June 12, 2023 • 33 mins
Each week, TechTarget Editorial's news team and industry experts provide candid insight and analysis of the biggest IT news headlines of the week. Today's show focuses on the key product launches at Cisco Live in Las Vegas. We talk to Cisco executives at the show and industry analysts from Enterprise Strategy Group.
Watch Tech News this Week on the TechTarget News Bytes Channel:
Stories featured on Tech News this Week:
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Unknown (00:00):
because it's just going to force you know Cisco to to
(00:02):
drive and push back and, anddeliver their own innovative
technology as well. Andultimately, it's going to be the
end consumer who benefits.
Antone Gonsalves (00:12):
Hi, and welcome to Tech News this week.
I'm your host, techtarget,editorial news director and
jungle and solvus. As the introsuggests, on today's show, we'll
cover the new technologiesintroduced at the Cisco lie con
live conference in Las Vegas. Iwas there, along with analysts,
(00:33):
Paul nashawaty. And Bob, LalaBert from tech targets
Enterprise Strategy Group. Hi,Paul. Hi, Bob. Thanks for
joining me.
Unknown (00:44):
Thanks, Tom.
Antone Gonsalves (00:45):
All right. So Cisco Live, show the company
taking a platform approach tomodernize this technology. For
applications are distributedacross the cloud edge and
Datacenter environments. Notabletechnologies presented at the
show included a full stackobservability platform, and a
(01:06):
new approach to networkmanagement called networking
cloud. Paul will start with youwhat's in full stack
observability? And what's theimpact on Cisco customers?
Unknown (01:18):
Thanks, Anton. Yeah, you know, I will have to say, an
exciting event. And like thetheme of the event, let's go was
the theme. Let's go right intothis. Because there's such a,
there was a lot to talk about,you know, with full stack
observability. I know we have ashort period of time here. But
folks, Cisco's full stackobservability platform and
digital experience monitoring,and they their beliefs of it
(01:39):
came out really during thisevent to show how customers can
build applications acrossecosystems as open extensible.
It really kind of drives on thearchitecture. But you know the
thing about it, it's reallyinteresting, because when you
start looking at full stackobservability, and what that
really means, Cisco has addedeverything into it, and which
(02:01):
allows for a very robustsolution. So what I mean by that
is, there's bi directionalintegration between AP, AP
dynamics and 1000. I support forDTN. It also closes that
observability gap with rapid,actionable recommend
recommendations and insights. Sowhen that is kind of applied to
how organizations are movingforward, and their cloud native
(02:23):
approaches, it really does allowfor those actionable insights
across the ecosystem.Furthermore, you know, you want
you see these cloud nativeenvironments, it really does
come down to maturity. This iswhere I think Cisco may have to
struggle a little bit when itcomes to full stack
observability. Because a lot oforganizations organizations have
(02:45):
a really not mature and theirobservability practices these
days, right. And by doing sothey have to start somewhere. So
they may be doing monitoring oralerting, but that all that all
those insights that are that areactionable when using AI to kind
of drive this and looking atacross the entire ecosystem that
may be too robust for a lot oforganizations. So I think the
(03:06):
crawl Walk Run approach thatthat Cisco is looking to achieve
with this is the right approach,flip on the things that make
sense. But meet the customerwhere they are.
Antone Gonsalves (03:15):
All right. And Bob, networking cloud, why don't
you give us a rundown on whatthat is and how it's going to
benefit customers, according toCisco?
Unknown (03:26):
Yeah, absolutely. Like, like Paul said, there was a lot
to unpack in the show, therewere a lot of announcements. One
of the bigger ones was theyreleased a networking Cloud
Vision, as most organizationsknow, they unveiled a security
cloud last year. So now they'removing on to the network side
and taking a look at how theycan effectively consolidate the
(03:47):
number of platforms they have,to their ultimate goal of
getting it down to basically twoplatforms, which is one which
would be a cloud based, and onethat's on premises. But the key
underlying features of that isthat they would also make it
simpler to make that transition.So similar to what Paul was
saying about meet the customerswhere they are, Cisco is very
(04:07):
cognizant of their installedbase and the fact that they need
to bring them on and noteveryone's going to jump to the
cloud. So they've been doing alot of things like doing single
sign on. So today with thatcloud, they've been talking
right, that's one of theannouncements that you'll be
able to go in and regardless ofthe platform, the domain you're
in, be able to get to it from asingle web based portal. So
(04:27):
things like that will help someof the other announcements
included things like the Merakicloud base, which has been
monitoring the catalyst switchesnow has the ability to show the
CLI show commands. So for thosepeople who are tightly I guess
attached to CLI commands andthose right the legacy CCI ease
(04:48):
who had been made a huge livingof Cisco reusing those Cisco
products are now going to havethat opportunity to be able to
still see some of those commandsbut also then, you know,
transition to cloud If they wantto. So like I said, this is a
big announcement, obviously,Cisco's had a lot of different
products that they've acquired.So that idea of they're going to
(05:10):
simplify that process provide aunified interface, right? I like
to refer to that as theprinciple of least astonishment.
So when you go from one domainto another, it still looks very
familiar, very intuitive for youto be able to use it. Right. I
think that's what Meraki hasalways prided itself on being
somewhat of the, the apple ofthe networking space. And so
(05:30):
they're looking to capitalize onthat and drive it forward. But
like I said, overall, you'reseeing a massive amount of
consolidation that will takeplace or or integration, if you
will, to a common platform thatwill enable organizations to
drive much greater operationalefficiencies, as those network
environments get far morecomplicated.
Antone Gonsalves (05:49):
Yeah, Cisco talked an awful lot about
simplicity. I mean, I can't evenI can't even count how many
times I hear that word from hisex. But when you look at Cisco's
become Cisco technology,overall, is highly complex. I
mean, they have multipleoperating systems, they have
multiple network managementplatforms, a lot of products
(06:13):
that overlap, all this has to bebrought together. In the
meantime, you have companies,for example, John Chambers,
their former CEO, he launched acompany last year now. And he
knows Cisco. And his companytakes it is going after the
wireless LAN. By making itreally simple. I mean, it's
(06:35):
basically as a service, right? Imean, you buy the they installed
the access points to switches,and they take care of
everything. You don't have thatpoint, you just use it, right.
And they charge a charge by theuser, by how many people on the
network? It's pretty simple. Myquestion is, Cisco seems to be
(06:58):
really good for largeenterprises who do a lot of this
stuff themselves, right? Theycan deploy the technology to get
the talent. But when you startgoing down to the healthcare
industry for major universities,right, for that mid size,
enterprise maybe a littlelarger, it seems like companies
like chambers, Niall, that hiscompany, have have a better
(07:23):
opportunity. So what's, what'syour take on all of that?
Unknown (07:28):
Yeah, I think Well, I think there's a couple of
things, right. I mean, it's,it's obvious easier when you're
starting from scratch to come upwith a solution that it's you
know, you're driving simplicityin from the get go than an
organization that's beendelivering products for over 20
years. And you have to keep inmind, right, a lot of times,
Cisco products are the sameproducts are installed for over
20 years. So what you talkedabout the complexity, and that's
(07:50):
exactly what they're trying toaddress with this, right?
They're trying to take all thoseplatforms, and boil them down to
just two right, can be able toyou'll see that consolidation of
the operating systems. You'llsee I think a lot of those, a
lot of those complexities beingtaken out now. It's going to
take time to do that. But theyare also keep in mind, you're
right, to a certain extent,they're working with some of the
(08:11):
world's largest organizationsand the world's largest
networks. So there are certainlyorganizations that take
advantage on the you know thatthat smaller and the colleges
universities that have Merakideployed, right, simple cloud
based consumption based models.Cisco is also moving to more,
you know, consumption basedmodels, simpler licensing. So
there's things that they aretrying to do to get them to
(08:32):
that, I think, from theperspective of delivering
something as a fully managedservice or a network as a
service. I think today you'reseeing a lot of their partners,
delivering those and providingthat that layer of abstraction
and delivering the upfrontassessment and design and, and
so forth. But yeah, there'scertainly going to be innovative
companies like Niall and others,who are going to be pushing
Cisco and driving them, youknow, to compete and deliver
(08:55):
more, more guests operationallyefficient and simpler systems,
for sure.
And I'll add on to that, Anton,you know, if you look at the the
overall picture of Cisco overallSimplicity is the direction that
they're going. And you know,when you look at what they've
announced for the differentbusiness units, openness,
extensibility scalability,flexibility is really the key
(09:15):
message to drive in the home.Give me another example of how
they're making the company alittle bit more user friendly, a
little bit more simple. Changingrebranding the name of ETSI the
emerging technology andincubation team to open up from
start to out shift out shift isthe new brand. So there in the
conversation was in we werespeaking specifically around the
(09:36):
service mesh, takes sto and takeCalista T Kalista allows you to
do use sto far faster and easierthan using sto by itself. So
it's it's about providing thatuser experience and making it
easier and simple to use. And Ithink Cisco is making those
changes and they're reacting tothe market dynamics.
(09:56):
Yeah, another great example tothat is also the like 1000 eyes
and what Going there andexpanding, going all the way
down to the WebEx endpoints andso forth. Right? So we're so
much of the talk today about thecorporate network is the
internet, being able to havevisibility into that
environment, but then all theway down to potentially a home
office provides a lot ofbenefits to organizations as
well.
Antone Gonsalves (10:17):
So So in your opinion, Cisco, will be asking
can be as competitive in thatmid market as a as some of these
other companies like thestartups like Nile but of
course, you also have a Aruba?Arista?
Unknown (10:32):
Yeah, there's, I'll tell you this. Anton, they're
certainly going to try it.They've been doing this for a
long time. They've got a lot ofcustomers SMB all the way
through enterprise. And right,they still have over 50% of the
market. So like I said, I thinkit's great to see the innovative
companies coming up right,driving that that unified
(10:53):
networking approach, driving alot of the the AI approaches and
things like that drivingsimplicity, because it's just
going to force you know, Ciscoto, to drive and push back and,
and deliver their own innovativetechnology as well. And
ultimately, it's going to be theend consumer who benefits.
Antone Gonsalves (11:10):
Okay, right. So you bring up a good point, to
a large extent, this is Cisco'smarket to lose, so to speak,
since they they there's such adominant force within within
networking and insecurity. Okay,next are interviews I conducted
at Cisco Live with companyexecutives, JL Valenti, Vice
(11:33):
President of Product Managementfor multi cloud networking and
Jeremy Foster, General Managerof Cisco's server portfolio.
Foster is also involved inhelping customers with their
sustainability efforts. Hi, I'mat Cisco Live, I'm here with JL
(11:55):
Blent. He is VP of ProductManagement for enterprise
routing, SD Wan and multi cloudnetworking for Cisco. JL, thanks
for joining me.
Unknown (12:07):
It's a pleasure. All right. Thank you for having me.
Antone Gonsalves (12:09):
Okay. So, first question I have, I think
we're going to focus a lot onmulti cloud in this session. It
seems multi cloud is the norm,it seems for large and mid sized
enterprises. Lots of options areout there. Start startups that
offer connectivity as a service,public clouds offering Internet
(12:34):
inter inter connectivity, andcome at vendors like Cisco. So
technically, how is Ciscodifferentiating itself? In the
market?
Unknown (12:44):
So that's a great question. And yes, indeed, multi
cloud is omnipresent, pervasivethroughout, you know, this
conference, but last year andthe year before, I think,
obviously, this is only growing.First, you mentioned enterprise
in a mid market. In fact, I canjust also witness that we've
(13:06):
seen it across every singlesingle business can be public
sector, and be DOD. In the US,but also anywhere in the world.
Cloud is, is pervasive cloud iseverywhere. And our clouds can
be you know, some have even whatthey call distributed hybrid
cloud, which obviously, you sayit's multi cloud, it's multi
(13:29):
cloud, so multiple cloudproviders, which Instructure, we
will discuss. But it's alsohybrid, in the context that not
everything is running on apublic cloud, but also in
private clouds, or data centersthat obviously behave as as
cloud and now need to bestitched together as people were
going through an evolutionprocess. In that, you know, in
(13:53):
that vein, so back to Cisco, A,we realized, obviously, years
ago, the criticality of helpingour customers, being able to
harness the possibilities andthe capabilities that are
available to them in the cloud,in the clouds, and do that in a
(14:16):
simpler way. So there's a lot ofcomplexity, I'm sure we'll talk
about that. So what we reallytry to bring is simplify the
whole process of, you know,defining and accessing to
workloads or applications in thecloud can be SAS, that can be
obviously yes, you know,applications, and do that in a
(14:37):
very secure way. And coming withthat you Sorry, I've heard about
the full stack observability.Not only do you need actually to
provision and define that butyou need to assure it so the
whole aspect of visibility ofmonitoring is critical to be
able to detect and respond tosome of the threats, outages
(14:59):
problems that you may encounter.And so we really take a holistic
approach to cloud throughout
Antone Gonsalves (15:05):
Cisco. Okay. So when you when you talk about
simplicity or simplifying thewhat's a very, very complex
environment? How do you? How doyou define simplicity? What does
what does that mean? So readersunderstand
Unknown (15:19):
it? Yeah, that's a very good question. So it means
really first building. So it'sall about architectures and
building constructs,abstractions, that, you know,
pretty much hide the complexityof the underpinning of
stitching, for example, accessto the cloud, making it actually
more normalized, more consistentacross cloud. One of the key
(15:41):
areas, obviously, for what Irun, I wake up, you know, you
mentioned enterprise writing andSD Wan is, you know, the first
first step was, hey, how do Iconnect users or branches? To
the cloud to different clouds?How do you do that, when you
have, obviously, different cloudnetworking? Obviously, the way
the constructs work on a jurordifferent from AWS different
(16:01):
from Google GCP? So how do youcreate a level of abstraction,
and a level of automation, thatgives people the comfort through
workflows, etc, to really beable actually to deploy at
scale, in a matter of reallyminutes or hours, not days, and
(16:22):
actually keep it, you know,fluid, flexible, because there's
a lot of dynamicity on top ofthe cloud even more. So now,
those applications are, youknow, spinning up, you know,
this horizontal scale, verticalscale, horizontal scale. So
you're spinning up between newcomponents, you have DevOps,
trying to bring new products tomarket to applications all the
time, how should we NETOPS keepup with all of those aspects is
(16:45):
very critical. And so you needthat level of automation, you
need a level of integrationbetween the different parts.
Antone Gonsalves (16:51):
So when you talk about needing an
abstraction layer, what isCisco's abstraction layer.
Unknown (16:57):
So those abstractions are really related to, for
example, the construct in thecontext of SD, when we call
cloud on ramp is really beingable actually to, you know, to
assemble all of the parts thatwould create automatically, the,
what we call the tunnels, thefabric, the the overlay, and
(17:18):
make it very simple, so thatusers to applications,
applications were in datacenters before now they are in
clouds, they might be now evendistributed in the same cloud in
different regions, you mighthave a database somewhere, and
you may have actually thedefault application in a
different place. How actually,do you create a construct
(17:40):
through actually, you know,those, the tunnels where we
could the fabric, so you try toabstract that to a level of
fabric and define policies thatyou can apply to say, hey, if
those sites are those users wantto access those applications,
this is what the policy shouldbe they can they cannot, they
should be able to have, youknow, you can define SLA s in
(18:02):
terms of what you expect interms of latency, you know,
jitter, packet loss, etc. To beable actually to have the system
automatically with its ownintelligent, determine what is
the best path to get to thatapplication, which can be
obviously replicated in multiplezones, as you said, to maintain
actually, the reliability, we dothat behind the scenes means
(18:23):
obviously, there are a lot oftechniques. DPR is one of the
techniques where we grow NP. Sothey do control planes that we
have very sophisticated to beable to, you know, centralize
the routing and define theroutes and dynamically as we
take into account, the policies,the situations of maybe
bottlenecks are happening at anygiven time in the network, etc.
So there's a lot of complexitythat we encapsulate bottle up
(18:46):
and make that available as asimplified pretty much workflow
and instead of constructs tocustomers.
Antone Gonsalves (18:51):
Okay, so given the complexity of their cloud,
do you do you find that most ofyour customers are turning to
manage service providers? Ifthey want Cisco technology,
they're depending on yourpartners, your managed service
providers? help them deploy us?Yeah, I
Unknown (19:08):
think so. Again, there are many different breeds,
breeds of managed serviceproviders. There are breeds of
managed services that were bornin the cloud. So they those ones
are, are new. They were builtmaybe 10 years ago, and you
know, or from the club 2008 20years ago, but 15 The thing is,
(19:29):
the more classic managed serviceproviders that connectivity
managed service provider, willcommunication service providers
to beat ups. Yes, indeed theyare. They've evolved they have
capabilities to be able actuallyto help those customers to the
cloud, most likely today's it isactually still a branch or site
to cloud and obviously, more andmore people work from anywhere.
(19:52):
So how do you get also users toclouds from you know, you know,
hybrid work type of environment,but more and more Those
applications then set us ondistributed, and the its data
center to cloud one cloud andmulti cloud, etc. And you start
seeing, you know, like aI mlapplications, data residing in
one cloud. The Tensor flow, forexample, from Google reside,
(20:15):
obviously, the GCP environment,how do you bring those together
in a very secure manner as well,because the data that is
underneath is very proprietary.And so being able actually to
create, you know, that thosecapabilities for, for those
customers require systemintegration, and managed
services, so that, you know,clearly a number of customers
(20:39):
don't have the people don't havethe skill set, or can keep up
with the, the evolution and thetide of all applications,
including SaaS applications thatare combined with that at times,
whenever he says, folks andServiceNow, in SAP, etc. So you
have those compositeapplications more and more, the
(21:00):
large one, the very large, or,you know, there are a number of
customers or a large data set,hey, I can do that myself. I
manage my network, I've managedintegration with the cloud, and
that's fine. And, and you have alot of obviously those, you
know, enterprises salary, themid market, but also the large
ones that went out in thisbusiness. That's not what we do.
(21:21):
And yes, indeed, we rely uponone or multiple partners, some
actually divided even theirmanager receives in regions and
obviously stitch that together.So you gotta beat off of
everything. And that's whatmakes also this market quite a
bit interesting.
Antone Gonsalves (21:35):
Sure. And it makes sense to me, given the
complexity, the scarcity oftalent, I would think that
enterprises are becoming moredependent than ever before on
their technology partners, aswell as service providers.
Absolutely. JL, thank you verymuch for your work. Appreciate
your time. Thank you. Hi, I'mhere at Cisco Live with Jeremy
(22:01):
Foster, General Manager of CiscoNetworking. Compute. Jeremy,
nice to see you.
Unknown (22:08):
Nice to be here. Okay.
Antone Gonsalves (22:10):
So I wanted to start off our conversation on
hybrid cloud, it seems that thehybrid clouds evolve and where
it's becoming possibly lessabout the data center and more
about the edge, as we'recompanies are saying, are
deciding that the data, it'd beeasier to start fresh at the
(22:33):
edge, rather than take legacytechnology and somehow adapt it
to the cloud. So what are youseeing? And what's your
observation
Unknown (22:43):
as to spend a lot of time with customers, I think,
you know, over the last sixmonths, and even at the show a
lot of talk about, you know,hybrid cloud and customer
settling into the fact that it'snot going to be everything in
one public cloud, where you runall your infrastructure to your
point, you have a lot of datacenters that are out there,
where you've already madetremendous investments as a
business. And then you have thisemerging edge edges, this huge
(23:04):
spectrum of things, right, thathave all different kinds of use
cases, whether that's a retailedge, or whether that's an
enterprise edge where you wantto try and do things, you know,
along new use cases likeinference at the edge, where you
want to get the as close to thewhere the data is being
generated as possible with thetechnology. And those are almost
like taking a mini data centerand building it out in lots of
(23:24):
locations. And so the theme fromcustomers is coming back is just
we want to live in this hybridmode. But that generates a ton
of complexity, because now Ihave to figure out how to
maintain my assets on prem, dothe things that I want to do
with my applications in newapplications, oftentimes, that
are developed in the cloud. Andthen obviously, go after these
new use cases that my businesswants to do. So it's, it's quite
(23:46):
a bit of things for them to dealwith. And that's really what
we're focused on is trying tohelp simplify that experience,
and help them operate in auniform manner across all those
different areas.
Antone Gonsalves (23:54):
And it turns out, sustainability is something
I want to ask you about, becauseit seems to be top of mine of on
everyone, tech vendors as wellas as well as their customers.
And, you know, when you when youtalk sustainability in the data
center, you're you're addingcomplexity to an already very
(24:15):
complex environment. So how isCisco helping customers address
that problem?
Unknown (24:21):
I think it is top of mind. It's good, because it's
good for the world. But it'salso good for business in terms
of helping customers be focusedon how can they optimize for
those environments and go afterreplacing those legacy
environments with newertechnology, and the newer
technology really helps drive alot of the power of that. And
that's one way that we can helpright, we can build the best,
most sustainable infrastructurein the industry. Certainly from
(24:46):
a compute perspective, that'snot anything new for us. We've
been doing that since 2009. Andif you take a typical customer
scenario, like one of the largefinancials that we were working
with, they have 500 servers inone area, their data center.
Those processors were a couplegenerations old Replace that
with UCS acts as 134 serversinstead of 500. So drives out,
not only 31% of the power drivesout 39% of the maintenance
(25:11):
costs, it drives out of 66Roughly percent of the actual
physical space that it wastaking up. Right. So there's a
lot of business benefits togoing after sustainability. And
I think the other area customersare struggling that we're trying
to help out with is the data.We've tracked the data inside of
a data center since forever ago,right? We've had power strips
(25:31):
that could report back to thefacilities people, how much
power are we using in aparticular area or zone of the
data center. But a lot of thatis data that is typically with
those operational teams, andmaybe not with the line of
business folks that are nowtrying to get an assessment of
how are we doing towardsreaching those ESG goals? So
we're trying to leverage thingslike intersite, where we are
collecting data across thatserver state and tracking it
(25:54):
over time. So customers canunderstand how much power Am I
using three months ago? How am Idoing towards reaching those ESG
goals? And then ultimatelyoverlaying that with where is
this power coming from? Is itcoming from solar? Is it coming
from wind? Is it coming fromcoal, and then helping them
understand what thatenvironmental impact is as they
reduce the power that they'reusing?
Antone Gonsalves (26:15):
Right? I believe Cisco, one of the
announcements in is Cisco Livewas the addition of
sustainability metrics withintheir nexus, guess. Fabric. Yes,
management fabric that wasinteresting. So is
sustainability when it comes toservice is always an upgrade
process to customers have toalways buy new servers or
Unknown (26:39):
not? It's a great question. And one other angle to
sustainability is how are youdoing the most with what you've
got. And so there's alldifferent kinds of features that
we've put into our solutions,whether it's dynamic power,
capping multiple power zones ofhow we lay out our UCS x
chassis, that are activelysettings, you change in software
to change how the power bus isworking, so that you can look at
(27:01):
the needs of those servers basedon what the application is
asking those servers to do. Anduse less power. In other words,
change how redundant yourrequirements might be, the more
redundancy you push out there,the more power you're going to
need. But you can actuallyoptimize those types of things.
So we're trying to make thosevery easy to expose those
capabilities for customers, anduse those, I'd say, in addition
(27:21):
to that, there's a lot of newfeatures and capabilities around
optimizations from say, Intel,AMD that we can put into our
software platform to let themleverage those features really
easily and at scale, becausethat's the that's the trick,
right? You want to be able tosee how are those applications
behaving? And how are thoseservers behaving, and then say,
Great, these are greatcandidates, for example, to use
(27:43):
those database optimizationsthat Intel just released, which
will then help increaseperformance and reduce the
amount of power that they need.So you can see,
Antone Gonsalves (27:50):
and cooling is a big is a is a big problem with
with servers. What is, you know,it's also a huge expense. So
what are the solutions forcooling? What is What do you how
are you addressing,
Unknown (28:10):
cooling, it's that's a real complex one in that, the
first thing we can do isoptimize our systems to deal
with in cool only the areas saywithin like our at UCSF chassis,
for example. It's this big, butif there's a lot of power
demands in a certain area wherea certain process or on a
certain servers going. And wehave over 20 cooling zones that
we've engineered in there with,I can't even remember how many
(28:31):
temperature sensors, but we wereable to actually dynamically
change where the fans and howmuch they're driving. So we use
the least amount of powerpossible to actually change that
in real time without thecustomer having to do anything.
So you can really optimize thosepower states there are cooling
states there. When you get tolooking into the future and say,
How are we going to leveragethings like liquid cooling,
there's a lot of impacts to howthat touches a customer, right?
(28:53):
You can, you can put liquidcooling and things like closed
loop liquid cooling that willstart building into some of our
servers when we when we need to.But right now, we don't need to.
But we will, as these newgenerations of processors
consume even more power. Andthat's an easy one for customers
to buy because it just sitsinside of the server. And then
you know, as you get furtherdown the line, looking at how do
(29:13):
we do liquid cooling across thethe data center is something I
think a lot of customers aretrying to figure out, because
it's operationally verydifferent than what we've done
in a data center. It requires alot of physical changes to how
they how they've actuallyplumbed their data center to get
water to the individual backs ofthe racks. And I think that's a
challenge we'll be tackling overthe next five years.
Antone Gonsalves (29:34):
I would change topic for a second. I wanted to
ask you about it your your takeon digital processing units dpus
Because that's being introducedinto servers to offload from the
work from the CPU. Do you seethat as something that could
dramatically change thearchitecture of data centers or
(29:58):
not?
Unknown (29:59):
I think Over time that will. And the thing about GPUs
is it's in our enterprisecustomer base is there's
different people with differentmarkets, right GPUs have already
dramatically changed how datacenters are designed in CSPs,
cloud scalars, and high endservice providers that
understand how to take thosetechnologies and have specific
enough needs at large enoughscales to use them. But it's not
(30:21):
something today that youraverage enterprise account, say
your healthcare organizationthat has a really great, awesome
data center has had to go pulloff the shelf and start using
those technologies. So I thinkas we make those technologies
more easy to consume, peoplewill use those technologies more
readily in say, the enterpriseat scale. And that'll be the
(30:41):
kind of changes that we see overthe next three to five years.
Antone Gonsalves (30:44):
Right now. Anyone who is primarily a cloud
technologies, someone a lot or alarger enterprise, large,
Unknown (30:51):
large scale type applications, are people who
have large data centers, becauseit's just for the same reasons
you're talking about, how do weincrease sustainability,
increased performance, GPUs area different way of doing
processing that can helpincrease the performance and
take, take all that workload offthe CPU, so you get to do more
with less.
Antone Gonsalves (31:09):
And obviously, the management tools have to be
there also don't have to change.Okay, so the, you know, the
cloud in the data center, how isyour data centers are shrinking,
because more workloads are goinginto the cloud? How else? Is the
data center changingarchitecturally? Because of the
(31:32):
because of the cloud?
Unknown (31:33):
Yeah, I think a couple of things, you look at the spend
in the industry over the lastfew years, clearly explosive
cloud growth, but on premrevenue spending is flat. And
it's not, it's not this binarything where everybody said, we
spent this much money in thecloud and all that money in the
data center instantly goes away.It's we all consume more
applications, we all generatemore data, we all do all that
stuff in our daily lives morethan than we did five years ago.
(31:55):
And so there's also the factthat people run their business a
lot of times based on what's intheir data center already. And
then it's not super easy toalready sometimes cost effective
to lift and shift all thoseapplications to the cloud. And
that's really why customers aresettling in on this hybrid cloud
mode. Because much like we do init, I think what we've seen over
the last years is people wantingto really aggressively go
(32:17):
towards cloud, because it's agreat new tool, it's extremely
operationally efficient, I canmove very fast and it sounds
like, you know, it offloads alot of that operational burden
on on a customer to make allthose things happen underneath
deployment of applications. Butthen the reality is, in some
cases, it's not free, there isno free, there's no free
(32:38):
workload, there's no free lunch,if you will. So then customers
are now saying like, every otherIT problem we try and solve, how
do we optimize around this?Well, I gotta look, and I've got
these assets on prem, I can runsome things on prem infected
customers with decent scale,maybe I can do that application
on prem, or I don't need to movethat application because my cost
profile will be better here.Every application is different.
(32:58):
And so that's, that's, you know,kind of a customer by customer
basis. But what we see is, wesee people balancing that out
and really trying to figure outwhere am I spending? Where can I
get the most for my dollars thatI'm spending being in the cloud
or beat on prem?
Antone Gonsalves (33:11):
Yeah, it's all about defining the problem. And
then and then solving it.Jeremy, thank you very much.
Unknown (33:17):
Thank you for your time, sir. Likewise.
Antone Gonsalves (33:22):
That wraps up this week's show. Thanks for
watching. I'll see you nextweek.
because it's just going to force you know Cisco to to
(00:02):
drive and push back and, anddeliver their own innovative
technology as well. Andultimately, it's going to be the
end consumer who benefits.
Antone Gonsalves (00:12):
Hi, and welcome to Tech News this week.
I'm your host, techtarget,editorial news director and
jungle and solvus. As the introsuggests, on today's show, we'll
cover the new technologiesintroduced at the Cisco lie con
live conference in Las Vegas. Iwas there, along with analysts,
(00:33):
Paul nashawaty. And Bob, LalaBert from tech targets
Enterprise Strategy Group. Hi,Paul. Hi, Bob. Thanks for
joining me.
Unknown (00:44):
Thanks, Tom.
Antone Gonsalves (00:45):
All right. So Cisco Live, show the company
taking a platform approach tomodernize this technology. For
applications are distributedacross the cloud edge and
Datacenter environments. Notabletechnologies presented at the
show included a full stackobservability platform, and a
(01:06):
new approach to networkmanagement called networking
cloud. Paul will start with youwhat's in full stack
observability? And what's theimpact on Cisco customers?
Unknown (01:18):
Thanks, Anton. Yeah, you know, I will have to say, an
exciting event. And like thetheme of the event, let's go was
the theme. Let's go right intothis. Because there's such a,
there was a lot to talk about,you know, with full stack
observability. I know we have ashort period of time here. But
folks, Cisco's full stackobservability platform and
digital experience monitoring,and they their beliefs of it
(01:39):
came out really during thisevent to show how customers can
build applications acrossecosystems as open extensible.
It really kind of drives on thearchitecture. But you know the
thing about it, it's reallyinteresting, because when you
start looking at full stackobservability, and what that
really means, Cisco has addedeverything into it, and which
(02:01):
allows for a very robustsolution. So what I mean by that
is, there's bi directionalintegration between AP, AP
dynamics and 1000. I support forDTN. It also closes that
observability gap with rapid,actionable recommend
recommendations and insights. Sowhen that is kind of applied to
how organizations are movingforward, and their cloud native
(02:23):
approaches, it really does allowfor those actionable insights
across the ecosystem.Furthermore, you know, you want
you see these cloud nativeenvironments, it really does
come down to maturity. This iswhere I think Cisco may have to
struggle a little bit when itcomes to full stack
observability. Because a lot oforganizations organizations have
(02:45):
a really not mature and theirobservability practices these
days, right. And by doing sothey have to start somewhere. So
they may be doing monitoring oralerting, but that all that all
those insights that are that areactionable when using AI to kind
of drive this and looking atacross the entire ecosystem that
may be too robust for a lot oforganizations. So I think the
(03:06):
crawl Walk Run approach thatthat Cisco is looking to achieve
with this is the right approach,flip on the things that make
sense. But meet the customerwhere they are.
Antone Gonsalves (03:15):
All right. And Bob, networking cloud, why don't
you give us a rundown on whatthat is and how it's going to
benefit customers, according toCisco?
Unknown (03:26):
Yeah, absolutely. Like, like Paul said, there was a lot
to unpack in the show, therewere a lot of announcements. One
of the bigger ones was theyreleased a networking Cloud
Vision, as most organizationsknow, they unveiled a security
cloud last year. So now they'removing on to the network side
and taking a look at how theycan effectively consolidate the
(03:47):
number of platforms they have,to their ultimate goal of
getting it down to basically twoplatforms, which is one which
would be a cloud based, and onethat's on premises. But the key
underlying features of that isthat they would also make it
simpler to make that transition.So similar to what Paul was
saying about meet the customerswhere they are, Cisco is very
(04:07):
cognizant of their installedbase and the fact that they need
to bring them on and noteveryone's going to jump to the
cloud. So they've been doing alot of things like doing single
sign on. So today with thatcloud, they've been talking
right, that's one of theannouncements that you'll be
able to go in and regardless ofthe platform, the domain you're
in, be able to get to it from asingle web based portal. So
(04:27):
things like that will help someof the other announcements
included things like the Merakicloud base, which has been
monitoring the catalyst switchesnow has the ability to show the
CLI show commands. So for thosepeople who are tightly I guess
attached to CLI commands andthose right the legacy CCI ease
(04:48):
who had been made a huge livingof Cisco reusing those Cisco
products are now going to havethat opportunity to be able to
still see some of those commandsbut also then, you know,
transition to cloud If they wantto. So like I said, this is a
big announcement, obviously,Cisco's had a lot of different
products that they've acquired.So that idea of they're going to
(05:10):
simplify that process provide aunified interface, right? I like
to refer to that as theprinciple of least astonishment.
So when you go from one domainto another, it still looks very
familiar, very intuitive for youto be able to use it. Right. I
think that's what Meraki hasalways prided itself on being
somewhat of the, the apple ofthe networking space. And so
(05:30):
they're looking to capitalize onthat and drive it forward. But
like I said, overall, you'reseeing a massive amount of
consolidation that will takeplace or or integration, if you
will, to a common platform thatwill enable organizations to
drive much greater operationalefficiencies, as those network
environments get far morecomplicated.
Antone Gonsalves (05:49):
Yeah, Cisco talked an awful lot about
simplicity. I mean, I can't evenI can't even count how many
times I hear that word from hisex. But when you look at Cisco's
become Cisco technology,overall, is highly complex. I
mean, they have multipleoperating systems, they have
multiple network managementplatforms, a lot of products
(06:13):
that overlap, all this has to bebrought together. In the
meantime, you have companies,for example, John Chambers,
their former CEO, he launched acompany last year now. And he
knows Cisco. And his companytakes it is going after the
wireless LAN. By making itreally simple. I mean, it's
(06:35):
basically as a service, right? Imean, you buy the they installed
the access points to switches,and they take care of
everything. You don't have thatpoint, you just use it, right.
And they charge a charge by theuser, by how many people on the
network? It's pretty simple. Myquestion is, Cisco seems to be
(06:58):
really good for largeenterprises who do a lot of this
stuff themselves, right? Theycan deploy the technology to get
the talent. But when you startgoing down to the healthcare
industry for major universities,right, for that mid size,
enterprise maybe a littlelarger, it seems like companies
like chambers, Niall, that hiscompany, have have a better
(07:23):
opportunity. So what's, what'syour take on all of that?
Unknown (07:28):
Yeah, I think Well, I think there's a couple of
things, right. I mean, it's,it's obvious easier when you're
starting from scratch to come upwith a solution that it's you
know, you're driving simplicityin from the get go than an
organization that's beendelivering products for over 20
years. And you have to keep inmind, right, a lot of times,
Cisco products are the sameproducts are installed for over
20 years. So what you talkedabout the complexity, and that's
(07:50):
exactly what they're trying toaddress with this, right?
They're trying to take all thoseplatforms, and boil them down to
just two right, can be able toyou'll see that consolidation of
the operating systems. You'llsee I think a lot of those, a
lot of those complexities beingtaken out now. It's going to
take time to do that. But theyare also keep in mind, you're
right, to a certain extent,they're working with some of the
(08:11):
world's largest organizationsand the world's largest
networks. So there are certainlyorganizations that take
advantage on the you know thatthat smaller and the colleges
universities that have Merakideployed, right, simple cloud
based consumption based models.Cisco is also moving to more,
you know, consumption basedmodels, simpler licensing. So
there's things that they aretrying to do to get them to
(08:32):
that, I think, from theperspective of delivering
something as a fully managedservice or a network as a
service. I think today you'reseeing a lot of their partners,
delivering those and providingthat that layer of abstraction
and delivering the upfrontassessment and design and, and
so forth. But yeah, there'scertainly going to be innovative
companies like Niall and others,who are going to be pushing
Cisco and driving them, youknow, to compete and deliver
(08:55):
more, more guests operationallyefficient and simpler systems,
for sure.
And I'll add on to that, Anton,you know, if you look at the the
overall picture of Cisco overallSimplicity is the direction that
they're going. And you know,when you look at what they've
announced for the differentbusiness units, openness,
extensibility scalability,flexibility is really the key
(09:15):
message to drive in the home.Give me another example of how
they're making the company alittle bit more user friendly, a
little bit more simple. Changingrebranding the name of ETSI the
emerging technology andincubation team to open up from
start to out shift out shift isthe new brand. So there in the
conversation was in we werespeaking specifically around the
(09:36):
service mesh, takes sto and takeCalista T Kalista allows you to
do use sto far faster and easierthan using sto by itself. So
it's it's about providing thatuser experience and making it
easier and simple to use. And Ithink Cisco is making those
changes and they're reacting tothe market dynamics.
(09:56):
Yeah, another great example tothat is also the like 1000 eyes
and what Going there andexpanding, going all the way
down to the WebEx endpoints andso forth. Right? So we're so
much of the talk today about thecorporate network is the
internet, being able to havevisibility into that
environment, but then all theway down to potentially a home
office provides a lot ofbenefits to organizations as
well.
Antone Gonsalves (10:17):
So So in your opinion, Cisco, will be asking
can be as competitive in thatmid market as a as some of these
other companies like thestartups like Nile but of
course, you also have a Aruba?Arista?
Unknown (10:32):
Yeah, there's, I'll tell you this. Anton, they're
certainly going to try it.They've been doing this for a
long time. They've got a lot ofcustomers SMB all the way
through enterprise. And right,they still have over 50% of the
market. So like I said, I thinkit's great to see the innovative
companies coming up right,driving that that unified
(10:53):
networking approach, driving alot of the the AI approaches and
things like that drivingsimplicity, because it's just
going to force you know, Ciscoto, to drive and push back and,
and deliver their own innovativetechnology as well. And
ultimately, it's going to be theend consumer who benefits.
Antone Gonsalves (11:10):
Okay, right. So you bring up a good point, to
a large extent, this is Cisco'smarket to lose, so to speak,
since they they there's such adominant force within within
networking and insecurity. Okay,next are interviews I conducted
at Cisco Live with companyexecutives, JL Valenti, Vice
(11:33):
President of Product Managementfor multi cloud networking and
Jeremy Foster, General Managerof Cisco's server portfolio.
Foster is also involved inhelping customers with their
sustainability efforts. Hi, I'mat Cisco Live, I'm here with JL
(11:55):
Blent. He is VP of ProductManagement for enterprise
routing, SD Wan and multi cloudnetworking for Cisco. JL, thanks
for joining me.
Unknown (12:07):
It's a pleasure. All right. Thank you for having me.
Antone Gonsalves (12:09):
Okay. So, first question I have, I think
we're going to focus a lot onmulti cloud in this session. It
seems multi cloud is the norm,it seems for large and mid sized
enterprises. Lots of options areout there. Start startups that
offer connectivity as a service,public clouds offering Internet
(12:34):
inter inter connectivity, andcome at vendors like Cisco. So
technically, how is Ciscodifferentiating itself? In the
market?
Unknown (12:44):
So that's a great question. And yes, indeed, multi
cloud is omnipresent, pervasivethroughout, you know, this
conference, but last year andthe year before, I think,
obviously, this is only growing.First, you mentioned enterprise
in a mid market. In fact, I canjust also witness that we've
(13:06):
seen it across every singlesingle business can be public
sector, and be DOD. In the US,but also anywhere in the world.
Cloud is, is pervasive cloud iseverywhere. And our clouds can
be you know, some have even whatthey call distributed hybrid
cloud, which obviously, you sayit's multi cloud, it's multi
(13:29):
cloud, so multiple cloudproviders, which Instructure, we
will discuss. But it's alsohybrid, in the context that not
everything is running on apublic cloud, but also in
private clouds, or data centersthat obviously behave as as
cloud and now need to bestitched together as people were
going through an evolutionprocess. In that, you know, in
(13:53):
that vein, so back to Cisco, A,we realized, obviously, years
ago, the criticality of helpingour customers, being able to
harness the possibilities andthe capabilities that are
available to them in the cloud,in the clouds, and do that in a
(14:16):
simpler way. So there's a lot ofcomplexity, I'm sure we'll talk
about that. So what we reallytry to bring is simplify the
whole process of, you know,defining and accessing to
workloads or applications in thecloud can be SAS, that can be
obviously yes, you know,applications, and do that in a
(14:37):
very secure way. And coming withthat you Sorry, I've heard about
the full stack observability.Not only do you need actually to
provision and define that butyou need to assure it so the
whole aspect of visibility ofmonitoring is critical to be
able to detect and respond tosome of the threats, outages
(14:59):
problems that you may encounter.And so we really take a holistic
approach to cloud throughout
Antone Gonsalves (15:05):
Cisco. Okay. So when you when you talk about
simplicity or simplifying thewhat's a very, very complex
environment? How do you? How doyou define simplicity? What does
what does that mean? So readersunderstand
Unknown (15:19):
it? Yeah, that's a very good question. So it means
really first building. So it'sall about architectures and
building constructs,abstractions, that, you know,
pretty much hide the complexityof the underpinning of
stitching, for example, accessto the cloud, making it actually
more normalized, more consistentacross cloud. One of the key
(15:41):
areas, obviously, for what Irun, I wake up, you know, you
mentioned enterprise writing andSD Wan is, you know, the first
first step was, hey, how do Iconnect users or branches? To
the cloud to different clouds?How do you do that, when you
have, obviously, different cloudnetworking? Obviously, the way
the constructs work on a jurordifferent from AWS different
(16:01):
from Google GCP? So how do youcreate a level of abstraction,
and a level of automation, thatgives people the comfort through
workflows, etc, to really beable actually to deploy at
scale, in a matter of reallyminutes or hours, not days, and
(16:22):
actually keep it, you know,fluid, flexible, because there's
a lot of dynamicity on top ofthe cloud even more. So now,
those applications are, youknow, spinning up, you know,
this horizontal scale, verticalscale, horizontal scale. So
you're spinning up between newcomponents, you have DevOps,
trying to bring new products tomarket to applications all the
time, how should we NETOPS keepup with all of those aspects is
(16:45):
very critical. And so you needthat level of automation, you
need a level of integrationbetween the different parts.
Antone Gonsalves (16:51):
So when you talk about needing an
abstraction layer, what isCisco's abstraction layer.
Unknown (16:57):
So those abstractions are really related to, for
example, the construct in thecontext of SD, when we call
cloud on ramp is really beingable actually to, you know, to
assemble all of the parts thatwould create automatically, the,
what we call the tunnels, thefabric, the the overlay, and
(17:18):
make it very simple, so thatusers to applications,
applications were in datacenters before now they are in
clouds, they might be now evendistributed in the same cloud in
different regions, you mighthave a database somewhere, and
you may have actually thedefault application in a
different place. How actually,do you create a construct
(17:40):
through actually, you know,those, the tunnels where we
could the fabric, so you try toabstract that to a level of
fabric and define policies thatyou can apply to say, hey, if
those sites are those users wantto access those applications,
this is what the policy shouldbe they can they cannot, they
should be able to have, youknow, you can define SLA s in
(18:02):
terms of what you expect interms of latency, you know,
jitter, packet loss, etc. To beable actually to have the system
automatically with its ownintelligent, determine what is
the best path to get to thatapplication, which can be
obviously replicated in multiplezones, as you said, to maintain
actually, the reliability, we dothat behind the scenes means
(18:23):
obviously, there are a lot oftechniques. DPR is one of the
techniques where we grow NP. Sothey do control planes that we
have very sophisticated to beable to, you know, centralize
the routing and define theroutes and dynamically as we
take into account, the policies,the situations of maybe
bottlenecks are happening at anygiven time in the network, etc.
So there's a lot of complexitythat we encapsulate bottle up
(18:46):
and make that available as asimplified pretty much workflow
and instead of constructs tocustomers.
Antone Gonsalves (18:51):
Okay, so given the complexity of their cloud,
do you do you find that most ofyour customers are turning to
manage service providers? Ifthey want Cisco technology,
they're depending on yourpartners, your managed service
providers? help them deploy us?Yeah, I
Unknown (19:08):
think so. Again, there are many different breeds,
breeds of managed serviceproviders. There are breeds of
managed services that were bornin the cloud. So they those ones
are, are new. They were builtmaybe 10 years ago, and you
know, or from the club 2008 20years ago, but 15 The thing is,
(19:29):
the more classic managed serviceproviders that connectivity
managed service provider, willcommunication service providers
to beat ups. Yes, indeed theyare. They've evolved they have
capabilities to be able actuallyto help those customers to the
cloud, most likely today's it isactually still a branch or site
to cloud and obviously, more andmore people work from anywhere.
(19:52):
So how do you get also users toclouds from you know, you know,
hybrid work type of environment,but more and more Those
applications then set us ondistributed, and the its data
center to cloud one cloud andmulti cloud, etc. And you start
seeing, you know, like aI mlapplications, data residing in
one cloud. The Tensor flow, forexample, from Google reside,
(20:15):
obviously, the GCP environment,how do you bring those together
in a very secure manner as well,because the data that is
underneath is very proprietary.And so being able actually to
create, you know, that thosecapabilities for, for those
customers require systemintegration, and managed
services, so that, you know,clearly a number of customers
(20:39):
don't have the people don't havethe skill set, or can keep up
with the, the evolution and thetide of all applications,
including SaaS applications thatare combined with that at times,
whenever he says, folks andServiceNow, in SAP, etc. So you
have those compositeapplications more and more, the
(21:00):
large one, the very large, or,you know, there are a number of
customers or a large data set,hey, I can do that myself. I
manage my network, I've managedintegration with the cloud, and
that's fine. And, and you have alot of obviously those, you
know, enterprises salary, themid market, but also the large
ones that went out in thisbusiness. That's not what we do.
(21:21):
And yes, indeed, we rely uponone or multiple partners, some
actually divided even theirmanager receives in regions and
obviously stitch that together.So you gotta beat off of
everything. And that's whatmakes also this market quite a
bit interesting.
Antone Gonsalves (21:35):
Sure. And it makes sense to me, given the
complexity, the scarcity oftalent, I would think that
enterprises are becoming moredependent than ever before on
their technology partners, aswell as service providers.
Absolutely. JL, thank you verymuch for your work. Appreciate
your time. Thank you. Hi, I'mhere at Cisco Live with Jeremy
(22:01):
Foster, General Manager of CiscoNetworking. Compute. Jeremy,
nice to see you.
Unknown (22:08):
Nice to be here. Okay.
Antone Gonsalves (22:10):
So I wanted to start off our conversation on
hybrid cloud, it seems that thehybrid clouds evolve and where
it's becoming possibly lessabout the data center and more
about the edge, as we'recompanies are saying, are
deciding that the data, it'd beeasier to start fresh at the
(22:33):
edge, rather than take legacytechnology and somehow adapt it
to the cloud. So what are youseeing? And what's your
observation
Unknown (22:43):
as to spend a lot of time with customers, I think,
you know, over the last sixmonths, and even at the show a
lot of talk about, you know,hybrid cloud and customer
settling into the fact that it'snot going to be everything in
one public cloud, where you runall your infrastructure to your
point, you have a lot of datacenters that are out there,
where you've already madetremendous investments as a
business. And then you have thisemerging edge edges, this huge
(23:04):
spectrum of things, right, thathave all different kinds of use
cases, whether that's a retailedge, or whether that's an
enterprise edge where you wantto try and do things, you know,
along new use cases likeinference at the edge, where you
want to get the as close to thewhere the data is being
generated as possible with thetechnology. And those are almost
like taking a mini data centerand building it out in lots of
(23:24):
locations. And so the theme fromcustomers is coming back is just
we want to live in this hybridmode. But that generates a ton
of complexity, because now Ihave to figure out how to
maintain my assets on prem, dothe things that I want to do
with my applications in newapplications, oftentimes, that
are developed in the cloud. Andthen obviously, go after these
new use cases that my businesswants to do. So it's, it's quite
(23:46):
a bit of things for them to dealwith. And that's really what
we're focused on is trying tohelp simplify that experience,
and help them operate in auniform manner across all those
different areas.
Antone Gonsalves (23:54):
And it turns out, sustainability is something
I want to ask you about, becauseit seems to be top of mine of on
everyone, tech vendors as wellas as well as their customers.
And, you know, when you when youtalk sustainability in the data
center, you're you're addingcomplexity to an already very
(24:15):
complex environment. So how isCisco helping customers address
that problem?
Unknown (24:21):
I think it is top of mind. It's good, because it's
good for the world. But it'salso good for business in terms
of helping customers be focusedon how can they optimize for
those environments and go afterreplacing those legacy
environments with newertechnology, and the newer
technology really helps drive alot of the power of that. And
that's one way that we can helpright, we can build the best,
most sustainable infrastructurein the industry. Certainly from
(24:46):
a compute perspective, that'snot anything new for us. We've
been doing that since 2009. Andif you take a typical customer
scenario, like one of the largefinancials that we were working
with, they have 500 servers inone area, their data center.
Those processors were a couplegenerations old Replace that
with UCS acts as 134 serversinstead of 500. So drives out,
not only 31% of the power drivesout 39% of the maintenance
(25:11):
costs, it drives out of 66Roughly percent of the actual
physical space that it wastaking up. Right. So there's a
lot of business benefits togoing after sustainability. And
I think the other area customersare struggling that we're trying
to help out with is the data.We've tracked the data inside of
a data center since forever ago,right? We've had power strips
(25:31):
that could report back to thefacilities people, how much
power are we using in aparticular area or zone of the
data center. But a lot of thatis data that is typically with
those operational teams, andmaybe not with the line of
business folks that are nowtrying to get an assessment of
how are we doing towardsreaching those ESG goals? So
we're trying to leverage thingslike intersite, where we are
collecting data across thatserver state and tracking it
(25:54):
over time. So customers canunderstand how much power Am I
using three months ago? How am Idoing towards reaching those ESG
goals? And then ultimatelyoverlaying that with where is
this power coming from? Is itcoming from solar? Is it coming
from wind? Is it coming fromcoal, and then helping them
understand what thatenvironmental impact is as they
reduce the power that they'reusing?
Antone Gonsalves (26:15):
Right? I believe Cisco, one of the
announcements in is Cisco Livewas the addition of
sustainability metrics withintheir nexus, guess. Fabric. Yes,
management fabric that wasinteresting. So is
sustainability when it comes toservice is always an upgrade
process to customers have toalways buy new servers or
Unknown (26:39):
not? It's a great question. And one other angle to
sustainability is how are youdoing the most with what you've
got. And so there's alldifferent kinds of features that
we've put into our solutions,whether it's dynamic power,
capping multiple power zones ofhow we lay out our UCS x
chassis, that are activelysettings, you change in software
to change how the power bus isworking, so that you can look at
(27:01):
the needs of those servers basedon what the application is
asking those servers to do. Anduse less power. In other words,
change how redundant yourrequirements might be, the more
redundancy you push out there,the more power you're going to
need. But you can actuallyoptimize those types of things.
So we're trying to make thosevery easy to expose those
capabilities for customers, anduse those, I'd say, in addition
(27:21):
to that, there's a lot of newfeatures and capabilities around
optimizations from say, Intel,AMD that we can put into our
software platform to let themleverage those features really
easily and at scale, becausethat's the that's the trick,
right? You want to be able tosee how are those applications
behaving? And how are thoseservers behaving, and then say,
Great, these are greatcandidates, for example, to use
(27:43):
those database optimizationsthat Intel just released, which
will then help increaseperformance and reduce the
amount of power that they need.So you can see,
Antone Gonsalves (27:50):
and cooling is a big is a is a big problem with
with servers. What is, you know,it's also a huge expense. So
what are the solutions forcooling? What is What do you how
are you addressing,
Unknown (28:10):
cooling, it's that's a real complex one in that, the
first thing we can do isoptimize our systems to deal
with in cool only the areas saywithin like our at UCSF chassis,
for example. It's this big, butif there's a lot of power
demands in a certain area wherea certain process or on a
certain servers going. And wehave over 20 cooling zones that
we've engineered in there with,I can't even remember how many
(28:31):
temperature sensors, but we wereable to actually dynamically
change where the fans and howmuch they're driving. So we use
the least amount of powerpossible to actually change that
in real time without thecustomer having to do anything.
So you can really optimize thosepower states there are cooling
states there. When you get tolooking into the future and say,
How are we going to leveragethings like liquid cooling,
there's a lot of impacts to howthat touches a customer, right?
(28:53):
You can, you can put liquidcooling and things like closed
loop liquid cooling that willstart building into some of our
servers when we when we need to.But right now, we don't need to.
But we will, as these newgenerations of processors
consume even more power. Andthat's an easy one for customers
to buy because it just sitsinside of the server. And then
you know, as you get furtherdown the line, looking at how do
(29:13):
we do liquid cooling across thethe data center is something I
think a lot of customers aretrying to figure out, because
it's operationally verydifferent than what we've done
in a data center. It requires alot of physical changes to how
they how they've actuallyplumbed their data center to get
water to the individual backs ofthe racks. And I think that's a
challenge we'll be tackling overthe next five years.
Antone Gonsalves (29:34):
I would change topic for a second. I wanted to
ask you about it your your takeon digital processing units dpus
Because that's being introducedinto servers to offload from the
work from the CPU. Do you seethat as something that could
dramatically change thearchitecture of data centers or
(29:58):
not?
Unknown (29:59):
I think Over time that will. And the thing about GPUs
is it's in our enterprisecustomer base is there's
different people with differentmarkets, right GPUs have already
dramatically changed how datacenters are designed in CSPs,
cloud scalars, and high endservice providers that
understand how to take thosetechnologies and have specific
enough needs at large enoughscales to use them. But it's not
(30:21):
something today that youraverage enterprise account, say
your healthcare organizationthat has a really great, awesome
data center has had to go pulloff the shelf and start using
those technologies. So I thinkas we make those technologies
more easy to consume, peoplewill use those technologies more
readily in say, the enterpriseat scale. And that'll be the
(30:41):
kind of changes that we see overthe next three to five years.
Antone Gonsalves (30:44):
Right now. Anyone who is primarily a cloud
technologies, someone a lot or alarger enterprise, large,
Unknown (30:51):
large scale type applications, are people who
have large data centers, becauseit's just for the same reasons
you're talking about, how do weincrease sustainability,
increased performance, GPUs area different way of doing
processing that can helpincrease the performance and
take, take all that workload offthe CPU, so you get to do more
with less.
Antone Gonsalves (31:09):
And obviously, the management tools have to be
there also don't have to change.Okay, so the, you know, the
cloud in the data center, how isyour data centers are shrinking,
because more workloads are goinginto the cloud? How else? Is the
data center changingarchitecturally? Because of the
(31:32):
because of the cloud?
Unknown (31:33):
Yeah, I think a couple of things, you look at the spend
in the industry over the lastfew years, clearly explosive
cloud growth, but on premrevenue spending is flat. And
it's not, it's not this binarything where everybody said, we
spent this much money in thecloud and all that money in the
data center instantly goes away.It's we all consume more
applications, we all generatemore data, we all do all that
stuff in our daily lives morethan than we did five years ago.
(31:55):
And so there's also the factthat people run their business a
lot of times based on what's intheir data center already. And
then it's not super easy toalready sometimes cost effective
to lift and shift all thoseapplications to the cloud. And
that's really why customers aresettling in on this hybrid cloud
mode. Because much like we do init, I think what we've seen over
the last years is people wantingto really aggressively go
(32:17):
towards cloud, because it's agreat new tool, it's extremely
operationally efficient, I canmove very fast and it sounds
like, you know, it offloads alot of that operational burden
on on a customer to make allthose things happen underneath
deployment of applications. Butthen the reality is, in some
cases, it's not free, there isno free, there's no free
(32:38):
workload, there's no free lunch,if you will. So then customers
are now saying like, every otherIT problem we try and solve, how
do we optimize around this?Well, I gotta look, and I've got
these assets on prem, I can runsome things on prem infected
customers with decent scale,maybe I can do that application
on prem, or I don't need to movethat application because my cost
profile will be better here.Every application is different.
(32:58):
And so that's, that's, you know,kind of a customer by customer
basis. But what we see is, wesee people balancing that out
and really trying to figure outwhere am I spending? Where can I
get the most for my dollars thatI'm spending being in the cloud
or beat on prem?
Antone Gonsalves (33:11):
Yeah, it's all about defining the problem. And
then and then solving it.Jeremy, thank you very much.
Unknown (33:17):
Thank you for your time, sir. Likewise.
Antone Gonsalves (33:22):
That wraps up this week's show. Thanks for
watching. I'll see you nextweek.
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!
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com