November 30, 2024 • 30 mins
Discover the cutting-edge world of technology with Carl Solder, Cisco Systems' CTO for Australia and New Zealand, as he takes us on an enlightening journey from the 1990s to today. With nearly three decades of expertise, Carl shares his insights on iconic innovations like the Catalyst 6500 switch and the challenges businesses face in adapting to rapid technological changes. This episode offers a comprehensive look at the relentless pace of tech evolution and how Cisco continues to lead the industry in transformative ways.
Explore the future of work as we discuss the shift to hybrid models and the innovative solutions Cisco is developing to support this new landscape. From enhanced collaboration tools to AI-driven devices, learn how organisations are reimagining their workspaces to boost productivity and foster collaboration. We also highlight the strategic acquisition of Splunk, emphasising the critical role of data in modern troubleshooting and observability, as well as the design of Cisco's new sales office in New York that embodies these forward-thinking changes.
Finally, we address the intersection of technology and sustainability, focusing on the exciting potential of smart buildings and simplified networking. Carl Solder shares how Power over Ethernet (PoE) is driving energy efficiency, helping achieve net-zero goals. Delve into Cisco's AI strategy and the emerging trend of repatriating workloads on-premises, offering insights into the economic and operational benefits of such a shift. This episode is a must-listen for anyone keen to understand the current and future state of technology in an ever-evolving digital landscape.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Michael van Rooyen (00:00):
Welcome to a special Cisco Live edition of
Securely Connected Everything.
I was fortunate enough to spenda better part of a week with
Cisco for the annual APJC CiscoLive event held in Melbourne
each year.
During that week, I had theopportunity to catch up with a
lot of interesting people fromCisco, of course, but also from
across the technology industry.
The great news for listeners ofthe podcast is that I got to
(00:20):
record many of thoseconversations at our Oro Live
podcasting booth, which we hadset up on site throughout Cisco
Live.
You'll get to hear a number ofthose discussions next season
when we kick off in late January2025.
I couldn't wait to Len to bringin today's guest, though, so
we're bringing him to you now asa special two-part Cisco Live
edition of Securely ConnectedEverything.
(00:41):
Without further ado, here ispart one of my discussion with
Cisco and ZCTO Carl Solder.
Today I have the pleasure inhaving a discussion with Carl
Solder, who is the CTO ofAustralia, new Zealand, for
Cisco Systems.
He's been at Cisco for a longtime and he'll certainly cover
that in our discussion, butwe're at Cisco Live able to get
(01:02):
some time with Carl and we'regoing to really talk about all
things that's happening inindustry, what Carl's seeing
across customers and the region,as well as all things that
Cisco's pushing forward in.
Carl, welcome, thank you forhaving me.
I really appreciate the time,as I said, but before we get
started, for those who don'tknow you, who are outside of the
circles, who listen into this,do you mind just spending a
little bit of time about yourcareer journey and what led you
(01:25):
to become the CTO of AustralianMuseum for Cisco?
Sure.
Carl Solder (01:28):
Graduate of Curtin University which was formerly
West Australian Institute ofTechnology.
Back in the day, Computerprogramming as it was known then
, now software engineering.
11 years in IBM and coming up28 years in Cisco.
I started in Cisco in themid-1990s when Cisco was setting
up sales offices aroundAustralia, I was the second
(01:49):
engineer in Perth to start.
I started off in sales as asystems engineer In 2002, I had
the opportunity to get hired bythe engineering teams out of the
US.
So I spent 19 years working inengineering.
I started off working on whatwas, I think now, arguably the
most successful product Cisco'sbrought to market, which was the
(02:10):
Catalyst 6500 switch.
Fantastic device, billions andmillions of dollars of revenue.
But that was my stompinggrounds in the early days.
But I started off working overthere in the engineering team
and eventually ended up, justbefore I left, leading a team of
about nearly 200 engineersworking in the enterprise
networking business.
So we were responsible forCatalyst 9,000 switches,
(02:32):
catalyst 8,000 routers, theCatalyst Enterprise wireless
portfolio, dnac as it was then,but now Catalyst Center, the
Viptela acquisition which is nowCatalyst SD-WAN Identity
Services Engine, so a whole hostof kind of mainstream
enterprise products.
So we were responsible forproduct strategy and technical
(02:53):
strategy defining what we weregoing to build, what we were
going to create next to servicethe needs of customers in the
market and then figuring out howto build as best a product as
we could.
But I had a bit of a life resetin the pandemic and decided that
it was time to come back toAustralia.
So I actually decided to retire.
I left Cisco, hung up my boots,moved back and a couple of
(03:15):
months in I got a call from theCisco country manager, ben
Dawson.
He said, hey, how would youlike to come back?
So they opened up a role as CTO.
He said, hey, how would youlike to come back?
So they opened up a role as CTOand I joined just short of
three years ago, february 2022.
Michael van Rooyen (03:36):
And I've been serving as the chief
technology officer for CiscoAustralia, new Zealand ever
since.
Wow, that's a fantastic careerand I think about just at the
heart of everything you've donethere.
It's really digital plumbing,right.
I mean, if I think about thoselife cycles, I mean that was
Cisco's bread and butter andyou've lived that whole life
cycle.
It is A bit of a pause, whichwas probably fortuitous, because
today, if you think about whereCisco's life cycle is going,
you know, if I really reflectwhere you guys are today you're
(03:57):
really back at that early 90s,late 90s, kind of that's the
feeling right there.
There's a lots happened, ofcourse, and we've got ai, which
no doubt we'll touch on lots ofconnectivity.
You know the world's continuedto be hyper connected, etc.
So it's probably a good time toto be a digital plumber as, so
to speak and and uh, you know,our industry's fascinating never
stops changing, right itdoesn't.
Carl Solder (04:17):
That's one thing that I have learned in my entire
career.
Change is the constant, soevery time you think you're just
getting comfortable with thelatest bit of tech, a couple of
weeks later you turn your headand there's something new you
have to learn about.
So you're absolutely right.
Michael van Rooyen (04:31):
And that just continues to evolve so much
.
What do you mean by that is?
We used to be able to pickthrough these tranches of you
know switching, and then we didwireless and then voice and you
know kind of map those TDM to IPtelephony.
You know all these phases buttoday, as you just said, there's
so many that are happening soquickly.
It can be two or three at thesame time.
Carl Solder (04:47):
It's faster, yes.
So that's the other observationI've made.
Just the pace of innovation,not only from Cisco, but just
the industry overall, is movingat a very fast pace, and I think
one of the biggest challengesthat most customers I speak to
is how can we consume this, howcan we take advantage of it?
And just being first of all,being aware, yes.
Once you're aware, then how doyou consume, yes, and how can
(05:08):
you actually use this to drivebetter productivity, better
experience for customers, betterprofitability for the business
they're all questions thatcustomers are always asking.
Michael van Rooyen (05:17):
Yeah, yeah, and because we are catching up
at Cisco Live and I know thatyou're involved globally with
your teams, et cetera, and spenda lot of time talking to the US
and the innovation and youlived, as you said in the intro,
around driving innovation.
While we're here at Cisco LiveMelbourne today, what are some
of the most excitingannouncements or some of the
innovations that Cisco's kind ofintroducing?
(05:38):
You guys normally release a fewthings at Cisco Live in the US
and then trickle through to herewe do.
We do.
Carl Solder (05:43):
Look, there's actually quite a lot to wrap
your arms around.
Yes, so I think Cisco's focusand you saw, if you saw the
keynote there were three areasof focus AI-ready data centers,
future-proof workspace anddigital resilience and they kind
of sit in the front of a wholehost of innovations that we're
(06:03):
bringing to market.
And they kind of sit in thefront of a whole host of
innovations that we're bringingto market In the background, in
my old world, when I tend tohave a mindset more from an
engineering side, I think aboutthe developments, the products
we're bringing to market.
We are reimagining our securityportfolio.
Yes, so we announced CiscoSecurity Cloud a little while
ago.
We've got a whole host ofinnovations related to improving
security posture fororganizations.
(06:24):
Hypershield is probably the bigone that we announced this year
.
That represents a brand new wayin which you can protect your
applications Identityintelligence all about
protecting you against sessionhijacking.
That's another one We'veintroduced.
In addition to that, on thecollaboration side, you might
have seen new collaborationdevices, new ceiling microphones
(06:47):
, room designers.
We're recognizing that in thisworld of hybrid work.
We actually published a hybridwork report earlier this year
and we, I think, polled some14,000 employees, a huge number
of employers, but we saw someinteresting statistics 80% of
those organizations were lookingto either mandate a full or
(07:10):
partial return to the office yes, but 50% of those organizations
didn't have their workspace setup to really accommodate the
kind of cubes.
Yes, I hate the cubes, and whatwe're seeing, certainly in the
US, and I think certainly we'restarting to see this here in
Australia is the reimaginationof the workspace, and so when
you do that, you start buildingmore collaborative areas, but in
(07:32):
tandem with that, you also needto have the right tech to be
able to support that.
Yes, and so that's a secondarea of innovation that Cisco is
also working on.
We're also developing newinnovations on the cloud side,
and that's kind of driven partlyby the continued adoption of
cloud and also the adoption ofAI.
So we bring in a market podsthat you can run your custom
(07:57):
workloads on.
We're doing things in thevisibility space with Thousand
Eyes, for example.
We recognize that between theuser and the cloud, where the
application is, there's manytouch points that potentially
could be problematic, and,especially with businesses
digitizing their front ends,that demand to drive as close to
100% availability as possiblemeans that when something goes
(08:18):
blip, something goes wrong.
It's all action stations andwhat can we do to identify where
the problem is?
If you think about a user ringsup and says I can't access my
application as a help deskoperator, is it a problem with
the device the person's using?
Is it a problem with the Wi-Fi?
Is it a problem in the LAN?
Is it a problem in the WAN withmy ISP?
Is it a problem in the cloud ofmy application?
Where do you start looking?
So having the tool sets to beable to zero in and give that
(08:42):
help desk operator a head startin identifying the problem so
they can remediate it.
They're all the innovationsthat we bring into market, and
Splunk is part of that.
Data is going to be a keymetric in that conversation.
The more data you have, themore you can see.
The more you can see, the moreyou can action.
So there's a method behindCisco's madness of acquiring
Splunk, but that's all going tobe fed into the observability
(09:04):
tool.
So madness of acquiring Splunk,yes, but that's all going to be
fed into the observability tool.
So data represents a reallyimportant part of our strategy
moving forward.
So there's a lot going on, as Isaid.
Security we've got the new AIassistants that are coming out.
That represents a brand newparadigm for how an operator
goes about enacting workflows.
If you think, in the past itused to be CLI.
I love my CLI and, in fact, allthe customers still like their
(09:27):
CLI.
They find it hard to extractthemselves from that.
But we've moved to the notionof controls and dashboards using
beautiful GUIs to simplifyworkflows.
But now you're getting a chatGPT style interface where you're
initiating prompts to enactworkflows and not having to
click on any buttons at all.
That's a brand new paradigmthat I think customers are going
(09:48):
to start seeing more and more,not only from Cisco, but from
everyone else in the industry.
Michael van Rooyen (09:54):
Yep.
Carl Solder (09:54):
So there's just so much going on right now.
It's exciting, but it's also alittle bit overwhelming.
Michael van Rooyen (09:59):
It would be.
Yeah, look, I mean that's a lotto digest and also the dynamics
of, probably, the new engineers.
We touched on CLI, a group ofCLI.
You know showing our ages andyou know cloud management.
You know all these greatbenefits.
But with all this AI investmentand all this AI drive, you know
we're going to have newengineers that are going to just
expect that to be the caseright, and the simplified and
(10:22):
asking the question.
I guess a side note would be Ithink we're going to have a
potential gap in troubleshooting, you know, because you have
this AI automation, which isfantastic.
But I think that the lack ofunderstanding the root cause of
the problem might be an outcomeof these changes.
But it is super exciting, right.
New product, lots of innovation.
You touched on collaboration.
You know, return to work's abig one, right.
Carl Solder (10:42):
Even just the plumbing to support that has had
a refresh because no one spentmoney on it, because everyone
was at home, right, yes, weactually have a what we like to
think it's a work place templatefor the future Our sales office
in Midtown Matten in New York,affectionately called One Pen
Plaza.
We went through a massiveretrofit in that floor space and
(11:04):
we converted the floor spacefrom being 70% cubes and 30%
collaborative space.
We flipped it to 70%collaboration and 30% cubes.
We actually seen a 40% increasein the number of employees we
can actually get back in theoffice and during the pandemic
everyone kind of went home.
But we've seen a 300% increasein employees coming back to the
(11:25):
office because of theenvironment and the workspace
that we've built, which is veryconducive and alluring to
actually wanting to come backand doing work.
But it's founded a whole bunchof collaboration technology but
there's also a lot of the actualendpoints that you have in the
building, which is drivinganother area of interest for
(11:46):
customers around net zero andsustainability and energy
consumption, or reducing energyconsumption and reducing
emissions where we areconnecting lights, reducing
emissions where we areconnecting lights, cameras,
audio, window shades, hot desks,usb ports, you plug your
laptops in and they're allpowered by PoE.
And when we power it via PoE wehave an ability to understand
(12:11):
energy consumption and we canmodulate and control that.
So, like, if users leave a room, we can detect through the
network that someone's left theroom, we can turn the lights off
.
If they didn't turn the lightsoff, we can lower the air
conditioning down, we can dropwindow shades.
So we're moving towards thisnotion of smart buildings, yes,
or programmable buildings, yes.
That's another area that we'rekind of looking at, but the
(12:31):
collaboration technology we'rebuilding marries up with that
very, very nicely.
Michael van Rooyen (12:35):
Wow, wow and I just think about that.
That's a sustainabilitydiscussion, right, and it's
driven by technology.
One of the key things I'd loveyou to touch on is knowing that
you spend so much time withcustomers of all varying sizes
and different industries, and Iknow that your passion around
innovation and innovating andsimplifying complex technologies
, because we've grown up with alot of complexity and we're
(12:57):
obviously having this nicetrajectory down towards
simplification, with a lot ofsmart technology behind the
scenes.
But how does your role allowyou to, you know, have a bit of
impact in the direction of Ciscoand what sort of conversations
you're having with customers?
What are the key things?
Carl Solder (13:12):
So I'll answer that in two ways.
Sure, one part in terms of therole I did in the US and how I
take advantage of that insightin my current role.
Michael van Rooyen (13:23):
Yes.
Carl Solder (13:23):
So the role I had in the US, as I said, was
focused on product and technicalstrategy.
Product strategy is ultimatelydeciding what we're going to be
building next.
When we decide, or in decidingwhat we're going to be building
next, there's a lot of inputs.
There's, for example, what ourcompetitors are doing, what are
new research projects that aregoing on that might have a
bearing on what we want to do,asks from customers, asks from
(13:45):
partners, research that we'redoing internally ourselves on
trying to solve problems for ourcustomers.
So we take all these inputs toformulate a product strategy.
Now, part of what we did was wealso solicit input from our
field.
So systems engineers, accountmanagers, product sales
specialists we do poll thoseindividuals and we get some
(14:07):
input from them.
Now, in my current role, knowingwhat I know about the way the
engineering engine works, andbecause I actually know a lot of
the individuals who aredeciding and building that
product strategy a part of myrole I like to think of myself
as the voice of the Australian,new Zealand customer.
So in my role, I do get totravel a lot, I do get to meet
(14:30):
with a lot of customers, a lotof partners, and I probably
unlike a lot of my peers whowork on segments of the market,
like I might be a healthcaresales specialist, I might be
working on the enterprise or thecommercial or small business.
I look at everything.
So I build up a uniqueperspective of, I guess, inputs
that I hear from customers and Ichannel that back to my former
(14:53):
colleagues in the US my formercolleagues in the US and I'm
pretty forceful around thatbecause I like to champion what
I think is going to be neededfor the Australian customers and
I work with the engineeringteams to make sure that the
voice of customers here isactually heard and is factored
into that product strategy.
Michael van Rooyen (15:11):
Yeah, and with those discussions is there
a common theme and I knowthere's probably so many angles
because of the criticality ofall our connectivity and
cybernet but is there one kindof key discussion that's really
dominating your currentdiscussions?
Look at a high level.
Carl Solder (15:26):
There's a couple of things that I think customers'
interests really at the momentartificial intelligence,
cybersecurity and the.
I guess how to drive moreefficiencies, operational
efficiency, reduce operatorburden, do more with less that's
kind of a common question, butultimately it drives down to how
can I make the network simpler?
(15:48):
Yes, simplicity counts and, Ithink, something that sometimes
gets lost in translation.
Engineers get excited.
They have all these wonderfulideas and they go off and build
all these wonderful things andsometimes we tend to forget the
basics.
We kind of work on everything.
But.
And for customers?
(16:09):
Yes, all these new innovationsare quite exciting and they
absolutely address solvingproblems in the business.
But for the IT practitionerit's how can I get things done
more quickly?
If a problem pops up, how can Iresolve that more quickly?
If I'm going to maybe upgradesome software or I'm going to
apply a patch or I'm going to dosome configuration, how can I
get it done in a more timelymanner without all the fuss?
(16:31):
That's ultimately whatcustomers are all about and if
you think, historically, we'vegone through software-defined
networking that had the promiseof simplifying networks, cloud
computing, taking away the needto have to worry about compute
infrastructure as a serviceplatform as a service, software
as a service, takinginfrastructure away.
I don't have to worry aboutracking and stacking and cabling
(16:52):
and upgrading software.
Those things were all attempts,but even though we've got all
of that in play, customers stillsay it's still complex.
Yes, so there's stillopportunities for companies like
Cisco to think about how we cancontinue to innovate, to drive
further simplicity.
That's ultimately, I think whatwe've started to recognize, that
(17:13):
if we can get those basicsright, then we're going to make
our customers happy.
Michael van Rooyen (17:17):
Yeah, and look for me having a look at
Partner Summit and also CiscoLive today and the key messaging
.
Cisco went through thistransition of incredible
engineering, some of the stuffyou're involved in 6500, you
touched on an amazing productall the way through to where we
are today.
To where we are today.
(17:38):
Probably the biggest feelingI'm getting out of Cisco is this
whole idea of this end-to-endsolutioning better together real
drive and security, how webring all these components
together so that you have thisbetter ecosystem and someone
kind of related to otherproducts where they work better
together.
Can you just touch on that,because I'm really picking that
as a big thing as opposed tothese silos that everyone used
to deal with Yep.
Carl Solder (17:54):
So I think the silo thing is something that's
manifested itself out across theindustry.
You look at multi-vendorenvironments.
You have the best firewall, youhave the best phone system, you
have the best switch, you havethe best access point, and when
you patch it together it becomesmore challenging.
I think, one thing that Iobserved in the US maybe
(18:16):
customers saw the realization abit ahead of where Australian
customers have seen this is thedifference between OPEX and
CAPEX.
So if you think back to why dowe build silos?
Because it was focused on whobuilt the biggest, the best, the
cheapest, and so you ended upbuilding all these best of breed
(18:37):
individual products, but as youstart to stitch them together
to your point, the OPEX sidebecomes much larger.
Now in the US we had this.
I think a lot of customers Ispoke to had this rule of thumb
that for every dollar you spendon CapEx, you spend $10 on OPEX.
So the question we would poseto them is if we can save you
(18:57):
20% on CapEx versus 20% on OPEX,what would you prefer?
And it's absolutely OPEX, ofcourse.
So, while you might spend alittle bit more on individual
products, but if they worktogether better as a system and
you can save more money on theOPEX side, that's what customers
want.
So that's this notion and thisdirection to move towards
platforms.
(19:17):
Yep, and certainly Cisco hasrecognized that.
And while we've builtbest-of-breed vertical stacks,
like you look at, for example,if you're looking at a cloud
stack, you might have Nexus.
You're looking at a WAN stack.
You might have the CatalystSD-WAN.
These are discrete stacks fromthe controller to the devices,
the OS it kind of isverticalized.
(19:37):
But can we build something thathas better horizontal
integration?
That's what we're starting todo.
We've probably taken the leadthere on the security side with
the security cloud, and we'vealso got an initiative and a
strategy on the networking sideknown as the Cisco Networking
Cloud, where we're going tostart doing that horizontal
integration drive betterplatform integration.
(19:58):
Again, focus on reducingoperator burden and improvements
in operational simplicity.
Michael van Rooyen (20:04):
Yeah, yeah, and look, I think it's a
fantastic journey and if I thinkabout the breadth of products
that the portfolio has, bringingthat all together, it really is
that amazing ecosystem.
And whilst historically peoplehad you had two camps, it was
opposed to end-to-end singlevendor best of breed.
You know, let's choose the bestfile.
In fact, we even you, would befamiliar with architectures
(20:24):
where we even have doublefirewalls as an example.
But it'd be two vendors to dothe same thing and that has lots
of challenges alone.
But I think the world isgetting more comfortable in
relation to having it all underone umbrella, having it all
under one banner, one managementplatform.
It's an incredible long journey.
It's been a vision for a while,but I think it's nice to see it
coming to fruition.
Now it does.
Carl Solder (20:44):
And it comes back to that CapEx OpEx thing.
You might have to spend alittle bit more money and it
might be that you don't get thebest of breed of everything.
But if you get betterintegration end to end, that's
going to yield a better outcomefor operators, and I use apple
as probably the best example inthe industry yes, you know, my
family has apple everythingexcept for my son, who has an
android phone.
So no facetime.
(21:06):
Yes, you know, no iMessage.
There's things that, like, youdon't quite get, and so it
becomes just that little bitmore difficult, but when you're
inside that ecosystem,everything's seamless, of course
, and so that's, I think, whatcustomers are starting to
recognize that if they stick toan ecosystem, they can drive
better outcomes.
Michael van Rooyen (21:25):
Yeah, 100%.
In the conversation a little bityou touched on AI and we'll
have to talk about it becauseit's so topical and your
customers are talking about it,and 18 months ago it would have
been a different conversationwe're probably having today, and
you talked about a lot of theintegration that the team's
doing around AI bots and AItools to simplify the operation,
as you said.
But can you just talk a littlebit more about?
(21:45):
Earlier you touched on one ofthe key pillars is AI data
centers, or building for AI datacenters.
Do you mind that's a reallyinteresting area for us?
I'm keen to hear about that.
Carl Solder (21:54):
So Cisco, its AI strategy is really centered
around a few aspects.
There is the infrastructurepiece.
So that is compute, that'sstorage, that's networking.
It's what are we doing to buildthe infrastructure that's going
to allow customers to run theircustom AI workloads.
So there, you're talking aboutcompute platforms with heavily
loaded with GPUs.
You're talking about highbandwidth, high speed Ethernet
(22:18):
400 gig, 800 gig.
Moving to the future, where itwill be terribly expensive, plus
, yes, and getting thatinfrastructure right so that
consumer customers can actuallyrun their our workloads.
So we're heavily focused onthat.
It's centered around our siliconone technology, silicon one is
foundational to that partneringwith industry vendors like amd,
(22:39):
like nvidia, like intel.
Those organizations are goingto have technologies that will
integrate into thatinfrastructure.
Yes, to bring that to life.
The other areas that we're alsolooking at is the embedding the
ai in our solutions.
So you're off the shelfsolution you take advantage of.
Show us a I don't know marketdashboard, for example.
It has radio resourcemanagement in the back, ai
(23:02):
monitoring, adjusting, finetuning your wi-fi settings
without the operator having todo anything, and then you have
the new emerging AI assistance,which is changing the
operational paradigm.
So while we kind of front andcenter and talk about AI-ready
data centers, that is a heavyfocus on the AI infrastructure
that we're building.
I also, as a CTO, I try to makesure customers don't forget
(23:25):
that there's also other aspectsto our AI journey that can also
benefit them as well.
Yeah, sure, sure.
Michael van Rooyen (23:30):
I guess you're in a good position there,
right?
People are going to need lotsof infrastructure for this and
as well as, how do you grappleit internally to make sure
people are getting the best outof the products?
And I think about the data setsthat you have available.
You know you've got suchnetwork, cyber, collaboration,
all this endpoint, thethousandized data.
It's just incredible how wecould use those data sets to
(23:51):
really drive the outcome.
What I'd be keen to get fromyour point of view as well, as
we talk a lot and I'll lead to asecond section to this question
, but we're seeing a lot ofcustomers needing to go back to
having workloads remote or closeto the edge.
That probably leads a bit to IoT, which we'll touch on in a
minute, but I'm also seeing thatmaybe a little bit in the AI,
not a full workload.
We need massive data centers,lots of power.
Are you seeing the same sort oftrajectory?
(24:11):
A little bit of compute wasgoing to the cloud, but now,
coming back a bit, we are.
We are.
Carl Solder (24:15):
So actually I was part of a set of CR roundtables
early this year country managerfrom idc who was there guest
speaking and they had mentionedsome research that they did.
It was posted middle of theyear.
Based on their survey, 80 oforganizations they polled said
that they'd be repatriating someor most of their workloads and
(24:38):
data.
But back on prem right right andit's actually I think I don't
know it's necessarily going togo from 100 to 0.
I kind of talk about it asrebalancing there's for a
multitude of reasons there'scompliance, there's sovereignty,
there's latency, there's cost,there's skills, there's security
, and I've spoken to quite a fewcustomers here in Australia who
are quote repatriating, and oneof those areas is AI.
(25:03):
So AI obviously very computeintensive.
If it's compute intensive, youknow that it's going to cost a
bit of money.
I actually refer customers toan interesting article that got
posted, probably about threeyears ago, by Andresen Horowitz,
which is one of the biggest VCson the planet.
And it was authored by a guycalled Martin Casado.
(25:25):
Martin Casado was a PhDdstudent at stanford.
He invented sdn softwaredefined networking.
He also invented openflow right.
He created a startup callednysera, got bought by vmware for
over a billion dollars and thatended up becoming what is now
known as nsx.
But he works for andres andhorowitz and he did he did this
study on cost of the cloud heactually his article the
(25:47):
Trillion Dollar Cloud Paradox,but he actually has an
interesting quote which I callout to customers and I normally
get a smile, and the quote isyou're crazy if you don't start
in the cloud, but you're crazyif you stay on it.
Right, but you're crazy if youstay on it, right.
And his argument is that theflexibility and the elasticity
(26:11):
of cloud and the cost metrics isreally great when you're
starting off.
Yes, but for many organizationsthat they've spoken to and in
fact it kind of holds true withsome customers I've spoken to as
well that the cost benefitanalysis that they did right at
the beginning of their cloudjourney doesn't necessarily
marry up to the actual coststhat they're bearing right now,
(26:33):
right.
So there is that desire from ITleaders to find ways to trim or
refine their cloud costs, andwe've had discussions with some
customers who've done kind of anew cost benefit analysis and
said, hey, actually it will becheaper for us to run it on-prem
than it will be for us tocontinue running the cloud.
So there's some things thatwork well in the cloud.
(26:54):
Yes, there's also someapplications and workloads that
might actually be better to runon-prem, and ai is one of those.
It's a great, maybe start offyour ai journey, but if you're
going to be serious aboutrunning AI, then maybe there's
an opportunity to considerrunning it on-prem.
And we've started to see moreso globally not so much in
(27:14):
Australia as organizations arenow starting to dip their toes
into actually building andputting into production
AI-enabled workflows.
A good percentage of those arenow starting to see that being
run locally or outside of in aprivate cloud as opposed to
being in the public cloud.
So I think over the comingmonths and years it's going to
(27:39):
be interesting to see where theindustry and customers' heads
are going to be at.
But there's definitely amomentum in some parts of the
customer base where they'rereassessing where they're going
to be running.
Michael van Rooyen (27:48):
Yeah, look, I completely agree and I see
that as well.
I think I see it in a couple ofdimensions.
First of all, some customersand I really like your thing
about starting in the cloud andnot necessarily stay there.
I think we're going to end upwith hybrid AI if that's a new
term, maybe, but effectively I'malso seeing customers' mindsets
say all these tools that havebeen developed that are in cloud
and you can ask questions andyou get these fantastic results
and you can do drawings andvideos and all that great stuff,
(28:11):
right.
I'm also seeing customerssaying well, we need to
understand that we need to buildour own LLMs and of our own
data, and the perception is thatthat should be on-prem, because
I need to know where itcontained it and own it and
everything whilst I'm on thecloud.
So I think there's thatdimension on the physical aspect
.
Second part is that you're right.
Not all workloads need to be inthe cloud to do that right.
(28:31):
You don't need a mass scale.
So much compute and performanceis happening at the edge, you
can probably get the rightresults for businesses at the
closest or far edge.
And then last dimension to thatis the IoT play right.
How do we ingest data to that?
Iot is going to be one of thebig feeds into AI models.
So we're seeing a lot ofcustomers look at that.
Carl Solder (28:49):
I think there's something else which is of
interest, and that's the generalmove towards building smaller
language models.
Yes, the small language models,the micro language models, the
domain specific language models,the ability to potentially run
a model on a laptop or a tablet.
Yes, that opens up a whole rangeof opportunities that you might
not be able to necessarilybuild with a more generic large
(29:12):
language model.
So I think there's so much.
I actually was talking tosomeone recently and I kind of
think of AI like the iPhone.
When the iPhone came out in2007, it was a very different
device then to what it is today100% and I don't think back then
people fully appreciated thepotential of what that could do.
And when you think back thenyou couldn't do online banking,
(29:33):
you couldn't order your pizza,you couldn't check where the bus
was, couldn't do your bankingthose use cases kind of got
derived on that journey.
And I think the same thing withAI.
We're at that moment where weknow there's a lot of potential,
but I don't think we trulyrealize exactly what it
potentially can do, and I thinkthat will be discovered.
Michael van Rooyen (29:54):
I hope you enjoyed part one of my
discussion with Carl Solder.
Cisco ANZ CTO.
Tune in next week for part twoof the special Cisco Live
edition of Securely ConnectedEverything.
Welcome to a special Cisco Live edition of
Securely Connected Everything.
I was fortunate enough to spenda better part of a week with
Cisco for the annual APJC CiscoLive event held in Melbourne
each year.
During that week, I had theopportunity to catch up with a
lot of interesting people fromCisco, of course, but also from
across the technology industry.
The great news for listeners ofthe podcast is that I got to
(00:20):
record many of thoseconversations at our Oro Live
podcasting booth, which we hadset up on site throughout Cisco
Live.
You'll get to hear a number ofthose discussions next season
when we kick off in late January2025.
I couldn't wait to Len to bringin today's guest, though, so
we're bringing him to you now asa special two-part Cisco Live
edition of Securely ConnectedEverything.
(00:41):
Without further ado, here ispart one of my discussion with
Cisco and ZCTO Carl Solder.
Today I have the pleasure inhaving a discussion with Carl
Solder, who is the CTO ofAustralia, new Zealand, for
Cisco Systems.
He's been at Cisco for a longtime and he'll certainly cover
that in our discussion, butwe're at Cisco Live able to get
(01:02):
some time with Carl and we'regoing to really talk about all
things that's happening inindustry, what Carl's seeing
across customers and the region,as well as all things that
Cisco's pushing forward in.
Carl, welcome, thank you forhaving me.
I really appreciate the time,as I said, but before we get
started, for those who don'tknow you, who are outside of the
circles, who listen into this,do you mind just spending a
little bit of time about yourcareer journey and what led you
(01:25):
to become the CTO of AustralianMuseum for Cisco?
Sure.
Carl Solder (01:28):
Graduate of Curtin University which was formerly
West Australian Institute ofTechnology.
Back in the day, Computerprogramming as it was known then
, now software engineering.
11 years in IBM and coming up28 years in Cisco.
I started in Cisco in themid-1990s when Cisco was setting
up sales offices aroundAustralia, I was the second
(01:49):
engineer in Perth to start.
I started off in sales as asystems engineer In 2002, I had
the opportunity to get hired bythe engineering teams out of the
US.
So I spent 19 years working inengineering.
I started off working on whatwas, I think now, arguably the
most successful product Cisco'sbrought to market, which was the
(02:10):
Catalyst 6500 switch.
Fantastic device, billions andmillions of dollars of revenue.
But that was my stompinggrounds in the early days.
But I started off working overthere in the engineering team
and eventually ended up, justbefore I left, leading a team of
about nearly 200 engineersworking in the enterprise
networking business.
So we were responsible forCatalyst 9,000 switches,
(02:32):
catalyst 8,000 routers, theCatalyst Enterprise wireless
portfolio, dnac as it was then,but now Catalyst Center, the
Viptela acquisition which is nowCatalyst SD-WAN Identity
Services Engine, so a whole hostof kind of mainstream
enterprise products.
So we were responsible forproduct strategy and technical
(02:53):
strategy defining what we weregoing to build, what we were
going to create next to servicethe needs of customers in the
market and then figuring out howto build as best a product as
we could.
But I had a bit of a life resetin the pandemic and decided that
it was time to come back toAustralia.
So I actually decided to retire.
I left Cisco, hung up my boots,moved back and a couple of
(03:15):
months in I got a call from theCisco country manager, ben
Dawson.
He said, hey, how would youlike to come back?
So they opened up a role as CTO.
He said, hey, how would youlike to come back?
So they opened up a role as CTOand I joined just short of
three years ago, february 2022.
Michael van Rooyen (03:36):
And I've been serving as the chief
technology officer for CiscoAustralia, new Zealand ever
since.
Wow, that's a fantastic careerand I think about just at the
heart of everything you've donethere.
It's really digital plumbing,right.
I mean, if I think about thoselife cycles, I mean that was
Cisco's bread and butter andyou've lived that whole life
cycle.
It is A bit of a pause, whichwas probably fortuitous, because
today, if you think about whereCisco's life cycle is going,
you know, if I really reflectwhere you guys are today you're
(03:57):
really back at that early 90s,late 90s, kind of that's the
feeling right there.
There's a lots happened, ofcourse, and we've got ai, which
no doubt we'll touch on lots ofconnectivity.
You know the world's continuedto be hyper connected, etc.
So it's probably a good time toto be a digital plumber as, so
to speak and and uh, you know,our industry's fascinating never
stops changing, right itdoesn't.
Carl Solder (04:17):
That's one thing that I have learned in my entire
career.
Change is the constant, soevery time you think you're just
getting comfortable with thelatest bit of tech, a couple of
weeks later you turn your headand there's something new you
have to learn about.
So you're absolutely right.
Michael van Rooyen (04:31):
And that just continues to evolve so much
.
What do you mean by that is?
We used to be able to pickthrough these tranches of you
know switching, and then we didwireless and then voice and you
know kind of map those TDM to IPtelephony.
You know all these phases buttoday, as you just said, there's
so many that are happening soquickly.
It can be two or three at thesame time.
Carl Solder (04:47):
It's faster, yes.
So that's the other observationI've made.
Just the pace of innovation,not only from Cisco, but just
the industry overall, is movingat a very fast pace, and I think
one of the biggest challengesthat most customers I speak to
is how can we consume this, howcan we take advantage of it?
And just being first of all,being aware, yes.
Once you're aware, then how doyou consume, yes, and how can
(05:08):
you actually use this to drivebetter productivity, better
experience for customers, betterprofitability for the business
they're all questions thatcustomers are always asking.
Michael van Rooyen (05:17):
Yeah, yeah, and because we are catching up
at Cisco Live and I know thatyou're involved globally with
your teams, et cetera, and spenda lot of time talking to the US
and the innovation and youlived, as you said in the intro,
around driving innovation.
While we're here at Cisco LiveMelbourne today, what are some
of the most excitingannouncements or some of the
innovations that Cisco's kind ofintroducing?
(05:38):
You guys normally release a fewthings at Cisco Live in the US
and then trickle through to herewe do.
We do.
Carl Solder (05:43):
Look, there's actually quite a lot to wrap
your arms around.
Yes, so I think Cisco's focusand you saw, if you saw the
keynote there were three areasof focus AI-ready data centers,
future-proof workspace anddigital resilience and they kind
of sit in the front of a wholehost of innovations that we're
(06:03):
bringing to market.
And they kind of sit in thefront of a whole host of
innovations that we're bringingto market In the background, in
my old world, when I tend tohave a mindset more from an
engineering side, I think aboutthe developments, the products
we're bringing to market.
We are reimagining our securityportfolio.
Yes, so we announced CiscoSecurity Cloud a little while
ago.
We've got a whole host ofinnovations related to improving
security posture fororganizations.
(06:24):
Hypershield is probably the bigone that we announced this year
.
That represents a brand new wayin which you can protect your
applications Identityintelligence all about
protecting you against sessionhijacking.
That's another one We'veintroduced.
In addition to that, on thecollaboration side, you might
have seen new collaborationdevices, new ceiling microphones
(06:47):
, room designers.
We're recognizing that in thisworld of hybrid work.
We actually published a hybridwork report earlier this year
and we, I think, polled some14,000 employees, a huge number
of employers, but we saw someinteresting statistics 80% of
those organizations were lookingto either mandate a full or
(07:10):
partial return to the office yes, but 50% of those organizations
didn't have their workspace setup to really accommodate the
kind of cubes.
Yes, I hate the cubes, and whatwe're seeing, certainly in the
US, and I think certainly we'restarting to see this here in
Australia is the reimaginationof the workspace, and so when
you do that, you start buildingmore collaborative areas, but in
(07:32):
tandem with that, you also needto have the right tech to be
able to support that.
Yes, and so that's a secondarea of innovation that Cisco is
also working on.
We're also developing newinnovations on the cloud side,
and that's kind of driven partlyby the continued adoption of
cloud and also the adoption ofAI.
So we bring in a market podsthat you can run your custom
(07:57):
workloads on.
We're doing things in thevisibility space with Thousand
Eyes, for example.
We recognize that between theuser and the cloud, where the
application is, there's manytouch points that potentially
could be problematic, and,especially with businesses
digitizing their front ends,that demand to drive as close to
100% availability as possiblemeans that when something goes
(08:18):
blip, something goes wrong.
It's all action stations andwhat can we do to identify where
the problem is?
If you think about a user ringsup and says I can't access my
application as a help deskoperator, is it a problem with
the device the person's using?
Is it a problem with the Wi-Fi?
Is it a problem in the LAN?
Is it a problem in the WAN withmy ISP?
Is it a problem in the cloud ofmy application?
Where do you start looking?
So having the tool sets to beable to zero in and give that
(08:42):
help desk operator a head startin identifying the problem so
they can remediate it.
They're all the innovationsthat we bring into market, and
Splunk is part of that.
Data is going to be a keymetric in that conversation.
The more data you have, themore you can see.
The more you can see, the moreyou can action.
So there's a method behindCisco's madness of acquiring
Splunk, but that's all going tobe fed into the observability
(09:04):
tool.
So madness of acquiring Splunk,yes, but that's all going to be
fed into the observability tool.
So data represents a reallyimportant part of our strategy
moving forward.
So there's a lot going on, as Isaid.
Security we've got the new AIassistants that are coming out.
That represents a brand newparadigm for how an operator
goes about enacting workflows.
If you think, in the past itused to be CLI.
I love my CLI and, in fact, allthe customers still like their
(09:27):
CLI.
They find it hard to extractthemselves from that.
But we've moved to the notionof controls and dashboards using
beautiful GUIs to simplifyworkflows.
But now you're getting a chatGPT style interface where you're
initiating prompts to enactworkflows and not having to
click on any buttons at all.
That's a brand new paradigmthat I think customers are going
(09:48):
to start seeing more and more,not only from Cisco, but from
everyone else in the industry.
Michael van Rooyen (09:54):
Yep.
Carl Solder (09:54):
So there's just so much going on right now.
It's exciting, but it's also alittle bit overwhelming.
Michael van Rooyen (09:59):
It would be.
Yeah, look, I mean that's a lotto digest and also the dynamics
of, probably, the new engineers.
We touched on CLI, a group ofCLI.
You know showing our ages andyou know cloud management.
You know all these greatbenefits.
But with all this AI investmentand all this AI drive, you know
we're going to have newengineers that are going to just
expect that to be the caseright, and the simplified and
(10:22):
asking the question.
I guess a side note would be Ithink we're going to have a
potential gap in troubleshooting, you know, because you have
this AI automation, which isfantastic.
But I think that the lack ofunderstanding the root cause of
the problem might be an outcomeof these changes.
But it is super exciting, right.
New product, lots of innovation.
You touched on collaboration.
You know, return to work's abig one, right.
Carl Solder (10:42):
Even just the plumbing to support that has had
a refresh because no one spentmoney on it, because everyone
was at home, right, yes, weactually have a what we like to
think it's a work place templatefor the future Our sales office
in Midtown Matten in New York,affectionately called One Pen
Plaza.
We went through a massiveretrofit in that floor space and
(11:04):
we converted the floor spacefrom being 70% cubes and 30%
collaborative space.
We flipped it to 70%collaboration and 30% cubes.
We actually seen a 40% increasein the number of employees we
can actually get back in theoffice and during the pandemic
everyone kind of went home.
But we've seen a 300% increasein employees coming back to the
(11:25):
office because of theenvironment and the workspace
that we've built, which is veryconducive and alluring to
actually wanting to come backand doing work.
But it's founded a whole bunchof collaboration technology but
there's also a lot of the actualendpoints that you have in the
building, which is drivinganother area of interest for
(11:46):
customers around net zero andsustainability and energy
consumption, or reducing energyconsumption and reducing
emissions where we areconnecting lights, reducing
emissions where we areconnecting lights, cameras,
audio, window shades, hot desks,usb ports, you plug your
laptops in and they're allpowered by PoE.
And when we power it via PoE wehave an ability to understand
(12:11):
energy consumption and we canmodulate and control that.
So, like, if users leave a room, we can detect through the
network that someone's left theroom, we can turn the lights off
.
If they didn't turn the lightsoff, we can lower the air
conditioning down, we can dropwindow shades.
So we're moving towards thisnotion of smart buildings, yes,
or programmable buildings, yes.
That's another area that we'rekind of looking at, but the
(12:31):
collaboration technology we'rebuilding marries up with that
very, very nicely.
Michael van Rooyen (12:35):
Wow, wow and I just think about that.
That's a sustainabilitydiscussion, right, and it's
driven by technology.
One of the key things I'd loveyou to touch on is knowing that
you spend so much time withcustomers of all varying sizes
and different industries, and Iknow that your passion around
innovation and innovating andsimplifying complex technologies
, because we've grown up with alot of complexity and we're
(12:57):
obviously having this nicetrajectory down towards
simplification, with a lot ofsmart technology behind the
scenes.
But how does your role allowyou to, you know, have a bit of
impact in the direction of Ciscoand what sort of conversations
you're having with customers?
What are the key things?
Carl Solder (13:12):
So I'll answer that in two ways.
Sure, one part in terms of therole I did in the US and how I
take advantage of that insightin my current role.
Michael van Rooyen (13:23):
Yes.
Carl Solder (13:23):
So the role I had in the US, as I said, was
focused on product and technicalstrategy.
Product strategy is ultimatelydeciding what we're going to be
building next.
When we decide, or in decidingwhat we're going to be building
next, there's a lot of inputs.
There's, for example, what ourcompetitors are doing, what are
new research projects that aregoing on that might have a
bearing on what we want to do,asks from customers, asks from
(13:45):
partners, research that we'redoing internally ourselves on
trying to solve problems for ourcustomers.
So we take all these inputs toformulate a product strategy.
Now, part of what we did was wealso solicit input from our
field.
So systems engineers, accountmanagers, product sales
specialists we do poll thoseindividuals and we get some
(14:07):
input from them.
Now, in my current role, knowingwhat I know about the way the
engineering engine works, andbecause I actually know a lot of
the individuals who aredeciding and building that
product strategy a part of myrole I like to think of myself
as the voice of the Australian,new Zealand customer.
So in my role, I do get totravel a lot, I do get to meet
(14:30):
with a lot of customers, a lotof partners, and I probably
unlike a lot of my peers whowork on segments of the market,
like I might be a healthcaresales specialist, I might be
working on the enterprise or thecommercial or small business.
I look at everything.
So I build up a uniqueperspective of, I guess, inputs
that I hear from customers and Ichannel that back to my former
(14:53):
colleagues in the US my formercolleagues in the US and I'm
pretty forceful around thatbecause I like to champion what
I think is going to be neededfor the Australian customers and
I work with the engineeringteams to make sure that the
voice of customers here isactually heard and is factored
into that product strategy.
Michael van Rooyen (15:11):
Yeah, and with those discussions is there
a common theme and I knowthere's probably so many angles
because of the criticality ofall our connectivity and
cybernet but is there one kindof key discussion that's really
dominating your currentdiscussions?
Look at a high level.
Carl Solder (15:26):
There's a couple of things that I think customers'
interests really at the momentartificial intelligence,
cybersecurity and the.
I guess how to drive moreefficiencies, operational
efficiency, reduce operatorburden, do more with less that's
kind of a common question, butultimately it drives down to how
can I make the network simpler?
(15:48):
Yes, simplicity counts and, Ithink, something that sometimes
gets lost in translation.
Engineers get excited.
They have all these wonderfulideas and they go off and build
all these wonderful things andsometimes we tend to forget the
basics.
We kind of work on everything.
But.
And for customers?
(16:09):
Yes, all these new innovationsare quite exciting and they
absolutely address solvingproblems in the business.
But for the IT practitionerit's how can I get things done
more quickly?
If a problem pops up, how can Iresolve that more quickly?
If I'm going to maybe upgradesome software or I'm going to
apply a patch or I'm going to dosome configuration, how can I
get it done in a more timelymanner without all the fuss?
(16:31):
That's ultimately whatcustomers are all about and if
you think, historically, we'vegone through software-defined
networking that had the promiseof simplifying networks, cloud
computing, taking away the needto have to worry about compute
infrastructure as a serviceplatform as a service, software
as a service, takinginfrastructure away.
I don't have to worry aboutracking and stacking and cabling
(16:52):
and upgrading software.
Those things were all attempts,but even though we've got all
of that in play, customers stillsay it's still complex.
Yes, so there's stillopportunities for companies like
Cisco to think about how we cancontinue to innovate, to drive
further simplicity.
That's ultimately, I think whatwe've started to recognize, that
(17:13):
if we can get those basicsright, then we're going to make
our customers happy.
Michael van Rooyen (17:17):
Yeah, and look for me having a look at
Partner Summit and also CiscoLive today and the key messaging
.
Cisco went through thistransition of incredible
engineering, some of the stuffyou're involved in 6500, you
touched on an amazing productall the way through to where we
are today.
To where we are today.
(17:38):
Probably the biggest feelingI'm getting out of Cisco is this
whole idea of this end-to-endsolutioning better together real
drive and security, how webring all these components
together so that you have thisbetter ecosystem and someone
kind of related to otherproducts where they work better
together.
Can you just touch on that,because I'm really picking that
as a big thing as opposed tothese silos that everyone used
to deal with Yep.
Carl Solder (17:54):
So I think the silo thing is something that's
manifested itself out across theindustry.
You look at multi-vendorenvironments.
You have the best firewall, youhave the best phone system, you
have the best switch, you havethe best access point, and when
you patch it together it becomesmore challenging.
I think, one thing that Iobserved in the US maybe
(18:16):
customers saw the realization abit ahead of where Australian
customers have seen this is thedifference between OPEX and
CAPEX.
So if you think back to why dowe build silos?
Because it was focused on whobuilt the biggest, the best, the
cheapest, and so you ended upbuilding all these best of breed
(18:37):
individual products, but as youstart to stitch them together
to your point, the OPEX sidebecomes much larger.
Now in the US we had this.
I think a lot of customers Ispoke to had this rule of thumb
that for every dollar you spendon CapEx, you spend $10 on OPEX.
So the question we would poseto them is if we can save you
(18:57):
20% on CapEx versus 20% on OPEX,what would you prefer?
And it's absolutely OPEX, ofcourse.
So, while you might spend alittle bit more on individual
products, but if they worktogether better as a system and
you can save more money on theOPEX side, that's what customers
want.
So that's this notion and thisdirection to move towards
platforms.
(19:17):
Yep, and certainly Cisco hasrecognized that.
And while we've builtbest-of-breed vertical stacks,
like you look at, for example,if you're looking at a cloud
stack, you might have Nexus.
You're looking at a WAN stack.
You might have the CatalystSD-WAN.
These are discrete stacks fromthe controller to the devices,
the OS it kind of isverticalized.
(19:37):
But can we build something thathas better horizontal
integration?
That's what we're starting todo.
We've probably taken the leadthere on the security side with
the security cloud, and we'vealso got an initiative and a
strategy on the networking sideknown as the Cisco Networking
Cloud, where we're going tostart doing that horizontal
integration drive betterplatform integration.
(19:58):
Again, focus on reducingoperator burden and improvements
in operational simplicity.
Michael van Rooyen (20:04):
Yeah, yeah, and look, I think it's a
fantastic journey and if I thinkabout the breadth of products
that the portfolio has, bringingthat all together, it really is
that amazing ecosystem.
And whilst historically peoplehad you had two camps, it was
opposed to end-to-end singlevendor best of breed.
You know, let's choose the bestfile.
In fact, we even you, would befamiliar with architectures
(20:24):
where we even have doublefirewalls as an example.
But it'd be two vendors to dothe same thing and that has lots
of challenges alone.
But I think the world isgetting more comfortable in
relation to having it all underone umbrella, having it all
under one banner, one managementplatform.
It's an incredible long journey.
It's been a vision for a while,but I think it's nice to see it
coming to fruition.
Now it does.
Carl Solder (20:44):
And it comes back to that CapEx OpEx thing.
You might have to spend alittle bit more money and it
might be that you don't get thebest of breed of everything.
But if you get betterintegration end to end, that's
going to yield a better outcomefor operators, and I use apple
as probably the best example inthe industry yes, you know, my
family has apple everythingexcept for my son, who has an
android phone.
So no facetime.
(21:06):
Yes, you know, no iMessage.
There's things that, like, youdon't quite get, and so it
becomes just that little bitmore difficult, but when you're
inside that ecosystem,everything's seamless, of course
, and so that's, I think, whatcustomers are starting to
recognize that if they stick toan ecosystem, they can drive
better outcomes.
Michael van Rooyen (21:25):
Yeah, 100%.
In the conversation a little bityou touched on AI and we'll
have to talk about it becauseit's so topical and your
customers are talking about it,and 18 months ago it would have
been a different conversationwe're probably having today, and
you talked about a lot of theintegration that the team's
doing around AI bots and AItools to simplify the operation,
as you said.
But can you just talk a littlebit more about?
(21:45):
Earlier you touched on one ofthe key pillars is AI data
centers, or building for AI datacenters.
Do you mind that's a reallyinteresting area for us?
I'm keen to hear about that.
Carl Solder (21:54):
So Cisco, its AI strategy is really centered
around a few aspects.
There is the infrastructurepiece.
So that is compute, that'sstorage, that's networking.
It's what are we doing to buildthe infrastructure that's going
to allow customers to run theircustom AI workloads.
So there, you're talking aboutcompute platforms with heavily
loaded with GPUs.
You're talking about highbandwidth, high speed Ethernet
(22:18):
400 gig, 800 gig.
Moving to the future, where itwill be terribly expensive, plus
, yes, and getting thatinfrastructure right so that
consumer customers can actuallyrun their our workloads.
So we're heavily focused onthat.
It's centered around our siliconone technology, silicon one is
foundational to that partneringwith industry vendors like amd,
(22:39):
like nvidia, like intel.
Those organizations are goingto have technologies that will
integrate into thatinfrastructure.
Yes, to bring that to life.
The other areas that we're alsolooking at is the embedding the
ai in our solutions.
So you're off the shelfsolution you take advantage of.
Show us a I don't know marketdashboard, for example.
It has radio resourcemanagement in the back, ai
(23:02):
monitoring, adjusting, finetuning your wi-fi settings
without the operator having todo anything, and then you have
the new emerging AI assistance,which is changing the
operational paradigm.
So while we kind of front andcenter and talk about AI-ready
data centers, that is a heavyfocus on the AI infrastructure
that we're building.
I also, as a CTO, I try to makesure customers don't forget
(23:25):
that there's also other aspectsto our AI journey that can also
benefit them as well.
Yeah, sure, sure.
Michael van Rooyen (23:30):
I guess you're in a good position there,
right?
People are going to need lotsof infrastructure for this and
as well as, how do you grappleit internally to make sure
people are getting the best outof the products?
And I think about the data setsthat you have available.
You know you've got suchnetwork, cyber, collaboration,
all this endpoint, thethousandized data.
It's just incredible how wecould use those data sets to
(23:51):
really drive the outcome.
What I'd be keen to get fromyour point of view as well, as
we talk a lot and I'll lead to asecond section to this question
, but we're seeing a lot ofcustomers needing to go back to
having workloads remote or closeto the edge.
That probably leads a bit to IoT, which we'll touch on in a
minute, but I'm also seeing thatmaybe a little bit in the AI,
not a full workload.
We need massive data centers,lots of power.
Are you seeing the same sort oftrajectory?
(24:11):
A little bit of compute wasgoing to the cloud, but now,
coming back a bit, we are.
We are.
Carl Solder (24:15):
So actually I was part of a set of CR roundtables
early this year country managerfrom idc who was there guest
speaking and they had mentionedsome research that they did.
It was posted middle of theyear.
Based on their survey, 80 oforganizations they polled said
that they'd be repatriating someor most of their workloads and
(24:38):
data.
But back on prem right right andit's actually I think I don't
know it's necessarily going togo from 100 to 0.
I kind of talk about it asrebalancing there's for a
multitude of reasons there'scompliance, there's sovereignty,
there's latency, there's cost,there's skills, there's security
, and I've spoken to quite a fewcustomers here in Australia who
are quote repatriating, and oneof those areas is AI.
(25:03):
So AI obviously very computeintensive.
If it's compute intensive, youknow that it's going to cost a
bit of money.
I actually refer customers toan interesting article that got
posted, probably about threeyears ago, by Andresen Horowitz,
which is one of the biggest VCson the planet.
And it was authored by a guycalled Martin Casado.
(25:25):
Martin Casado was a PhDdstudent at stanford.
He invented sdn softwaredefined networking.
He also invented openflow right.
He created a startup callednysera, got bought by vmware for
over a billion dollars and thatended up becoming what is now
known as nsx.
But he works for andres andhorowitz and he did he did this
study on cost of the cloud heactually his article the
(25:47):
Trillion Dollar Cloud Paradox,but he actually has an
interesting quote which I callout to customers and I normally
get a smile, and the quote isyou're crazy if you don't start
in the cloud, but you're crazyif you stay on it.
Right, but you're crazy if youstay on it, right.
And his argument is that theflexibility and the elasticity
(26:11):
of cloud and the cost metrics isreally great when you're
starting off.
Yes, but for many organizationsthat they've spoken to and in
fact it kind of holds true withsome customers I've spoken to as
well that the cost benefitanalysis that they did right at
the beginning of their cloudjourney doesn't necessarily
marry up to the actual coststhat they're bearing right now,
(26:33):
right.
So there is that desire from ITleaders to find ways to trim or
refine their cloud costs, andwe've had discussions with some
customers who've done kind of anew cost benefit analysis and
said, hey, actually it will becheaper for us to run it on-prem
than it will be for us tocontinue running the cloud.
So there's some things thatwork well in the cloud.
(26:54):
Yes, there's also someapplications and workloads that
might actually be better to runon-prem, and ai is one of those.
It's a great, maybe start offyour ai journey, but if you're
going to be serious aboutrunning AI, then maybe there's
an opportunity to considerrunning it on-prem.
And we've started to see moreso globally not so much in
(27:14):
Australia as organizations arenow starting to dip their toes
into actually building andputting into production
AI-enabled workflows.
A good percentage of those arenow starting to see that being
run locally or outside of in aprivate cloud as opposed to
being in the public cloud.
So I think over the comingmonths and years it's going to
(27:39):
be interesting to see where theindustry and customers' heads
are going to be at.
But there's definitely amomentum in some parts of the
customer base where they'rereassessing where they're going
to be running.
Michael van Rooyen (27:48):
Yeah, look, I completely agree and I see
that as well.
I think I see it in a couple ofdimensions.
First of all, some customersand I really like your thing
about starting in the cloud andnot necessarily stay there.
I think we're going to end upwith hybrid AI if that's a new
term, maybe, but effectively I'malso seeing customers' mindsets
say all these tools that havebeen developed that are in cloud
and you can ask questions andyou get these fantastic results
and you can do drawings andvideos and all that great stuff,
(28:11):
right.
I'm also seeing customerssaying well, we need to
understand that we need to buildour own LLMs and of our own
data, and the perception is thatthat should be on-prem, because
I need to know where itcontained it and own it and
everything whilst I'm on thecloud.
So I think there's thatdimension on the physical aspect
.
Second part is that you're right.
Not all workloads need to be inthe cloud to do that right.
(28:31):
You don't need a mass scale.
So much compute and performanceis happening at the edge, you
can probably get the rightresults for businesses at the
closest or far edge.
And then last dimension to thatis the IoT play right.
How do we ingest data to that?
Iot is going to be one of thebig feeds into AI models.
So we're seeing a lot ofcustomers look at that.
Carl Solder (28:49):
I think there's something else which is of
interest, and that's the generalmove towards building smaller
language models.
Yes, the small language models,the micro language models, the
domain specific language models,the ability to potentially run
a model on a laptop or a tablet.
Yes, that opens up a whole rangeof opportunities that you might
not be able to necessarilybuild with a more generic large
(29:12):
language model.
So I think there's so much.
I actually was talking tosomeone recently and I kind of
think of AI like the iPhone.
When the iPhone came out in2007, it was a very different
device then to what it is today100% and I don't think back then
people fully appreciated thepotential of what that could do.
And when you think back thenyou couldn't do online banking,
(29:33):
you couldn't order your pizza,you couldn't check where the bus
was, couldn't do your bankingthose use cases kind of got
derived on that journey.
And I think the same thing withAI.
We're at that moment where weknow there's a lot of potential,
but I don't think we trulyrealize exactly what it
potentially can do, and I thinkthat will be discovered.
Michael van Rooyen (29:54):
I hope you enjoyed part one of my
discussion with Carl Solder.
Cisco ANZ CTO.
Tune in next week for part twoof the special Cisco Live
edition of Securely ConnectedEverything.
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!
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy And Charlamagne Tha God!
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.