethingthat exists today.
So the next step is reallyapplying that into real systems.
Speaker 1 (28:21):
So go ahead, and this is what you really highlighted.
This thing of that definitionof the sureness of the format
means that everything else we docan always rely on that
consistency across platforms,across software, across hardware
platforms, across software,across hardware.
(28:44):
Even RAID is well.
Again, not a lot of innovationshappening in RAID, but yet the
hardware and softwareinnovations that allow us to do
more with understanding how tooptimize and extend the life of
hardware and protect anothergroup that we have as a
data-focused area.
I had a little mini server backI'm an older fellow so I've been
I've had used to have thoselike tiny servers started in
distributed computing and we hadextra ones that were left over
(29:08):
because they've beendecommissioned and I thought, oh
, you know I'm a nerd, I want todig in and do some new nt
windows, nt goodness.
So I found an old server andlike, oh, we had a stack of the
old drives and so I took fivedrives and I put them into the
tower and I powered on theserver to get ready to put in
the cd and it came up to theoperating system screen and I
thought, oh, and I logged inwith my regular production id
(29:32):
and it was like wait a second,because, just by sheer
coincidence, I picked up fivedrives that were originally from
it, stuck them back in and itbecame the server.
It was that six months earlierhad been unplugged and well,
security problem because wedidn't really think about how to
clean that data.
The fact that five disparatedrives had enough information to
(29:53):
be able to tell the controllerto rebuild the operating system.
In this it went back in placeand again it's like what do we
store with metadata?
Where do we store index indices?
Where do we do all these things?
Those common standards, then,mean that if we want to do
things like emulation andinnovation, it has to be done on
(30:14):
the basis of a foundation.
Cinder blocks are always thesame shape for a bloody reason,
so that we can build houses ontop of them.
Speaker 3 (30:22):
Exactly.
Speaker 1 (30:23):
That's boring.
Speaker 3 (30:25):
Yeah, that's right, you mentioned a lot of
innovation happening.
Let's touch on this for asecond because, as we might have
seen the press now Kiyokya, wd,samsung, you name it right they
just announced 120 terabytemedia on Flash 256 come running.
(30:46):
After this, we have seen thefirst pre-announcement of one
petabyte drives Media drivescoming right.
So, while RAID's not evolving alot, let's discuss, because
rebuild times is one of the mainthings people are worried about
.
If you have too long rebuildtimes, you're exposed.
Yes, there's erasure coding.
Yes, there's rate DP and tripletech.
(31:07):
These are all out there, but wehave to look at the market and
always evolve and adjust.
So you might see moreinnovation coming in the rate
group division about ratestandards going on as these
capacity evolve, so there mightbe another version coming out.
We never stop, we alwaysinnovate.
Just to make the point here,it's not dead and done.
We're done with it.
We will innovate going forwardbecause the technology evolves
(31:31):
over time.
It doesn't stop.
Speaker 1 (31:34):
It's.
I'll say SNEA is not unlike themoon mission that people think
it was a trip to the moon.
But what they don't realize isit was the beginning of decades
of innovation because ofresearch done on that trip to
the moon.
That's, that's why thesecommunities have to exist, and
even we talked about differentstandards bodies, iso, ieee.
We have cross, you know,organization collaboration.
(31:58):
That's going on.
We've had folks on the podcastalready that are doing work
within and across the standardsorgs, and this can be done
because built on standards meansinnovation happens faster.
I just want to hammer that likeI should put that on my
forehead.
Standards drive innovation.
It may seem boring to slow downand define it right at the
(32:21):
start, but boy oh boy does theflywheel kick in because of the
work that's being done by thisgroup.
Speaker 3 (32:29):
That reminds me, eric , a great suggestion to remind
our audience here.
We do want more members.
Okay, there are a lot ofcompanies out there, especially
on AI, for example, startcompanies popping up left and
right.
We live and die by our members.
As Michael mentioned before, wehave a great set of companies
(32:50):
working together already, but wewant more.
We need innovation.
We need people coming in todrive standards.
If you haven't been looking atSNEA yet, please look at our
website.
Look at our membership.
We just changed our structureMuch more easy to join SNEA now
and participate in certain workgroups at every level of this
one.
If you have suggestions on newstandards, don't think you're on
(33:12):
your own.
Come to us.
We're happy to work with you.
Bring you in.
We are open for new ideasuggestions.
So don't think we're closed off.
We want more members.
Just to point this out to theaudience here.
Speaker 1 (33:25):
The beauty of how we can safely share how we're doing
things that still allow us to,in our individual companies,
continue to work as the way wedo but the fact that you can
probably save a ton of time Justbeing in that room and seeing,
like you said, the talks thathappened at SDC and they're on
(33:45):
the YouTube channel as well.
There's quite a few that arethere, so folks should go check
those out.
You get to interact withengineers who are building these
tools and seeing real customerexperiences.
There's very rare opportunitiesto get in front of that kind of
community.
So, oh boy, I could spend allday with you guys.
(34:06):
But I definitely want to say,as a wrap for what we, what you
see ahead and you know,specifically around format, but,
sneha, in general, with 2025,curtis, I'll let you lead us out
what's got you fired up about,what's new and coming.
Speaker 2 (34:22):
What's got you fired up about what's new and coming?
Oh well, we've mentioned itseveral times already, but AI
(34:45):
has just introduced so manychallenges.
With every AI tool that theyuse, stores, provenance
information in a differentformat, and there's
opportunities for identifyinghow we can take some of these
tools that the world is buildingon for AI and start creating a
little bit more structure aroundhow those things can work
(35:09):
better together.
Especially, you know, there'sso many things throughout the
storage infrastructure behind AIthat, as everybody is focused
on the GPUs and the compute andthe processing, there's this
entire storage layer thatunderlies that.
You've got to get that datainto your AI processing and
(35:31):
there is so much opportunity forbuilding out that storage layer
for accelerating your AI worldaccelerating your AI world.
Speaker 1 (35:47):
And, michael, what's got you excited to wake up on
Christmas morning knowing thatyou've got a year ahead and
presents?
Speaker 4 (35:51):
to open.
Well, I'd have to say thatevery day is pretty much like
that, opening a brand newpresence every day through the
year, and the reason I'm sayingthat is this last year we had a
very remarkable activity withthe Plugfest.
There was a whole community puttogether and many of the
(36:12):
contributors are not traditionalSNEA members.
So we're reaching out intoareas for cloud storage.
So we're reaching out intoareas for cloud storage that and
(36:33):
it's not just the storage, butit's the gateways and other
infrastructure that goes on-premand various software.
So, and to get very specific,with the Plugfest, we conducted
that in September.
We had a November 21st webinar.
So I'd love to like, in yourcomments, we can publish that
link for that.
Speaker 1 (36:47):
Yeah.
Speaker 4 (36:48):
That was a report to the industry on the activity
there and that gives a lot ofthe detailsFest and we're going
(37:08):
through the review processwithin SNEA to create a new
organization for that, and sothat was just recent.
We've got our PlugFest callscoming up in January as well as
the open source test suite, sowe have separate calls for that.
And then just yesterday I had awonderful conversation with one
(37:32):
of the candidates we're lookingat for leading the open source
test suite development, and whyI point that out is what excites
me is to have thoseconversations where these are
time commitments and it's noteasy.
People are extremely busy, butwhen you have something vitally
(37:56):
important that is exciting andthen having that discussion
about okay, where is theindustry going, how can we
influence that and provide asignificant impact.
That's the kind of discussionjust individually we have, but
also with our forum we had abirds of a feather session
(38:19):
looking at strategies and thenthe final thing is I'm doing
some outreach to open sourcetest people.
So it's a very long list of newpeople.
So my goal personally is todouble the size of our test
matrix by April for the nextPlugfest.
Speaker 1 (38:39):
Fantastic, and that's where real testing happens.
I always say there's no greaterlie than things like I have
read and accept the terms andconditions and all the tests are
in the PR.
Thank you, yeah, nice,excellent, well, definitely,
we've got a lot ahead and I wantto make sure that we bring it
(39:02):
back.
So, george, my German and mybeautiful cornucopia background
friend, what is ahead for us andwhat has you excited about 2025
?
Speaker 3 (39:18):
So being the board of directors for SNIA is an
amazing role.
The collaboration is awesomeacross the companies.
Also, being part of themembership committee, with our
new structure we just rolled outand we haven't seen it.
Please check it out.
We now made it easier for youto join, and Michael just
mentioned virtual feather event.
(39:38):
I hosted my first one at the HTCin September and I was just so
happy to see that evencompetitors in a room can talk
about customer outcomes and howwe can benefit, and the concern
about companies saying, oh, Iwould share my confidential IP.
You don't.
(39:59):
You don't have to share any IPwhatsoever.
This is about the customeroutcome.
It's not about we want your IPto publish your IP and you no
longer differentiate in themarket.
No, we talk with them and sayyou can be first to market.
If you have some innovation,bring it up.
We want to work with you onthese things.
What excites me right now is2025 coming up on this one
(40:20):
Working with these new parts ofweather events.
They will drive new standardsacross all these competitors
working as one team.
There's no A B companies.
What's the point of the endcustomer?
How do we get there on thesethings, and I'm looking forward
to the Colorado event coming upnext year, the STC next year
October coming up for even moreinnovation and driving new
starts in the future.
Speaker 1 (40:42):
That's what excites me about the future.
We almost need a podcast justcalled the humans of SNEA,
because that is the reason why Ilove being there and, as you
said, we know how to safelyexchange ideas without breaking
the boundaries of of IP.
You know, we, we're allprofessionals and, in fact,
they're the people that are theclosest to the metal but also
(41:04):
the closest to the customer, andbeing able to bring those
together is such a fantasticthing.
So, yeah, definitely, folkslisten to to these, this amazing
crew.
So, michael George and Curtis,thank you for sharing time with
us today.
And if people want to reach out, of course, connect with all of
us through SNEA events.
There's great virtual events aswell as in-person events.
(41:24):
Head to SNEAorg and see whereall the action is.
There's lots of stuff that'scoming up, and quickly.
So for folks, if they did wantto connect with you directly,
what's the best way, curtis?
If we want to spend more Curtistime, how do we do so?
We want to spend more Curtistime.
Speaker 2 (41:41):
How do we do so?
If they want to connect with medirectly, my email address is
widely published through anumber of industry organizations
and they're free to look me upthrough any of the multitudes of
documents I've contributed tosend me an email.
Of course, there's always acommunication mechanisms like
LinkedIn and other social mediaplatforms.
Speaker 1 (42:04):
Fantastic George.
Speaker 3 (42:06):
Same here.
I've been in this for 20 plusyears, many companies but there
are open websites like LinkedIn.
You mentioned other ones.
Our published names are inthere.
If not, surely they can reachout to every three of you to
reach out to us and, by the way,there are no dumb questions.
Anything you might have, reachout to us.
Speaker 1 (42:22):
Happy to have worked with you, bingo.
I love that.
That's exactly it, michael.
Speaker 4 (42:27):
Yeah, so easiest way is cstichair at sneaorg Very,
very quick.
Speaker 1 (42:37):
And definitely anybody who is at sneaorg is a
friend of mine.
So, folks, thank you forsharing an amazing year together
.
We saw a lot of cool changes in2024, much more coming ahead,
much more great conversationslike this.
So, even if you're seeing thisafter the new hero is already
launched, there's many morecoming.
So head to the website.
The podcast, as well, isavailable for audio subscription
(43:00):
, and we're also on YouTube.
The podcast, as well, isavailable for audio subscription
, and we're also on YouTube, sothese beautiful smiles can be
seen ad infinitum, which is alsothe reason why none of us
really need to give out an emailaddress.
We're everywhere.
But come to come to SDC.
It is truly one of the best,one of the best groups of folks
and a great.
It's a nerd fest in thegreatest possible way, and with
that, thank you very much, team.
(43:22):
Enjoy the experts on data.
Thank you, yeah, this is.
This is a lot of fun.
I love my job, I love all thatwe do, and we'll see you all at
the next snea event thank you.
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
Ridiculous History
History is beautiful, brutal and, often, ridiculous. Join Ben Bowlin and Noel Brown as they dive into some of the weirdest stories from across the span of human civilization in Ridiculous History, a podcast by iHeartRadio.