The State of the Network Engineering Union, with William Collins & Scott Robohn

Send us a text

What if the future of network engineering isn't just about cables and hardware, but about mastering the cloud and cybersecurity? Join me, Andy Lapteff, as I sit down with industry giants William Collins and Scott Robohn to uncover the secrets of thriving in this ever-evolving field. William, now a cloud automation expert at Alkira, takes us through his journey from systems administrator to a key player in multi-cloud network software, and shares insights from his own podcast, the Cloud Gambit. Meanwhile, Scott shares wisdom from his 35-year career in networking, offering a unique perspective on how to navigate the industry's seismic shifts with optimism and foresight.

As the realm of network engineering transforms, we examine the commoditization of networking as a utility and what it means for professionals in the field. Amidst the pressures of standardization and falling prices, the conversation turns to emerging technologies like 5G, 6G, and network security that promise new opportunities. We emphasize the timeless importance of core skills such as TCP/IP, even as the industry leans more towards software. We also delve into the 'cowboy' days of the 90s and 2000s, contrasting those times with today's structured and strategic approach, providing invaluable insights for both newcomers and veterans in the field.

Technology never stands still, and neither should your skill set. We explore how to adapt and thrive amidst technological change, discussing how automation can enhance job satisfaction and ease the learning curve. We expand on the importance of foundational skills and certifications, like Cisco's CCNA, remaining relevant even as cloud and cybersecurity take center stage. Tune in to uncover how network engineering can be your stepping stone into lucrative emerging fields, learn how to leverage Python and Terraform, and understand the intricate dynamics behind cloud adoption. Let us guide you through the maze of AI in network engineering, as we dissect the hype and highlight the real innovations that could reshape our professional landscapes.

Find everything AONE right here: https://linktr.ee/artofneteng

Episode Transcript

Speaker 1 (00:00):
This is the Art of Network Engineering podcast.
In this podcast, we exploretools, technologies and talented
people.
We aim to bring you informationthat will expand your skill
sets and toolbox and share thestories of fellow network

(00:21):
engineers.

Speaker 2 (00:23):
Welcome to the Art of Network Engineering podcast.
My name is Andy Laptev, I am atAndy Laptev on the Twitters and
tonight I am joined by two.
What are you?
Are you behemoths of industry?
Are you what's the term Notbehemoths?

Speaker 3 (00:38):
Titans, titans, titans of industry.
Let's start over.

Speaker 2 (00:42):
Tonight, we are joined by two titans of industry
William Collins and Scott Robon.
Did I pronounce that right?
I should know your last nameGot it.
It's like putting a robe on YepJust keep it on.

Speaker 4 (00:53):
Awesome.

Speaker 2 (00:55):
So a special shout out thank you to William for
jumping on last second.
A couple of folks are sick onthe Art Network Engineering and
couldn't make it, so Williamjumped in.
Spoiler William has anotherrecording with us coming up soon
, all about, I think, cloud.
So, william, we get to hang outa couple times, which is great,
and then I'm going to see bothof you next week at a USNUA

(01:16):
event, so we are like new BFFs.
I'm really loving all the timewe're spending together.

Speaker 3 (01:21):
Let's do some quick intros here.

Speaker 2 (01:23):
So, william, for folks living under a rock, who
are you?
Where do you work?
What do you do?

Speaker 3 (01:28):
Yeah, William Collins , so I've been doing.
I started out like system admin, data center, grinding my gears
, load balancing, CSS, all thosethings back in the day, and
then when cloud adoption hit, Ijumped pretty hard into cloud.
So I've been doing this stufffor probably around 20 years,

(01:48):
really focused heavily on cloudautomation, AI and security
stuff.
And I work for a startup.
So I work for Alkira.
We build multi-cloud networksoftware.
We're kind of like a networkplatform.
So I did work in enterprise fora very long time, which is
where I really grinded and gotto know technology.
And then startup life has been,I think, going on three and a

(02:10):
half four years now.

Speaker 2 (02:12):
So so your career trajectory is very much in line
with what we're going to talkabout today.
It sounds like you did all thethings right and I am learning
from you every day.
I continue to.
What's the name of the startupyou work at?

Speaker 3 (02:23):
Alkira.
Yeah, Alkira.
So the Khan brothers, Amer andAtif Khan, formerly Viptela,
which is now Cisco SD-WAN.
So that was their startup andthey started Alkira following
that.

Speaker 4 (02:37):
I like to point out that they're former Juniper
before they were former Alkira.
You know what's funny aboutthat.

Speaker 3 (02:43):
I went to an event and this is a funny story.
I go to an event.
Atif walks in and like therewas like a convergence, this
massive convergence of people ina giant circle and they were
all ex-Juniper just in thisgigantic circle, in this event.
It was absolutely hilarious.

Speaker 2 (02:58):
That's what they do.
They're kind of like attractedto each other.
I see over your right shouldera very cool logo sign.
So what podcast did you create?
And host.

Speaker 3 (03:11):
I always forget to promote this thing.
So the Cloud Gambit podcast I'mthe host.
Recently I have a co-host, soYvonne Sharp joined me.
So we are having all sorts ofconversations, everything from
like venture capital, howstartups get funded, to the
technology itself.
I talk to a lot of founders,startup folks like that and just

(03:32):
interesting conversations thatare kind of like network
engineering and cloud adjacentas well, Because it's good to
know about the people around you.
It pays dividends later on.

Speaker 2 (03:42):
I was joking with William before the show that if
I knew Yvonne was in the marketI would have grabbed her.
But you did, and that's a greatget.
I love, I love Yvonne and I'mreally enjoying the show.
The addition of Yvonne hasdefinitely not that the show
wasn't incredible before, but Ireally love your, your
conversations.

Speaker 3 (03:58):
Yeah, she's the cloud psychologist for sure.
Or philosopher cloudphilosopher.
I got to get it right.
She is the cloud philosopher,For sure.
Or philosopher cloudphilosopher.

Speaker 2 (04:05):
I got to get it right she is the cloud philosopher,
cloud whisperer, awesome.
Well, thanks, william.
Great intro, great stuff.
Scott, who are you?
What do you do?
Where do you work?

Speaker 4 (04:11):
Hey, thank you for having me on.
The short version is I'm anindependent consultant in
networking no-transcript butI've got almost 35 years in the

(04:46):
networking business.
I'm one of these very fortunateindividuals that I got in just
before the commercial internetstarted really getting traction
and there was alwaysopportunities to learn, learn,
learn new stuff.
And, andy, when you wanted totalk about this tonight, yeah,
let's talk about this.
I think there's way more hopeand way more positives that we

(05:08):
can talk about here, even whenwe're dealing with being
disrupted personally and in anindustry.

Speaker 2 (05:13):
Yeah, perfect segue.
So let's jump into the topic,the roundtable.
Whatever we're doing here, youknow, scott, one of the reasons
I brought you two in is you guystravel all over the place.
You're always, it seems likeevery time I see you, you're
somewhere else and you're, youknow, with networking or with
cloud folks, or so you guys getaround a lot more than I do to

(05:35):
all these different events andScott's experience.
You know, we talked abouthaving you on the show, scott,
and we started talking aboutwhat are we going to talk about?
And I know that you go aroundtalking about network operations
a lot with teen ops and it's,it's a great show, it's a great
conversation.
Listen, I'm a network operator,so like it's, it's near and
dear to my heart and I love yourvision and your goal of like.
How can we and I don't want tobastardize the, the, the mission

(05:57):
statement, but my understandingis you're trying to like,
codify, you know what are bestpractices that we could follow
in network operations, becauseit really seems like every place
is kind of doing their ownthing and kind of doing some of
this, some of that and, like youknow, even just.
I mean, I know that we havebodies of standards and you know
, but it really just seems likeeverything is a snowflake

(06:21):
everywhere.
So I really like theconversations you've been having
, but again, having you on, I'mlike, well, I don't just want to
have the teen ops conversation,what are we going to talk about
?
And I started thinking like so,like because of our show and
because of, I guess, what I do,like I have a ton of people I'm
working with, like fourdifferent people right now that
approach me like hey, I want toget into network engineering.
You know, I got out of themilitary or I'm in college or

(06:41):
I'm studying cyber, like I wantto get into network engineering.
And I found your show, I foundyour.
Whatever, what's your advice?
And this happens to me all thetime, and my advice five, seven
years ago I'm not sure is thebest advice to give today.
Because it seems to me and Idon't know if this is my bias,
which is why I want to talk toyou guys but it seems like

(07:03):
there's a lot changing in IT, inthe networking space, right,
like cloud came in and disruptedeverything.
I think before that, you know,automation was a big disruptor.
And when, for me, when thesenew quote unquote new things
came in to my life.
I was drowning in just keepingthe lights on at places.

(07:26):
So when they said, well, nowyou have to learn automation,
you got to be kidding me.
I'm doing four maintenancewindows a week.
I'm exhausted.
This has been going on foryears.
When am I supposed to do that?
And then it was simultaneouslycloud.
We merged with another company.
We were in one cloud, they werein four.
Now we're a multi-cloud.
Now you have to be cloud-versed.
So again, this is probably justmy bias, but I was introduced to

(07:48):
things like automation andcloud in production environments
of just.
There wasn't enough time tolearn this stuff.
So when I got out of production, when a vendor grabbed me and
said, hey, let's come over hereand work on making customer
experience better, I'm likegreat, I am not studying anymore
.
I am sick of this shit, becauseI was burned out, honestly and

(08:08):
I just stopped studying.
So I'm going to shut up soon.
But where I'm going with allthis is I stopped studying.
I went to a vendor.
I felt great, now I'm justgoing to smile and be charming
and be the voice of the customerand make things better for this
company's customers.
And then all these rounds oflayoffs start happening in tech
and I get hit by one of them andmonths go by and I'm like, how
the hell am I not getting a job?

(08:30):
I'm like a smart person.
I've worked on some of theworld's biggest networks.
I'm a publicly known person.
I think I have a goodreputation, I have integrity and
I can't freaking land anything.
What is going on?
So bring it back to our topic.
My experience has been there'sbeen a lot of transition, a lot
of things going on.
People come to me and ask mewhat should I do to get into

(08:52):
networking or what should I doto stay skilled up, and I guess
I just like to start maybe withyou, scott.
You've been around a long time,you've seen a lot of this.
You said something before theshow of like this has happened
before, so do we want to startthere?
Like, is what's happening now?
A is my perception that there'sa lot of change happening right

(09:13):
now in IT and in networkengineering, specifically
happening Like that's not justmy cognitive bias.
Right, like the commoditizationof networking.
Now it's a utility, nobodycares, let's automate everything
away.
We haven't even gotten to AIyet, so we'll get there.
Right, but is that happening?
Is this a thing that'shappening right now from your
perspective.

Speaker 4 (09:28):
Yeah, you're not wrong.
Right, You're not just feelingsomething.
You're being impacted bychanges in the industry and
without doing a whole historylesson that I'm probably not
really qualified to give beyondthat 30-year window, look, we
can look at all sorts ofexamples in manufacturing in
particular from the industrialrevolution, of how things have

(09:49):
changed that have beenlabor-saving mechanisms but
definitely have impactedpeople's lives and disrupted
them, and there's good and badin all of that.
Right, I'm not want to, I'm nottrying to put a pretty face on
oh, it's all just great andpeople are never impacted.
Right, you gotta be honestabout that In our in our day and

(10:12):
age.
Right, I mean, let's go back toprogrammers.
Right, there were punch cardsand there was programming in
assembler and other much lowerlevel languages.
That's not how you program mostmachines today.

Speaker 2 (10:27):
Yeah.

Speaker 4 (10:27):
So there's been change, there's been disruption
and evolution there, right, andwe could come up with a you know
many different examples, Ithink, for the conversation.
Let's just assume that there'salways going to be change of
some sort.
It could happen more quickly orless quickly.
I think, andy, what you'rereacting to realistically is an
acceleration of some of thatchange and a couple of things

(10:49):
that have piled up on you andothers all at once.
I'll breathe and see if Williamhas something to add there too.

Speaker 3 (10:56):
Yeah, I mean those are all good points.
You said something about yousaid, commoditization.
You know networking as autility.
I think that's an interesting.
So one thing I try to do is Itry to take alternate viewpoints
, sometimes just to tease outthe conversation and maybe even
challenge folks sometimes,because I actually just had a
long conversation about thiswith someone like a week ago,

(11:17):
and it's funny like thisindustry is evolving in like so
many ways and there iscommoditization.
But like, what are?
Like, if you think aboutcommoditization, what are
actually the signs so you couldtake um, uh, like, maybe
standardization, so networkingprotocols or you know a lot has
been standardized to make, youknow, make it easier for vendors
to, you know, offer competingproducts and such right.

(11:40):
You know there's um downwardprice pressures.
You know that's another sign,maybe like, like you know you
have competition, so you havethis like downward pressure on
prices, you know, for manynetworking and telecom services.
But then you have, you know,like you said, cloud sort of
came along and it acceleratedthe aspects that can be
commoditized just like so fast,like so all at once.

(12:03):
Commoditized just like so fast,like so all at once.
So all of these things are likekey indicators of you know some
broader picture ofcommoditization, but you also
and this is where I think it's abeautiful thing so like looking
at the, you know the other endof the coin.
There's a lot of factors thatlike counter this.
So one area that I am just noteven well versed in at all, I
don't, I actually don't I wantto stay away from this area, but

(12:26):
if you look at like 5G and 6G,like those types of technologies
, they are very innovative,they're very much differentiated
.
So that's like an alternateindicator.
And then you have like networksecurity.
You know threats evolve, thethreat landscape, your attack
surface, there's endpoints,everywhere everything can be

(12:48):
attacked.
And now you have chat GPT, soit's accelerated like 100%,
probably a lot more actually Idon't even want to know what
that number is.
But there's this ongoing demandfor more advanced security
solutions and more advancedsecurity practitioners.
And it's funny because a lot ofthe security folks that I've
rubbed shoulders with over theyears, a lot of them aren't

(13:09):
technical, like on the networkengineering side.
So you need skills like networkengineering, like, are you
going to go ask somebody thatdoesn't understand network
engineering to look at a packetcapture?

Speaker 2 (13:20):
I think not, not going to happen and the pushback
I would offer.
On that I agree with everythingyou're saying.
Right Is so for thecommoditization, like networking
was my job.
I was the person who built,created.
You need a VLAN, you come to me, right.
Then cloud comes along, youpush a button, you get a 172.16
and you're done and the networkjust magically exists.

(13:40):
So I hear what you're sayingand I agree with you that the
paradigm shift that cloudbrought to the networking world.
Like they figured out what wedidn't.
We be the networking industry.
Right, like all right.

Speaker 3 (13:52):
Well, now it's easy and now.

Speaker 2 (13:53):
it's quick and just push a button, right.

Speaker 3 (13:55):
So I don't know how.

Speaker 2 (13:56):
I can show value to my company when they're like
well, we can just do this incloud buddy and push a button.

Speaker 3 (14:00):
To answer your original question, though, what
I was going to say as far asgetting so, there's a big
difference between somebody thatis like newly wanting to come
into network engineering andsomebody that's already there
that's saying, hey, I need tostay relevant in this industry,
and I think and I want to talkto both of those audiences
because I think, yeah, sure, Imean, they're both talking to us
, right Wondering.

(14:21):
And if you're newly coming in, Istill think if you want network
engineering to be your trade,then you need to learn the core
fundamentals.
You need to learn TCP IP.
I have two TCP IP books on myshelf.
I still look back at themconstantly.
You need to know.
You know because routing pro,you know, bgps, you know runs on

(14:41):
TCP, you know.
It's one of those things whereyou have this stack and you
really need to learn thefundamentals and that carries on
into cloud.
So if you think about cloud Imean when I started deploying
connectivity to the cloud whatdo you think we're doing?
Like, okay, you have BGP, youhave circuits, you have path

(15:03):
prepending from A side and Bside.
You have all these things thatyou're still doing.
There's another side of this.
It's running these protocolsand you have to know how
failover and all these mechanicswork, still, even in the cloud.
So those skills, they are stillrelevant, but I think where
you're, the sort of the jobdisplacement is happening is
more on the hardware end.
I think a lot of it istransition to software and I

(15:25):
want to send this back over toyou, scott.

Speaker 4 (15:28):
Yeah well, so you give me some great things to
argue with you about.
No, sorry, Friendly, friendlyarguments.

Speaker 3 (15:34):
I told Cable to bring the popcorn.
Yeah, that's why we're hereI'll get spicy.

Speaker 4 (15:37):
Here we go.
I think the one big item herethat is kind of the definition
of commoditization is where weare in the growth profile of
building out the Internet.
You think in the 90s, right, andinto the 2000s, even though we
had the bubble, there was stilla lot of infrastructure being

(16:00):
built out.
We couldn't keep up rightinfrastructure being built out.
We couldn't keep up Right.
And you know, some of youprobably heard me talk about
that cowboy mentality right,where didn't matter what you did
, as long as you got it done.
You know you didn't have toshare documentation, you didn't
have to create documentation.
You know, if you love being theman, you know that was a great
environment for you to toachieve in, but maybe not the

(16:23):
best environment for a maturesystem.
Now, I'm not saying there's zerogrowth today, but it's not like
it was in the 90s, into theaughts, and now we're looking at
largely capacity upgrades ofinfrastructure that already
exists, so there's not as muchnew stuff to build and it is
limiting the hardware.
There's price pressure on thehardware right, and I've seen it

(16:47):
at the vendors I've worked forand I've seen it in the
disaggregated model right, wherenow I can buy, you know,
different packaging of switchesand routers from Acton, from
Edgecore, you know from othersuppliers and I can pick an OS
right, there's IP Infusion,there's DriveNets there there's
there's IP infusion.
There's drive nuts, there'sArcus, there's Vios and then

(17:10):
there's micro tick and I don'tcompletely understand micro tick
.
But there are all these othernon mainstream options that it's
great for competition, but thedrives prices down and therefore
it impacts the people whoactually make those.
You know, switches and routersand firewalls work.
So that's, that's a another bigfactor I'd throw into what you
laid out, william yeah, that'sfair yeah, um, the 5g, 6g thing.

(17:34):
Wow, let's do another episode onthat you know how.

Speaker 3 (17:38):
You know what I mean, though that is so
differentiated.
Like I tried to do, I tried toget into that space a little bit
.
It is deep.

Speaker 4 (17:47):
So it is, but it's a different set of technologies.
It's not like we understand BGP.
There are other people thatunderstand the 5G air interface
right and know what has tohappen, you know, with signal
propagation, and it looks likedeep magic to those of us on the
outside, but it's really justanother bucket of technologies
that can be learned.
And I'll just say when I landedat Nokia, one of the coolest

(18:09):
things about working there wasaccess to the mobility portfolio
and being able to learn in thatspace.
I've been working my way intomobility stuff since my first
mobile backhaul projects atJuniper as a tech engineer, and
this is part of the alwayslearning thing.
Right, and I think this playsright into the topic here

(18:29):
Mobility was not a core strengthof Juniper as a vendor, but
there were mobility solutionsthat they played a role in that
gave enterprising youngindividuals the ability to learn
if they are motivated.
Oh, mobile backhaul, what's that?
Okay, well, what's what's onthat?
What is this?
What does the radio do?
What does the baseband unit do,etc.

(18:50):
So finding things that you canat least get a finger in and
maybe put a hand or dive infully, you know, wherever you
can, is a great way to acquirenew skills and make yourself
more valuable.

Speaker 2 (19:03):
I want to touch on two things quickly.
So, william, foundationalnetworking knowledge, like you,
more valuable.
I want to touch on two thingsquickly.
So, william, foundationalnetworking knowledge.
Like you said, if you want tocome in somebody new, it'd be
great if we had a vendoragnostic accepted training
program, but it seems like theCCNA is the gold standard.
Is that sound correct to youand you, scott?

Speaker 3 (19:20):
like when you, when you tell somebody what to study.

Speaker 4 (19:22):
It's very practical.

Speaker 2 (19:29):
Like it's most encompassing.

Speaker 3 (19:30):
I never took the net plus, so I don't know what's in
it, but I get the sense that theccna covers more material and
more deeply I don't know if itgets more into nuts and bolts,
for sure and you have toactually configure things you
have to do subnetting andcomprehend like things that are
really valuable, that are goingto be with you no matter what
vendor you use.

Speaker 2 (19:43):
So and this and this will tie into that.
Like Scott, you had said aboutthe acceleration of change, you
know called up to me and I thinkthat's a really good point.
And then somebody in the chatjust said cloud AI, automation
they all seem to happen all atonce, like they all just hit us
at once, right, so, and that'sbeen so.
That was a Zetharian, I think,in the chat and that's kind of
how I felt too and I'll show youlike I mean it's crazy, I won't

(20:06):
do it, but I have a pile ofbooks right here trying to
consume, right, but I talkedearlier about I kind of felt
like some of my skills aged out.
Oh, you got your pile, yeah,yeah.
So you know I was, I alwayswanted to get my CCMP.
I never did, so I startedworking on that.
So I have a 900-page Encorebook that I'm halfway through.
I'm working on coding because Irefuse to code until I'm like,

(20:28):
uh-oh, I better learn this stuff.
So now I ran a Python scriptthe other day.
I have three different Pythonbooks here, one from Eric Cho,
one from John Capobianco, onefrom Edelman's folks.
My friend sent me an EVPNVXland book.
Right, I got a Cisco PI ATSbook I was reading about machine
learning tonight at mydaughter's gymnastics so like,

(20:51):
let's say this, but, but.
But that's the world we're inand it sounds schizophrenic when
you say it out loud Like so, ifyou're new and you're coming in
, get your CCNA.
Now, what I've been tellingpeople is you might want to
learn some Linux and some Pythonand some cloud as well.
Now, that sounds like a lot tothrow at somebody, because it
took me a year and a half to gothrough the Cisco Netacad, fail

(21:14):
the exam three times, finallypass the thing and then
interview for six months forthat first knock roll.
And if you were to tell me thenand, by the way, you're going
to have to learn programming andDevOps and Git and AI and all
the stuff we're telling peopleto learn.
So that's kind of where I amand I would just like to touch
on that somehow, if we can,whether you're new or whether

(21:36):
you're a guy like me, who youcall them the tradnet ops guys.
Right, scott, that's right.
Yeah, the old school guys whodid the job.
They were route switched, thatwas it.
Well, here we are.
Times are changing and ourskillset needs to be updated.
So what are those skills?
Can we talk about what you needtoday, either new or a trad net

(21:56):
ops like me.

Speaker 4 (21:57):
Can I, can I just pull you back to one thing.
You said, you said you, Iforget how you put it.
Your skills went out.
How did you forget how you putit?
Your skills went out?
How did you, how did you put?

Speaker 2 (22:06):
make the statement um I feel like I feel like my
skills kind of aged out or gotaged out.

Speaker 4 (22:13):
That's what you're relevant yeah, yeah, and I would
strongly encourage you to notthink about it that way.
Everything is additive and like, even if you don't ever have to
teach somebody IPv4, subnettingagain, you know that's a,
that's a foundational skill thatyou have and you'll be able to
tie other new things to it.
I don't know what those thingsare Right, but like I, I bounced

(22:37):
around a lot in my career.
There are jobs where I wish Ihad stayed in longer.
I should have given things achance longer.
That being said, gave me a lotof exposure to a lot of
different things, and I've youknow, on one hand, the way our
brains are wired.
It wants to make connections,it wants to see patterns, but
it's also really good at findingconnections to help us cope,

(22:58):
survive and thrive.

Speaker 2 (23:00):
So, yeah, I don't I'm not trying to pull you down but
I just that you're not agingout, you're just adding it's a
good reframe, yeah, well, likepeople like william and my buddy
tim mack, and like all theseguys that and girls who pivoted
to cloud and and abandon usnetworking people you want to
know something funny about that.

Speaker 3 (23:20):
so you you're talking about like AI and cloud sort of
or automation and AI all thesethings hitting at once.
And it's funny like if you canfind an opportunity within a
role like whatever you're doingwhere you can take one of these
things, you can apply it to yourjob and add value.
This is when it becomes veryreal, and this is exactly what

(23:40):
happened with me in automation.
So we were, I had written soanybody out there that ever had
to do, you know, generate CSRsand do the whole cert install
process with a CSS load balancerknows that this technology has
come a long way.
That was one of the worst thingsI ever had to do in tech.
It was awful every time, and sothat motivated me at the time.

(24:01):
I had to do so many of thoseearly in my career that I was
like there's got to be somethingI can do or else I'm going to
get out of tech.
It was like that miserable.
I hated doing that Worse thanlike fast food.
So that is what forced me topick up Bash scripting it was
actually a mix of Bash and Perland even some NetSNMP for some

(24:22):
things but bringing thesetogether and writing some
scripts to do that, and once Igot it working and I actually
had some help from a coworker atthe time that helped me with a
few things as well that is whatgot me zeroed in.
I was like, wow, this made myquality of life so much better,
and back then that was just onething.
But nowadays there are so manyopportunities where automation

(24:45):
can and it doesn't.
You don't have to be a softwaredeveloper.
I'm not a software developer,because I work with software
developers that are top notch,like in the Silicon Valley space
, and you see their code baseand then you see what I've
written and you're like, okay,you're comparing like apples to
like you know, dirt pretty much.
There's that much of adifference, so it's all in

(25:08):
context.
So finding something to apply,like where you can garner value
from it, is one way.
Like it's really hard to justgo out and say I want to learn
automation, I'm just going to goand make myself learn and I
don't really have a real problemto apply this thing to.
That's really hard, that's anuphill battle and I'm not
discouraging people not to codebecause of that, but it's just.

(25:31):
It's easier when, when you finda way you know to to apply it
to real world things.

Speaker 4 (25:37):
Okay, that was it, yeah and you know to take that,
william.
One of the ways to figure outwhat do I start with or what do
I do next is to figure out okay,what can I focus on for a short
period of time a relativelyshort period of time and get my
first job with it right.
So this is the new and you tostart learning stuff from the

(26:06):
book and on the job and alsounderstand what the customers
are dealing with.
It's never fun and peoplealways want to yell at somebody,
but doing those things at thesame time are super useful, and
I didn't do this right away inmy career, but I was a tech
engineer for five years andnobody ever called me to thank

(26:26):
me for keeping the internet up.
Today, all you're hearing aboutis the problems.
But you learn so much more whenyou see things break and then
going and reading about the OSImodel or just sitting in a class
.
Right, yeah, you could dotroubleshooting labs in a class
and that's helpful, but when yousee how things break in the

(26:46):
real world, you really get tounderstand.
Okay, this is how things reallywork.
Somebody asked in the chat ifCisco is really the de facto
certification engine today, andI would just say, from a
practical perspective, there's alot of deployed Cisco gear and
if you're going to startsomewhere, you're going to pick
something.
It's a, it's a, it's areasonable place to start.

(27:07):
I'll also plug.
You know, the juniper certprogram, the arista cert program
.
I know those fairly well.
I don't know the nokia certprogram like I should, having
been an employee, though aristaand juniper are doing really
good things there.
Um, but again, by the numbers,getting a CCNA is going to be
the most recognizable andprobably give you the biggest

(27:28):
chance or that first cut ofskills.
Unless you know you're goinginto a Juniper shop or an Arista
shop and then you could swapCCNA for JNCIA or the Arista L1,
for example.
So be aware, tailor it to theenvironment you think you're
going to be working in.
Hey.

Speaker 2 (27:45):
William, when you had to learn coding, you didn't
freak out right Like was itintuitive to you what this was
so long ago that network peopledidn't even.

Speaker 3 (27:57):
it wasn't even in conversations.
Like this was like whenenterprises were still
configuring like Linux servers.

Speaker 2 (28:03):
Well, that's why I'm asking like you seem like an
outlier, that you're just likeoh yeah, I'm just going to
figure out how to do some bashscripts, right it kind of goes
even further than that.

Speaker 3 (28:11):
So I had a major.
So I guess maybe this is toomuch info for this podcast, but
I had a major back surgery whenI was very young like major.
So my spine was so curved thatif they wouldn't have operated
it would have been I wouldn't beable to do much now.
So I have a pretty nicetitanium rod in my back.
So when I was recovering fromthis surgery it was really rough

(28:33):
and I was basically stuck athome.
I couldn't do anything formonths physical therapy and
everything and I happened tocome across a and this is like
it was a lifesaver at the time,but it was a how to install
Slackware magazine with like abunch of floppies in it.
I was like this is a coolscience experiment.
So I actually wiped out myparents I don't even know what

(28:54):
version of Windows it wasrunning but they went to the
grocery store and they left meat home and I destroyed that
machine.
We never got Windows back on it.
They were like, look, you justkeep your thing over there,
we'll go invest in a newcomputer.
That's cool that you're doingsomething you know, like not
playing video games, have fun.
So, yeah, I installed Slackware.
I got it running.
It was just a terminal, therewas no GUI.

(29:16):
And then I remember the firsttime, I remember where I was
sitting, but I put like four orfive commands in this file and
ran it and it did these thingsand I thought for some reason at
that time I was like, wow, thatis really cool, this could be a
powerful thing.
I was in middle school.

Speaker 2 (29:32):
Isn't it amazing that the formative experiences we
have and how they shape ourperception and, more powerfully
than that, our narratives thatwe carry?
I'm comparing my story withcoding to yours.
I tried to be a computerscience major in college because
I love computers.
I took stuff apart, right LikeI had that curious tactile kind

(29:52):
of thing, and they threw me inC++ and calculus and I got my
ass handed to me, even withtutoring, and I'm like, well, I
guess I'm not smart enough to dothis.
Now, why am I bringing that upYears later, working in
production, being told I have tolearn Python?
I had a very strong neuralpathway built in my head and a
story tied to it that I can'tcode, I'm not smart enough.

(30:14):
This isn't for me.
And Scott reminded me earlierwhen he reframed like your
skills aren't aging out, it'sall additive.
And then you had said somethingsimilar about like well, if
you're going to get in cloud,you need foundational networking
knowledge.
So I just love how much.
I don't know if psychologicalis the right word, but it's such
an inside game, a lot of thisstuff.
Like you had a good experiencebeing introduced to code.

(30:37):
Great, I didn't right.
But it just takes someonereframing it Like listen man,
it's just additive.
This isn't C++, and you guysboth know.
Last week I stole code from aclass I was taking that Dwan
Lightfoot teaches and I was ableto use a Python, a very short
Python script, and NetMeco tolog into a Cisco router.

(30:58):
I was like huzzah, this isgreat.

Speaker 3 (31:00):
So we're talking about subnet like core
fundamental skills that I thinkwe were talking about, the CCNA,
and then we were talking about,okay, cloud and AI and automate
all this stuff hit at once.
Oh, it's all this new stuff tolearn.
But again, perspective.
So I don't want to say I talkto network engineers every day,
but I talk to a lot of folksfrom a lot of big companies and

(31:22):
I get a picture into some commonproblems that network engineers
have to deal with.
And it's funny because one ofthe biggest, I'd say, pain in
the neck problems is IPaddressing in cloud.
You have all these networks thatwere built using the same slash
, 20.
And then it's like, okay, weneed these to talk now, we need

(31:44):
to talk to the data center, weneed to do this and we need to
do that.
And if there's job security outthere and you're putting NATs
everywhere, you're not going tohave developers that are
managing all these NATs and allthis complex routing.
And I mean that just goes toshow you that if these cloud
practitioners like towards thebeginning, if they would have
had a little, if they would haveasked for help, but at that
time they didn't need networkengineering help oh, networking

(32:06):
is going away.
You don't need those skills.
We're full stack developers, weknow everything and we'll come
to find out that.
You know, with some of thesebad experiences, you realize,
okay, maybe I didn't know asmuch as I thought I knew in the
cloud, and we need skilledpeople in storage.
We need skilled people innetworking.
We need skilled people insecurity.
You know we actually need thosethings because the alternate

(32:33):
universe of if we don't havethose things is we've built all
this stuff and none of it cantalk and we can't get anything
to work, we can't get ourapplications to run and hey,
that's a problem.

Speaker 2 (32:46):
You're going to get me on a soapbox because we've
we've got rid of so many networkoperators the past couple of
years At least that's myperception with the layoff stuff
.
And then we're going to replacethem with software to find this
and this other overlay and thisAI automation stuff, and like,
hey, look out, you know, yeah,our solution, our software, our
magical software that's going tomanage all your stuff for you.
Sure, it's expensive, but it'sa heck of a lot cheaper than all
those humans that you just gotrid of.

(33:07):
Like isn't that great.
So, to your point, william,like when the cloud came by, oh,
it's going to be easy to pressa button.
And then we come to find, well,we don't need as many people.
There's these two narrativeswith automation.
I hear is you're going to freeup your people to do more higher

(33:28):
value work, right, well, okay,or you can do the work of 10
people with one person who'sreally good at these automation
tools.
And, similar to that cloudproblem, they ran them with the
IPs.
I'm afraid we're going to haveless knowledgeable networking
people running biggerenvironments with this fancy
tooling.
And if all hell breaks loose orthey find like, oh crap, we
messed all the IP addresses,like there's not going to be

(33:49):
enough people to fix all thenetworking crap that they
screwed up by software defining,automating and AI-ing
everything because they weregoing too fast and they were
chasing the money and theydidn't realize all these
mistakes they're making alongthe way.
I'm off my soapbox now, but it'sa parallel that I feel is like
the same trend happening withgoing lean with people and

(34:10):
maximizing.
If you're a publicly tradedcompany and you have to show
profit every quarter forever,it's impossible, right?
I mean the whole system again,another soapbox.
That isn't this show, but thatis impossible if anybody has any
sense of.
I'm not an economist, so youknow you're going to cut your
people when you can and cook thebooks to make it look good
until you can spin up the newthing.
But I'm concerned that you knowwe're getting rid of network

(34:34):
people and replacing them withsoftware long term.
That doesn't seem great to me,am I totally crazy?

Speaker 4 (34:41):
I don't think you're crazy.
And there is that crowd or thatpersona, an IT leader, IT
director, IT VP or CIO that saysI'm sick and tired of having to
have CCIEs to run my network,and that's a quote, and I've
heard that.

Speaker 2 (34:59):
So many times.
But networks are more complexthan they've ever been.
That's true, right.
I'm not arguing with you, butright that's the counter.

Speaker 4 (35:04):
I'm not supporting, I'm reporting.

Speaker 2 (35:08):
I'm with you, right, and I get their concern.

Speaker 4 (35:10):
It's a thing.
Right, it is a thing.
But the other side of the cointhat I'm hearing more and more
of over the last few months is Ican't hire networking people.
I can't.
I've had this open rec for ayear and none of the new college
grads want to do this.
They all want to work withWilliam on cloud stuff and sorry

(35:33):
.

Speaker 2 (35:33):
And I'm not saying that's not true, but, william, I
don't know if you would rejoinyet but he has a buddy who's
like a mid-level guy who isapplying for a net ends job and
there's like 750 applicants.
So they say they can't getpeople.
But then there's people out ofwork competing against hundreds,
if not thousands, of people.
So like I don't know if bothare true, right, I don't know.

Speaker 3 (35:54):
Actually from the hiring perspective.
So when I was at my last role Iworked for a Fortune 50
healthcare provider.
I was building two teams at thetime, so one of them was a
cloud Like we were four cloudsdeep.
At the time it was crazy.
We had more cloud work than weknew what to do with and a lot
of the problems we had on theinfrastructure side pertained to
networking.
But then we were also doingsome data center projects.

(36:16):
So one thing I found reallyquick when the floodgates of
resumes came in and I startedoriginally hiring for the data
center centric role, theseresumes like every single one
was inflated so bad.
And then when you start talk.
It was like one in like 50where you had like an honest

(36:36):
resume and that is so timeconsuming because HR is fielding
these things.
They're looking and validatingmaybe that you have the
certifications, graduated fromwhere you said you did, and all
this stuff.
But if you're coming on andyou're wanting to work in a data
center and you say you have 10years of experience and you have
a CCNP and you can't walk methrough, okay, like bits into

(36:58):
this layer two, okay, now we'reframes.
Now, okay, now it went in thisinterface and this routing
protocol, now we're packets,like if you don't know the, if
you can't talk a story through anetwork device, but you have
all this experience and allthese certifications and all
this stuff, like, okay, whathappened?
Like these are that's about asbasic as you can get.
We're talking like VLANs andVRFs and you know pretty

(37:21):
fundamental network engineeringstuff.
So that was one of the problemsin hiring was resumes that were
just fudged to.
You know, mount Everest.

Speaker 2 (37:30):
It was just crazy and that's a hard thing to work
around because time yeah, and Ididn't mean to cut Scott's legs
out from under him, because I'veheard the same thing that they
can't find people.
And there are college kids thatdon't want to be in networking,
like, oh, I'm going to be incyber, you're going to secure
networks that you don'tunderstand how they work, right,
like, I guess, is that a thing?
I don't know, it doesn't soundlike a thing, but I've heard the

(37:52):
same thing, scott, so I'm notsaying what you're saying isn't
true.
They can't find people.
You, I know that there's plentyof people talking to college
students and graduates who arelike I don't want to go into
networking, it sounds awful, Idon't want to do that, it's not
sexy, I want to be AI or dev.

Speaker 4 (38:05):
We always chase the new squirrel right, the new
shiny object.
Where else did I want to takethat?

Speaker 3 (38:12):
Sorry, I really derailed that.

Speaker 4 (38:14):
No, it's okay.

Speaker 2 (38:16):
No, it's a great conversation.
I think software is taking ourjobs.
I think that publicly tradedcompanies don't care because
they have to extract everydollar they can and people are
expensive.
And that concerns me.
And the only reason I say itout loud here is because we've
been running a networkengineering-focused podcast that
tries not to lie to theiraudience, who seems very
invested in what we're tellingthem.

(38:37):
And if we come on here everyweek and say it's great and
everything's wonderful and thisis the best job in the world and
not cover the entire, we haveto objectively cover the news
right and what's happening.
I mean, I was joking with AJthe other day.
I think we have one, twonetwork engineers left on our
show.
Aj went to security and I wentto a vendor and somebody's in SE

(39:01):
.

Speaker 4 (39:02):
Let me reframe that for you again.
Oh good, so he went to Zscaler.
He went to Zscaler.

Speaker 3 (39:08):
Don't give him a free plug.
Yeah no, we're talking realityhere.

Speaker 4 (39:12):
We're talking reality here, right, but he didn't lose
his network engineering skills.
He's reapplying them and he'slayering new skills on top.
And to park on the securityfield thing for a bit here Wow,
what a forced ecosystem ofinnovation.
If you want job security,security is a really good place

(39:34):
to be.
Or skilling up in security andlike even just calling it
security as one thing is notfair.
Um, there are so many differentsubcategories, like I was.
I was fascinated, like whenjuniper bought net screen.
You know, I was a carrierrouting guy, right, and I knew

(39:54):
bgp and I knew sonnet and ATMand I knew all these big iron
things.
But security guys know portsand protocols and they
understand.
Okay, this is what thisapplication does on the network
and it's a really interesting.
So going into being a firewallperson, right, you need that

(40:15):
basic understanding of what'shappening at the network layer,
but you've got to be looking upat the applications and how
they're using the network.
So that's another valid placeto skill up.
That area sounds like they'realways looking for people and I
hear I've been involved in acouple of different state
security-focused organizationswhere they're always working

(40:37):
with colleges and universitiesto entice new grads into cyber,
especially in the DC area.
You know it's what everybodywants to do.

Speaker 3 (40:45):
And if you know network engineering, I think
that's almost a cheat code asfar as like elevating yourself
to security, because if you knowhow the nuts and bolts work I
mean that was one of the biggestchallenges in the past that I
had working with like parallelsecurity teams is we're just not
talking the same language, likethey're not.
Some of the security teams Ihad to work with over the years
were not technical at all.

(41:06):
That means a lot ended upgetting punted over and fell on
the network engineering side andnow you're seeing these two
fields actually consolidate.
There's a massive consolidationgoing on.
So it's almost like a cheatcode.
If you've got networkengineering under your belt and
you do want to level up to adifferent area for job security,
security is a great place to go.

Speaker 4 (41:29):
So that cheat code would have helped me with your
World of Warcraft referenceearlier today that I did not get
.
Is that what you're telling me?
Sorry?
I can't help you there.
Nope, I'm not that guy, I hadto Google it.

Speaker 2 (41:41):
So I guess to stay relevant right in the job market
, you have to try to see aroundthose corners, like where are
things headed?
And I don't know if it's true,but I thought I had read that
what's behind a lot of Ciscolayoffs the past year is they're
reinvesting in cyber and AI.
So you reminded me, scott, withhow security is a big thing.
They've decided they're the twomarkets to go after.

(42:02):
So I don't know again mycognitive bias, but like, oh
well, if the biggest networkingvendor with the biggest market
share is making this huge shiftto these two areas, maybe that's
the target, maybe that's whereI should be.
And when you guys were talkingabout like cyber and cloud, I

(42:23):
just had a thought I never hadbefore.
It almost seems like networkingit's not necessarily a place
that people might stay right,like it's a jumping off point.
You come in, you pay your dues,you learn the network and then
you go to cloud or you go tocyber or you like.
You know.
I know that's a generalization,but I know that we're having a
hard time keeping getting peoplein networking and keeping them

(42:45):
in networking and if networkingis a foundational skill for
these sexier, higher paying jobs, it's going to be hard to
retain talent, I think,long-term.
I don't know if there's anytruth to that, it's just a silly
thought that popped into mybrain.

Speaker 4 (42:59):
I mean William.
Speak to that because you'velived part of that life at a
minimum.

Speaker 3 (43:03):
Well, one piece of advice and just, you know,
hindsight to 2020.
So when I look back and when Imade the big pivot to cloud and
I actually had a cloud role likecloud was like my sole
responsibility.
I'll tell you what.
When they say, the grass isalways greener on the other side
, and then you get over thereand you're like, oh boy.
So one of the reasons I wantedto get out of network

(43:25):
engineering was just aculmination of different.
You know, working in largeenterprise is tough.
When you see, like networkbudgets getting cut for other
newer, fancier, shinier thingsand you have major problems, you
don't have the budget to fixthem.
You can't get the change.
You know, you can't get changemanagement to work with you to

(43:47):
really take certain things downlong enough to make real changes
and really go back and fix someof that technical debt that's
kicking you in the butt.
I just got so tired of that,along with the fighting fires
that ensued from not being ableto fix this stuff, that I was
just like I, I can't, I needthis pivot, because it was
driving me crazy, if I'm honestI don't know that sounds so

(44:09):
familiar publicly, but this is atrue story.
Um and so like I get over tocloud, and cloud was in the
beginning it was nice and shiny.
A lot of times it was flyingunder the radar of change
management, like hey, I don'tknow what we're doing.
You know, we're our own likewe're Luxembourg, we're our own
bosses here.
Come on, and so is that sort ofevolves, though in enterprises

(44:32):
they realize, okay, this is abig part of our business we're
actually using, you know, movingsome of the crown jewels up
here.
Then they start bringing allthe enterprise things and
wrapping it up around the cloudand all the things you do in the
cloud.
You end up with a lot of thesame challenges I was sort of
outlining there on the networkengineering side.
It's a little bit different,but still the same thing.

(44:52):
But another thing I wanted tosay, and I don't want to change
the subject too much but when wetalk about networking being
networking, it doesn't matterhow high up the OSI model you go
, all the way up to applicationnetworking, we're still TCPIP,
BGP, we're tunneling the crapout of everything everywhere, at
every level, reverse proxyingthings, like it's never going to

(45:16):
go out of style.
This is all networking.
But I'll tell you what.
The thing that's changed is thetools you use to interact with
it.
So instead of hammering outthose things on those you know,
the Cisco, Juniper, Arista, CLIsyou're basically defining your
network configuration in a file.
You're saying, hey, I want tocommit that to a Git repo.

(45:36):
Some sort of trigger happensthat you set up a long time ago
that you probably still don'tunderstand, and that's going to
go and do some things.
It's going to have someautomated checks involved and
it's going to deploy someinfrastructure.
So it's really the tools andpractices have changed more than
the actual technology on theother end.

Speaker 2 (45:55):
That's kind of my hot take for the infrastructure as
code is something to add inthere, right like that's yeah
sure yeah terraform open tofu.

Speaker 4 (46:03):
Yep, yeah, I would.
I would say so if we, if wehave that plate of what are four
or five things that, if we wantto survive in networking, we
need to understand better.
You know that tooling, githuband python definitely fall in
the category of how I interactwith the equipment.
Right, and I'm don't don'tpaint me as the you know, the

(46:25):
CLI lover that I am.
I just wanted to go into WillFerrell on Saturday night live,
but the CLI is still there andfor all the threats of turning
off the CLI that I startedhearing 15 years ago at that
company that starts with a J, tomy knowledge it hasn't been

(46:48):
turned off yet.
But I have a broader variety ofways to interact with and
instantiate services on switches, routers, firewalls, et cetera.
So you're absolutely right,william.
Right, there's more ways tointeract with it.
The other two very practicalthings you know.
To Andy, to your comment of youknow being able to see around
the corner this isn't lookingaround the corner.
This is the train coming rightat you, or both of us, all three
of us, learning some Linux andlearning a cloud platform like

(47:13):
Linux, github, python and acloud platform platform Once you
have networking skills, whetheryou've gotten them from
experience or you've done yourCCNA and your CCNP or JNCIX.
Those four areas and AWS isprobably a really good place to
start for a cloud platform, justlike the CCNA is probably a

(47:35):
really good place to start foran entry-level cert.

Speaker 3 (47:37):
Yeah, and getting in there and doing it.
So cloud training as far as I'mconcerned, it's free.
The free tiers I hosted a blogfor a year using free tier stuff
.
You can do a lot with the freetiers, and especially Oracle.
If you want to learn some OCIand you want to have a lot of
infrastructure for free, oci hasa very amazing free tier.

(47:59):
So all the resources likeinstead of.
You know, back when I waslearning networking, I had to
order all this stuff on eBay.
Well, I actually had to get mymom to order all this stuff on
the internet, get it deliveredto the house and all this stuff
and set it up, and sometimes itworked and sometimes it didn't
and wow, that was a time hog.
But now you can start learningright now, where you mean you

(48:19):
could spin up an aws account anddeploy a vpc and a subnet and
an ec2 instance while we'retalking right now, in a few
minutes.

Speaker 4 (48:26):
You know it's all free, you know I think of um,
who was the individual whopresented at the kentucky nug
that you and I were both at, andhe did his yeah yeah, his
automated his cloud lab becausehe got sick and tired of
replacing hardware in his homenetworking lab.
That presentation is gold andI'm sure he would talk to people

(48:48):
if they were interested in thisis how I built it and that's a
perfectly safe place to learnand touch all of the things that
we were talking about.
He did it on GCP, but I'm sureyou could do something similar
on AWS or OCI.
So, yeah, stuff like that in alab and in a free tier and not

(49:08):
having to go on eBay and spendyour last penny, you know you go
buy more Marvel comic books.
You don't have to spend it onservers.
You know those are really goodways to kick tires in a very,
very risk-free environment.
A hundred percent.

Speaker 2 (49:23):
William, when you were talking about pivoting from
networking to cloud and howsexy cloud was and people left
you alone, pivoting fromnetworking to cloud, and how
sexy cloud was and people leftyou alone, and cloud does seem
to be.
You know, a first-class citizenlike you know where us network
guys are, just you know slummingit.
And for a second I thought,well, I wonder, you know, like
I've always heard that, becauseinfrastructure networking is a

(49:46):
call center and not RevGen, thatthat's why we're so, you know,
loathe, despised, kicked likedogs, right, but cloud seems to
be treated differently, withmore respect, with a little bit
of you know, but cloud's also acost center, right, like whether
you're building on prem or incloud, like why is?

Speaker 4 (50:05):
cloud treated.

Speaker 3 (50:06):
Where's Corey Quinn Cloud is the biggest cost center
.
Cloud is the biggest costcenter Right.

Speaker 2 (50:11):
But it seems to be, and maybe it's just the speed at
which you can deploy right and,like I mean, I get the benefits
of cloud.
I'm not having thatconversation, but cloud just
seems so revered by decisionmakers and I always thought it
was because we were a costcenter, which is why
infrastructure were such dogs.
Always thought it was becausewe were a cost center, which is
why infrastructure were suchdogs.

(50:32):
But if cloud costs, you knoworders of magnitude more than
on-prem, I guess it's just thespeed at which you can deploy is
why it's such a darling child.
I don't know, I'm just tryingto get my head around the
difference between cloud andon-prem and their heads.
But if they're both costcenters and if cloud's more
expensive, that framing is wrongfor me because it seems like

(50:52):
going to cloud is better right,like networking or going to
cloud.
There's more jobs there.
It's better, companies areinvesting in it, like that seems
to be a good place to go yeah,a few things there.

Speaker 3 (51:03):
Do you have something to say?

Speaker 4 (51:04):
scott, scott, I do, I do there's a, there's a, so
I'll just I draw a parallel here.
But it's back to growth, right,the internet had a lot of
growth.
A lot of people could get jobs,do well, and then it
commoditized.
We've seen 15 years of cloudgrowth and now we're just we're
starting to hear the drumbeat ofcloud repatriation.

(51:25):
Right, and maybe it's not ascool as it was was, but it was
still this thing that I had tohave, you know, as a corporate
manager, director, etc.
And now ai is becoming thatnext thing.
Yeah, so, and it'll besomething else after ai, right?
Um, but that's, that's likethat, I think.

(51:45):
I think that's what drives alot of the trends, and it's not
all based on solid science, asit were.
We're more impacted by emotionand marketing.
That I think, we want to admit.

Speaker 3 (52:00):
Oh, that is exactly where I was going, scott,
emotion and marketing.
Okay, let's go.
I remember the first time I wasin a meeting with a C I see
something.
I can't remember what the Ctitle was, but there was a big
mandate.
It was hey, cloud computing ishere, we're going to be out of

(52:21):
our data centers inside of like18 months.
And of course, everybody that'stechnical in the room is like
see all those zero downtimemainframes over there, rows and
rows and rows of them.
Yeah, we're just going to.
Yeah, not how it works.
So there was such a push fromthe cloud providers.
There was all these promises.

(52:43):
There was this big.
One of the big things was, oh,capex to OpEx.
I don't know how many meetingsI was in just talking about
CapEx to OpEx.
It's changing the world, it'sgoing to make all the business
stuff better, blah, blah, blah,blah, blah.
Well, it turns out accountingwith a lot of these companies.
That has been around for areally long time.
You want to know what's coming.
You want, like, you have this,it's just a different.

(53:05):
It's not as easy as saying, ok,we're just going to go complete
OpEx, we're going to do pay asyou go, we're not going to sign
an EDP for a long-term commit.
We're just going to bite thebullet and just keep spending.
Well then, your spending is outof control.
It's up and down.
Nothing's predictable.
There was this huge thing oflike, okay, cloud is our way to
get out of our own data centers.

(53:26):
We're not going to have datacenters anymore, we're not on
premises, that's not going to bea thing anymore.
And then, as you, as things getreal and things only get real
when you move that firstapplication, that complicated
application that has data spreadout and you know, across
multiple data centers, you havemultiple stateful things in the

(53:47):
way.
You, you know you have allthese different things that you
don't know about.
Oh, this needed this over hereto talk and it had this over
here.
And oh, one of the people thatwrote this part of the
application isn't here anymore,which has been kind of so taking
these giant monolithicapplications and peeling them
apart and making them, becauseif you do lift and shift, you've

(54:07):
like it's like you've alreadylost.
So you really want to use thecloud in the way the cloud was
intended to be used, which isnot lift and shift.

Speaker 4 (54:16):
No, I want it to work just the way it worked in my
private data center.
And they don't understand thatcloud is about operations as
much as it is about technology.

Speaker 3 (54:26):
How much money do you have?
If that is your goal, youbetter have a lot.
So I guess what I'm saying isthere was a lot of perception
and just again like not livingin reality.
And then when you move thatfirst application life, you know
it becomes real.
And then, okay, you try andmove another application and one
company I work for we had likeover a thousand applications and

(54:50):
it's like, okay, we like, okay,we're gonna, okay, we have this
as expedited timeline to moveall these applications to the
cloud.
And you know we could barelyget one off the ground because
of how complicated theseapplications that have been
around forever.
Um, it's not.
It's not as simple as snappingyour finger.
You know there's no amount ofprofessional services and all
these agencies that you canbring in that's going to make
that any better if you're goingto take down that application

(55:12):
and continually break stuff inorder to make it move faster.
So, as these lessons startbecoming real and real, and to
Scott's whole theme ofoperations, operations is just
the core for things runningsmoothly and you know it's just
reality Reality hits you andit's like, okay, well, maybe

(55:33):
we're not going to just move allof our eggs in the cloud basket
.
You know what.
Maybe it's a good idea we haveour data centers.
Let's run the workloads thatmake sense in our data centers,
that are cost-effective there,and then the things that work
really well in cloud, that weunderstand, that we really
repurpose, let's run those incloud.
It doesn't have to be all ornothing and it never I don't

(55:55):
believe it ever will be becauseI believe AI and the hardware.
I think there's going to be aresurgence in hardware in the
data center to some capacitybecause it's going to be more
cost effective to run certain AIworkloads.
Just a prediction.
I could be completely wrong.

Speaker 4 (56:15):
No, you're spot on.
Jensen is making hardware coolagain.
Um, he really is.
With that jacket, yeah, I wantthat jacket, but uh, yeah, the,
the gpu now as the nextfundamental unit of compute,
right, that's, uh, that's goingto be something to watch for the
next few years at a minimumpeople laughed at me when I
mined ethereum in my home miningrig.

Speaker 2 (56:30):
And now look at it.

Speaker 3 (56:32):
Who mines crypto?

Speaker 2 (56:36):
Well.
I can't anymore, but I'm allabout.
Next, william, you're going totell me AI is not going to solve
all of our problems.
That's what you're going to.
Right Cloud didn't solve allour application problems.
Next, you're going to go on therecord as saying we're going to
go on the record as sayingwe're going to get into all this
, all the same issues with AI.
That's where you're going,isn't it?

Speaker 3 (56:52):
How dare you A lot of new.
Yeah, but we really didn'treally touch that much on AI
actually.

Speaker 2 (56:59):
You know what we did, the community of service.
Is anybody sick of hearingabout AI?
I don't know.
Well, I mean, we're at the hourI was going to sum up, but if
listen, we can do whatever thehell we want.
If we want to touch on AI andthen wrap it up, we can.
I don't have much to say aboutit.

Speaker 4 (57:20):
I'm just reading a machine learning book, trying to
get my head around what thehell's going on.
It's all about getting throughthe hype right now, you know,
and I feel like in tech, but innetworking in particular, we're
really susceptible to these.
Here's the message about thenew thing, and the new thing is
it and it happened with SDN it'shappened with things before SDN
.

Speaker 2 (57:40):
Can you define who we is, Because I think it's people
who don't manage networks right.

Speaker 4 (57:46):
I think it's a slider bar, but I think it's part of
like we've been beaten intosubmission on it and I think
there's a marketing motion that,because it's so hard to get
ears and eyeballs on messages,there's this pressure for
marketing to say and do whateverthey can and I noticed I didn't

(58:07):
say reasonably can whateverthey can to get attention.
And then once they get attention, we can work the truth in there
somewhere.
I feel like it's worse thanit's ever been in my career and
I think that's very fertileground for lots of hype around
AI.
I think, because of where weare over the past 30 years,

(58:28):
we're now ready to and we'retired, we're exhausted and we
just, like you know, we eitherbelieve it or we don't pay any
attention to it.
That's something I thinkvigilance we all really need to
be hyper, hyper vigilant about.
It's like okay, I hear thismessage, I've listened to this
podcast, what's really happeninghere?
And it might not be an easyanswer, right, it's going to

(58:50):
take some digging for us andwatching you know what happens
in the next nine months, 12months, et cetera.

Speaker 2 (58:56):
It's probably just another tool right.
There's a ton of hype, we don'tknow where it's going to go.
So you know I've played aroundwith chat GPT a bit.
But the coolest like networkingcentric thing I did recently
was I spun up a couple of SRLinux nodes in Docker with
Container Lab.
I'm throwing around words tosound smart and show that I've
been studying, but then Iinstalled their SR Linux GPT and

(59:22):
this was huge, because I grab atoken from ChatGPT, throw it in
there, give them five bucks forthe API calls and then I can
just talk to the device ineveryday language.
And we all know I mean, if youcome up in Cisco and then you
have to go to another vendorwith different syntax, that's
the friction.
It's like oh my God, what isthis and how do I do the thing.
But I could talk to it and Isaid ask AI in the prompt in the

(59:46):
damn CLI, how do I configuretwo IP addresses on two
interfaces?
How do I configure BGP toneighbor up and how do I
advertise loopbacks across?
And just what I said to you andit gave me everything I needed.
To me that was like I don't knowif a watershed moment's the
right term, but like, oh my God,I am talking to a device via

(01:00:06):
the CLI in natural language andit's giving me the exact syntax
commands I need, which is thejob.
Like, hey guys, we're going tothis other vendor, learn the
crap, oh God.
And then you get adocumentation.
And every vendor documentationis a nightmare, right, but I can
talk to it and it gives me whatI need, like holy crap, even if
it's not perfect, if it's close.
So I got really excited when Idid that.

(01:00:28):
But again, there'shallucinations and a lot of hype
and all that stuff.

Speaker 4 (01:00:31):
But I think that's probably the number one most
practical thing right now fornetwork engineering.
Natural language processing andyou and I have seen it in other
places too right, we saw it atthe Nokia Data Center rollout
day that we did for Tech FieldDay, I saw it at Meter and this
is becoming very mature andadvanced very quickly.
And the being able to pivotinto other language support.

(01:00:56):
You know, think about the easefor product owners right when I
need to do other regionalization, for you know other languages,
so I think this is one real areathat's going to get interesting
.

Speaker 2 (01:01:09):
And we're both former Juniper people.
I remember seeing a missed demowhere they said in natural
language show me all the usershaving a bad wireless experience
right now.
Like what?
And then it pulled it up andthat's you know.
I didn't get much further inthat demo but like oh my God,
you're talking to yourinfrastructure and it's giving
you meaningful information.
That's pretty amazing to me.

Speaker 3 (01:01:27):
I think one interesting thing to watch out
for is following the money,because one thing that you
notice right now is if you lookat big enterprises, what are
they buying?
As far as AI is concerned, well, there hasn't been that moment
of okay, this is a productthat's actually an AI product.
It's not an existing product onthe market.

(01:01:47):
It's just throwing, you know,machine learning or something on
top of it and making that partof the product portfolio.
But if you look at theoutcropping of new startups like
I try to keep track of all thestartups, just to kind of
Because if you look at thestartups that are coming up,
these AI startups that aregetting the big checks from all
the you know, the deep-pockedfolks out on the West Coast

(01:02:09):
seeing some of them the problemis like AI startups can die
really quick because they gothrough and just obliterate
their you know their burn ratesjust so high.
Sure, you know, because it's somuch compute, you know it's so
so much intensive resources torun, to test and to build, so
things fizzle out pretty quick.
But the ones that are actuallygenerating value, that are going

(01:02:30):
to eventually become theproducts that enterprises are
going to buy to do reallytailored or specific things.
I think is an interesting thingto just watch out for, because
really we want to know where isthe value proposition for a big
company.
Is it just going to continue tobe these existing products that
are layering on?
Okay, it's a feature.
Ai really is a feature, it'snot a product.

(01:02:51):
How is that going to shape upover the next five years?

Speaker 4 (01:02:55):
Well, and to the example Andy cited with Mist
right, they weren't selling AI,they were making a better Wi-Fi
management system, right, and soyou could call it a feature, or
you can call it just anotherset of computational techniques
that I have available to me tomake my product better, faster,
stronger, and I think that'sprobably the second.

(01:03:18):
I think we're going to see itembedded more and more in
products just as a featureenabler, and we're not even
going to know.
It's AI or that's using an LLMor a small language model to do
X or Y, but developers are goingto be.
It's AI, you know, or that'susing an LLM or a small language
model to do X or Y, butdevelopers are going to be able
to make great use of it.
To you know, do better thingsfaster.

Speaker 3 (01:03:37):
Yeah, the small, you know, the small language models
are going to be huge becauseit's a specific set of data,
it's not a broad set of data.
The training game changes andit's cheaper, and you know, the
list goes on.
That's going to be a huge partof the future.

Speaker 2 (01:03:51):
All right, guys, this has been a fantastic show.
I could stay on with you allnight.
I learned so much I got acouple things that I just
circled here for key takeaways.
I really liked, scott, what yousaid about things being
additive, and you apply that toa lot of different things, right
.
New skills that we're learningare additive to networking.
Even where we ended just nowwith the AI conversation, ai

(01:04:14):
might be additive to otherproducts that are there, that
just make them a little moreusable.
It's a different interface,right.
So I really like how you said.
It's really important, I guess,how we frame things in our
perceptions and how we talk toourselves and me saying I'm a
network dinosaur because I don'tknow all the hotness Well, no.
Talk to ourselves and me sayingI'm a network dinosaur because
I don't know all the hotnessWell no, you just have to add on

(01:04:35):
to that.
And networking is always there.
If you're going to pivot tocloud or cyber, you need a
network foundation.
William, you said somethingabout making sure you're adding
value, which I think is a reallysmart thing.
You got into automation to addvalue to your company.
That's always going to be agood thing, especially if you're
a call center, like, oh, he'sdoing extra things and this is
helping us, like that's reallysmart, I really like that.

(01:04:55):
And then I really like whatScott said about like go where
the growth is.
Right.
It was the internet, it was andmight be cloud and maybe it's
AI now, but we're having theseconversations about you're a
network person, you're trying tostay employed.
You're trying to stay employed.
You're trying to see what to donext.
You can add on to your skills,which never ends, right.

(01:05:16):
Networking is always changingand tech is always changing.
And if you can try to staywhere the growth is and add some
value while you're there, Ithink that that's really smart.
You added infrastructure as code, which I think is a good call
out right, and this list neverends GitHub, infrastructure as
code, python right, likeTerraform.
I mean that's to me.
I get all nuts and freak out.
I'm like it's too much, I can'tdo any of it.
But just pick one right, likeautomation's important.

(01:05:37):
So recently I've picked Pythonand that's what I'm going to do
and I'm starting to makeprogress, because if I, I do
nothing.
So maybe just pick one and geta small win.

Speaker 1 (01:05:49):
And how deep do you?

Speaker 2 (01:05:50):
have to go right.
Yeah, like you said, you don'thave to be a developer.
But if I can look at a Pythonscript and kind of know what
it's doing and grab code to dothings, maybe that's enough.
Right, like I don't need to bea full stack developer as a
network person.
Right, like, if I'm going toget into cloud?
I mean, I've seen Terraform,it's not terrible.
Infrastructure's code's kind ofcool.
You know it's not that foreignfrom what you're doing in

(01:06:12):
Notepad and CLI.
So I think there's a lot of wecan transfer a lot of these
things and they're all veryrelevant to what we're doing.
So thank you guys for joining.
I really appreciate it.
You can find all of stuffnetwork engineering at our link
tree, link tree for slash art ofnet eng, something I always try

(01:06:35):
to call out as our community.
There.
It's a discord server called.
It's all about the journey.
Right now we're running a CCMPstudy group, running a CCNA
study group.
There's new people coming in.
There's triple CCIEs who are inthere helping.
We recently did wirelesschapters in CCMP and Bart Castle
from CBT Nuggets came in andspent like an hour and a half
teaching us wireless like forfree people.
So that's awesome.

(01:06:55):
Yeah, we can continue to havethese conversations.
So if you're looking for acommunity to join, if you're
looking for support, go tolinktreecom.
Forward, slash artineteng,click the all.
It's all about the journey link, as always.
Thank you so much for joining.
It's been a great conversationand we'll see you next time on
the Art of Network Engineeringpodcast.
Hey everyone.

Speaker 1 (01:07:14):
This is AJ.
If you like what you heardtoday, then make sure you
subscribe to our podcast andyour favorite podcatcher.
Smash that bell icon to getnotified of all of our future
episodes.
Also follow us on Twitter andInstagram.
We are at Art of NetEng, that'sArt of N-E-T-E-N-G.
You can also find us on the webat artofnetworkengineeringcom,

(01:07:34):
where we post all of our shownotes.
You can read blog articles fromthe co-hosts and guests, and
also a lot more news and infofrom the networking world.
Thanks for listening.
We'll see you next time.