Are data centers becoming less relevant? ZKast with Mike Bushong Juniper of Juniper Networks

0:00 [Music]

0:05 welcome to zcast everybody i'm zscaravel

0:08 from zk research as always i'm your host

0:10 i'm joined today by

0:12 mike bushong vp of cloud ready data

0:14 center at juniper networks mike why

0:16 don't you say hi to everybody

0:18 hey thanks for having me today and hello

0:20 everyone out there

0:21 yeah so we're gonna be talking uh data

0:23 center which uh you know it's a pretty

0:25 exciting topic there's always lots of

0:27 stuff going on in that area before we

0:29 get started though i do want to give a

0:30 shout out to e-week my media partner all

0:33 seacasts are done in conjunction with

0:34 the ewiki speaks program so thanks to

0:37 e-week uh mike now

0:39 you work and live and eat and everything

0:41 in data centers um what's going on with

0:44 data centers right we've had this

0:46 massive

0:47 number of people work from home

0:49 we've moved more and more applications

0:50 to the cloud and so i think there's some

0:52 conventional wisdom out there that data

0:55 centers are becoming less relevant but

0:56 uh you know tell us why that's not the

0:59 case

1:00 yeah i think the the whole data center

1:02 thing um when cloud came in right the

1:04 whole idea was the cloud is going to

1:06 wipe through and take out all of the

1:07 data centers that everything essentially

1:09 moves to amazon

1:11 microsoft google oracle whoever

1:13 i just that hasn't been true right i

1:15 think what most people are looking at is

1:17 kind of these hybrid environments

1:19 certainly workloads are moving to the

1:20 cloud i don't want to say that i'm like

1:22 a one of the people who says that cloud

1:24 isn't happening it's absolutely

1:26 happening

1:27 but i think what's what's really going

1:29 on is there's a bit of a bifurcation in

1:30 the market um

1:32 for companies that are maybe at the

1:34 lower end they could they might look at

1:35 cloud first or cloud only because their

1:37 needs are you know fairly

1:39 constrained right they've got a kind of

1:41 a well understood set of applications

1:43 for everybody else they've got kind of a

1:45 bunch of different applications and the

1:46 whole idea that merely lifting and

1:48 shifting an application from on-prem

1:50 into the cloud was going to yield some

1:52 huge financial impact that hasn't really

1:55 happened and so it's left people kind of

1:57 in this environment where they've got

1:58 you know the you know what they've got

2:00 on prem they've got you know new

2:01 emerging you know let's say sas

2:03 workloads and they're leveraging cloud

2:05 and what this leads to is not just you

2:07 know sort of these hybrid environments

2:08 but but hybrid and multi environments

2:10 right on-prem public public the net of

2:13 all of that is that you know data center

2:15 all of a sudden is no longer about the

2:17 four walls that you're kind of managing

2:18 and the equipment that you have in it

2:20 it's kind of this operational domain

2:22 this administrative domain that extends

2:24 out across different different you know

2:26 areas in the network and then with all

2:27 of that

2:28 that makes data center hot again because

2:30 it's all about data center operations i

2:32 think we're actually going through a bit

2:33 of a data center renaissance frankly and

2:36 then you mentioned the work from home

2:37 thing right you get more people

2:39 consuming more applications trying to do

2:41 more things in a distributed way it's

2:44 like that's actually built up a bunch of

2:46 demand for the services that are being

2:48 offered at a data center so it's not

2:50 just that operations and innovation is

2:52 alive it's that kind of the current

2:54 environment that we find ourselves in

2:55 it's really stimulating growth in this

2:57 overall space then we see that when you

2:59 start looking at you know kind of how

3:00 companies are doing who are in this this

3:02 data center proper area

3:04 i like what you framed it up because

3:06 data center maybe is now become

3:09 a term to describe operations versus a

3:12 physical place uh but i do think that

3:14 you know that uh you know i think we

3:16 analysts and you know i write for a lot

3:18 of public media tend to over swing the

3:21 pendulum sometimes we say everything's

3:23 moving to the cloud or everything's

3:24 moving mobile or whatever and that's

3:26 just not true um i do i will say though

3:29 that every business i talk to wants a

3:31 cloud model that doesn't necessarily

3:34 mean public cloud in fact my research

3:36 shows that 95 of companies will adopt a

3:40 hybrid cloud there are just certain

3:41 scenarios where things will run and by

3:44 the way the 95 numbers enterprises so i

3:46 do think you're right small businesses

3:47 can go all cloud but when you talk about

3:50 larger enterprise they have specific

3:51 workflows data sovereignty issues they

3:53 got regulatory issues compliance issues

3:56 and you just can't move everything to

3:58 uh the

3:59 public cloud in fact somebody asked me

4:00 the other day do i see a day where

4:03 everything's in public cloud and my

4:05 answer was

4:06 maybe but i'll long be retired so i

4:08 don't care

4:09 when that happens now the other big

4:11 disruptor in data centers has been this

4:14 thing that we call edge and this has

4:16 given rise to what i call

4:18 uh and gardeners in terms of distributed

4:21 cloud which describes an operating model

4:23 to your point

4:25 where you have one

4:27 logical cloud environment that spans

4:29 public clouds private clouds and edge

4:31 locations and that's fundamentally

4:33 different because historically what you

4:34 had with data centers well they were

4:36 centralized centers of data

4:39 right but now you've got these

4:40 distributed environments so can you you

4:42 talk a little bit about edge and how

4:44 that's really evolved the role of the

4:45 network and data centers

4:47 um yeah so if we take cloud just

4:49 generally i mean cloud is is you know i

4:51 think it's really an operational

4:52 construct um it's how do you how do you

4:55 manage things the whole idea of moving

4:56 to different consumption models and

4:59 and you know being able to spin up

5:00 applications when and where you need we

5:02 look at you know the kind of classic

5:04 cloud it was the idea that we would

5:05 bring users and data to the application

5:08 right things were hosted centrally and

5:09 then you know you could sort of reroute

5:11 everything into these central locations

5:13 what we see with distributed edge is the

5:15 kind of changing that a little bit let's

5:16 take the applications let's take the

5:18 data and push it closer to the users

5:20 in some cases it could be because

5:22 there's you know performance

5:23 requirements or latency requirements it

5:25 could be

5:26 um you know things like 5g are actually

5:28 changing the whole expectation of user

5:30 experience and and what users are

5:32 demanding and so it's no longer

5:33 sufficient to just route everything

5:35 through these national data centers what

5:37 you've got to do is essentially push the

5:38 data centers out closer to the users

5:41 now this is actually not that different

5:43 and i'll say spiritually right

5:44 architecture it's very different but

5:46 spiritually it's similar to this whole

5:47 idea of scale out right instead of

5:49 having you know these big massive things

5:51 what i want to do is create smaller

5:53 components and pull them and sort of

5:54 distribute the load across a larger

5:56 number of elements

5:58 when we do that right we trade off the

6:01 performance piece um

6:03 we get performance advantages but what

6:05 we pick up the cost of that is

6:06 essentially you know distributed

6:08 operations it's all the administrative

6:10 things that you have to do to manage

6:11 these devices well it's true in data

6:13 center as well if i move away from a

6:15 large national data center and i replace

6:17 that with dozens

6:19 you know maybe hundreds of

6:21 of you know distributed data centers and

6:23 this distributed edge what i pick up is

6:25 an operational burden and so what we see

6:26 with with the whole move to distributed

6:28 edge is it's a it's a shift and it's

6:30 emphasizing once again the operational

6:33 tenants of cloud which brings us back to

6:35 kind of the you know the opening

6:37 comments around cloud it's how do i get

6:39 cloud-like operations so i can reliably

6:41 and repeatably deploy applications

6:44 to preserve a consistent user experience

6:47 when my applications and the domains

6:49 over which i have to manage them are

6:51 distributed in all kinds of different

6:53 physical locations

6:54 um with a you know diversity of

6:56 connectivity a diversity of applications

6:58 like that's really the game in this

7:00 whole distributed edge space yeah in

7:02 fact i would argue when you talk about

7:04 operations a lot of businesses

7:06 haven't really moved to true cloud

7:09 operations i think when you think about

7:11 if i pick up a legacy workload and i do

7:13 a lift and shift and public cloud to a

7:14 public cloud i maybe change the location

7:17 of that but i'm not really changing the

7:19 way i operate right so i don't get the

7:21 benefit of the scale out and things like

7:23 that and so this

7:24 you know now what i actually think is

7:26 we're actually moving to an era where we

7:29 do have cloud operations and that's

7:30 giving rise to true multi-cloud and i

7:32 use the word true multi-cloud because

7:34 historically when people did multi-cloud

7:36 they were using multiple centralized

7:38 clouds but to me that's not the same as

7:40 multi-cloud where i could have one

7:42 operational construct across

7:44 all my clouds so uh you know i do think

7:47 the environment does become

7:50 um

7:51 more you know a little more you know

7:53 complicated here and so that certainly

7:54 changed the role um you know of

7:56 operations and security and things now

7:58 um you recently made another a number of

8:01 announcements right uh

8:03 you know the uh through uh uh the

8:05 abstract position that that juniper made

8:07 can you can you just walk us through

8:09 those

8:10 uh yeah so i mean just as a brief

8:12 reminder of what abstra does right so

8:14 this whole idea of abstract is intent

8:16 based networking um it's the idea that

8:18 rather than working configuration

8:19 statement by configuration statement you

8:21 ought to be able to declare

8:23 top down here's what i want my my data

8:25 center to do

8:27 so we did we took that intent model and

8:28 we've extended it actually to the edge

8:30 so we've said that you know there are

8:32 architectures that are different than

8:35 these centralized data centers right

8:37 typically called collapsed fabrics where

8:38 you you don't necessarily need a whole

8:40 bunch of devices use a couple of access

8:42 switches maybe some some tours

8:44 um and what we're doing then is

8:46 essentially extending the benefits of

8:47 intent-based networking to those types

8:49 of of deployments this is big right as

8:52 we start thinking about you know smaller

8:53 data centers uh use cases like iot

8:57 we start looking at uh things like telco

8:59 cloud where you've got you know let's

9:01 say carriers that are you know

9:02 distributing their data centers out in

9:04 support of their 5g services

9:06 there's a whole different kind of

9:08 collection of things that we can cover

9:10 within that that basic space um that's

9:13 that's a big part of this overall

9:14 announcement right extending the

9:15 goodness of abstract to new deployments

9:18 um beyond that we did some stuff on the

9:20 security side which we can talk about in

9:21 a bit um and then

9:24 of course you know when you pick up

9:25 something like intent based networking

9:26 it's really is a different operating

9:28 model which raises a question especially

9:30 like in your enterprise clients right it

9:31 raises a basic question you know there's

9:33 this thing that i want that is different

9:35 than what i have

9:36 how do i get there

9:38 and so we put together some migration

9:39 services that are really aimed at how do

9:41 we transition from environments as we

9:43 know it today to environments as as we

9:45 want them tomorrow

9:47 yeah i wouldn't drill down the security

9:49 side because juniper obviously has a

9:50 very big

9:51 uh security business as well you know

9:53 when you when you think of a traditional

9:55 data center it was a closed environment

9:58 you could control a lot of things uh

10:00 just by locking down

10:02 the walls right so now that we're moving

10:04 more and more things to the edge uh

10:06 we're using public clouds we're kind of

10:08 mixing public and private um

10:11 that that really makes curing a lot more

10:13 complicated so can you you talk about

10:15 those challenges and then tell me what

10:17 you know how juniper's addressing those

10:19 sure i mean you just spelled it out if

10:21 i've got

10:23 different types of users accessing

10:25 different types of applications all run

10:27 in different types of data centers that

10:30 are all supported by different types of

10:31 vendor equipment and different types of

10:33 vendor software

10:35 like what's the problem

10:36 um i mean that like like that's that

10:38 should be easy

10:40 it should be it sounds amazing right

10:41 like

10:42 what that sounds like to me by the way

10:43 is job security so for all of you out

10:45 there like that's you know good good

10:47 stuff

10:48 um

10:49 i think the real challenge is you know

10:51 and this is when you talk about true

10:52 multi-cloud i mean you mentioned it it's

10:54 the idea that i want uniform operations

10:56 and i want diverse underlying you know i

10:59 guess infrastructure

11:01 what apps just trying to do is provide

11:03 unified operations i want to do things

11:05 in a common way regardless of

11:08 the underlying infrastructure regardless

11:10 of the vendor regardless of the the

11:12 particular domain type regardless of the

11:14 underlying you know users and

11:15 applications whatever let me provide

11:17 kind of a consistent way of applying

11:19 policy across this diverse environment

11:21 what we're doing in this particular

11:22 release is we're taking this idea of

11:25 group-based policy the idea that we have

11:27 policy and control that's applied at

11:29 like a tenant or an application level

11:31 and then we can and essentially what

11:32 apps does is it abstracts that policy

11:34 from the underlying devices

11:36 then we go and apply that policy across

11:38 whichever devices are are required in

11:40 order to support that that overall

11:42 policy structure and then

11:44 we'll validate that the policy that's

11:46 deployed is essentially um working as

11:49 intended and the way we do that right so

11:52 so underpinning this idea of intent this

11:54 this declarative model this way of

11:56 pushing let's say configuration and

11:58 state information into devices i'm

12:00 underpinning that we actually create a

12:02 model we say here's the the network as

12:05 it's intended and then we deploy a bunch

12:07 of sensors and we say here is the

12:09 network as deployed

12:11 and when the as intended and the

12:12 deployed match we give you a big green

12:14 light everything is good

12:16 when as deployed and um and as intended

12:20 uh do not match then we can flag that

12:22 there's a problem so let me give you

12:23 kind of an easy example

12:25 you pay you uh you create policy for a

12:27 particular application but of course

12:29 multiple applications run across these

12:32 different you know

12:33 on the uh the switches the data center

12:34 fabric what if somebody else makes a

12:37 different policy instruction that

12:40 overlaps with the policy for for

12:42 application a and application b what we

12:44 want to do is flag when there's

12:46 overlapping policy we want to be able to

12:47 tell you when something is different

12:49 than as intended so that when you get in

12:51 this complex environment of you know

12:53 many to many for everything we're trying

12:55 to do

12:56 let us tell you if there's if there's

12:58 conflicts so you can resolve that

13:00 conflict ahead of time the worst time to

13:02 to possibly find out that your policies

13:05 are not working as intended

13:07 is in the middle of your ransomware

13:08 attack

13:09 right so segmentation becomes a a key

13:13 mitigation you know activity when you

13:15 think about some of the security stuff

13:16 going on you can't afford to find out

13:18 that it's not going to work at the point

13:20 of at the point of attack at the point

13:22 that something's gone wrong what

13:23 aperture's doing is not just

13:24 provisioning it but telling you up front

13:26 you know is your policy assured is it

13:28 working the way you expect it to work

13:31 uh thanks for that answer mike uh i know

13:33 one of the big value propositions of

13:36 appstra

13:37 and there are other intent-based

13:38 solutions but i believe appstore is the

13:40 only real multi-vendor one so can you

13:43 can you talk a little bit about why and

13:45 how that's grown in importance obviously

13:47 you know everybody wants to be able to

13:48 choose the vendor they use but

13:50 i think historically haven't been able

13:52 to so what have you seen there from your

13:53 customers and how they've been

13:55 leveraging that

13:57 sure um i mean so yes the idea behind

14:00 intent based is that if we can work from

14:02 a declarative model over the top then

14:03 we're abstracted naturally from the

14:05 underlying devices this means that we

14:07 are in a very good architectural

14:08 position to support multi-vendor

14:11 why does that matter um look everybody

14:13 is these days talking about supply chain

14:16 if you want to um

14:18 if you want to be able to order devices

14:20 from vendor a or vendor b or vendor c

14:22 but your operations are sort of fixed

14:24 then you're in a position that you can't

14:26 you actually can't do that what app

14:27 store allows us to do is to say you know

14:29 here's my data center the way it's

14:30 constructed i don't care about the

14:32 underlying devices my teams are trained

14:34 my services are deployed applications

14:36 are deployed um

14:38 regardless of what the underlying

14:39 vendors are then depending on what the

14:42 lead times are then you can go and take

14:43 advantage of you know vendor a vendor b

14:45 vendor c

14:46 now look supply chain won't always be a

14:48 constraint right we we think that you

14:50 know that things will improve at some

14:51 point

14:52 but at that point then the question is

14:54 do you want your vendors to compete for

14:56 your business if you want economic

14:58 leverage for instance then keeping your

14:59 vendors competing for for incremental

15:01 capacity is a is a huge thing that that

15:04 gives you um flexibility right it could

15:06 be that they're competing based on

15:08 feature functionality it could be based

15:10 on you know price and performance but

15:12 this gives you the flexibility to make

15:13 the decisions that you need to make

15:15 um you know so you can avoid these you

15:17 know seven nine ten year type

15:19 engagements where you say like i'm

15:21 committed

15:23 longer than i'm going to be in the role

15:24 i'm committed to this particular vendor

15:26 what we're essentially doing is

15:27 unlocking that and we've seen huge

15:30 traction here i mean we have um we've

15:32 got folks who come and they say look you

15:34 know appreciate that juniper has

15:35 hardware however i've got cisco or

15:38 arista or dell or sonic or you know or

15:42 cumulus or juniper whatever i've got

15:44 different suppliers you know can you

15:46 solve my operational challenges

15:48 and we've got accounts today where it's

15:50 it's literally it's it's you know it's

15:52 pure software over the top

15:54 breathing additional life into an

15:56 existing investment which allows them to

15:57 get more out of what they've already

15:59 spent and then gives them a plausible

16:01 path to go and extract additional value

16:02 over time

16:04 yeah it's intended in the whole shift of

16:06 software that took place in the network

16:07 industry years ago was supposed to allow

16:09 better network uh vendor

16:11 interoperability and it never really led

16:13 to that in fact i've had people describe

16:15 uh their vendor lock-in situation as

16:16 being a little like hotel california

16:18 where you can check out but you can't

16:20 ever leave and uh so you know unless you

16:23 wait the best time to do that there

16:25 is always with cio turnover right so

16:28 that can be long uh cycles now um i i

16:31 did want to ask you about

16:32 the acceptance that you're seeing with

16:34 intent-based networking and autonomous

16:36 operations i know when apps are first

16:39 launched when i when i talked to it

16:41 people about this concept of autonomous

16:42 operations to use the technical term it

16:44 gave them the heebie-jeebies i mean a

16:46 lot of them were very concerned about it

16:48 and uh so what do you what do you see

16:50 now and i'm assuming you know work from

16:52 home has played you know a role in in

16:55 getting to look at your product but are

16:56 you seeing more acceptance of it and

16:58 what are some of those low-hanging fruit

17:00 use cases where if people were to

17:01 adopted they could take advantage of it

17:03 faster

17:05 yeah i think people in general are

17:06 getting more comfortable with automation

17:08 i still think it's a long journey

17:10 you know the processes that surround

17:11 automation are sometimes immature

17:13 especially in some of the enterprise

17:14 type accounts

17:15 so i think there's there's still work to

17:17 go and and move forward

17:19 but in terms of like basic use cases i

17:20 think you start looking at how quickly

17:22 can i stand up you know additional

17:23 capacity how do i expand how do i add

17:26 you know new racks how do i add new

17:27 servers how do i add new pods depending

17:29 on the size of the data center you start

17:32 looking at this whole idea of mean time

17:33 to innocence we have a

17:35 good customer in the retail space where

17:37 they have an online shopping platform

17:39 and at the end of each day they've got

17:41 you know there's items that sit in the

17:44 um in their in their shopping carts that

17:45 don't get closed and of course the the

17:48 uh the the teams come back and say you

17:49 know we didn't get we didn't convert

17:51 these to sales there's a network problem

17:52 these things timed out you know what's

17:54 the mean time to innocence they spend

17:56 their time literally every day

17:58 you know having to prove that there

17:59 wasn't some network issue that caused

18:00 these purchases not to go through what

18:02 we can do on the abstract side is by by

18:04 demonstrating that the network is

18:05 working as intended and and that there's

18:07 no issues it's like that mean time to

18:09 innocence is merely

18:10 at that point it's like a dashboard view

18:12 you know here things are working as

18:13 intended that's a that takes an

18:15 operational practice that would have

18:16 been manual triggered by people

18:18 forced to forcing other people to go in

18:20 run a bunch of commands and prove that

18:22 things are working as you want them to

18:24 now that becomes just part of the system

18:26 you know it's it's this idea that that

18:28 we can

18:29 provide information provide insight um

18:32 like that that's how you get to some of

18:34 these autonomous operations it's not

18:35 about removing keystrokes and making it

18:38 like oh i can type in some command

18:39 faster it's about you know helping you

18:41 make decisions much faster in this case

18:43 you know the decision is you know do i

18:45 need to go and troubleshoot the network

18:47 or are things working again the way

18:48 they're supposed to be working so maybe

18:50 you should go and look at other things

18:51 that might be causing um a problem on

18:53 the retail side like that's a the amount

18:56 of time that that that saves it's not

18:58 just the the time it takes to execute

19:00 commands but it's the time it takes for

19:02 for people to coordinate across

19:03 functions that's how you drive a lot of

19:05 value you know not just into the it or

19:07 the networking team but into the

19:09 enterprise at large

19:10 yeah i've likened the transition to

19:12 autonomous networking a little bit like

19:14 the transition to autonomous vehicles

19:16 right nobody

19:18 except maybe elon musk would jump into a

19:20 car with no steering wheel no controls

19:22 and no driver right but we're all

19:24 willing to drive a car now and actually

19:26 like features like parallel park assists

19:28 and lane change alerts and some of those

19:30 things and so you know baby steps will

19:32 get us there and one day down the road

19:34 we'll have fully autonomous operations

19:36 but i do think there is some trust that

19:38 needs to be built up here so a last

19:39 question for you mike if i if i look at

19:41 juniper's kind of bigger

19:44 messaging right

19:45 juniper talks a lot about experience

19:47 first networking and uh so can you tell

19:50 us what that is and then how does

19:52 abstract fit into that

19:55 yeah i mean the whole idea behind

19:56 experience first is this it's the notion

19:58 that the things we optimize for are

20:00 changing where it used to be all about

20:02 you know

20:03 price and performance it was about you

20:05 know transport moving packets like

20:07 that's table stakes right when when you

20:09 can move the packets the question is

20:11 what do you do from there it's no longer

20:12 about is your network up right and we

20:14 user experience is the new uptime it's

20:17 not

20:17 it's just not the case that having you

20:19 know

20:20 your high availability is enough to

20:22 declare success um you've got to look

20:24 and say well you know are things working

20:25 the way that you want them to work the

20:27 applications function in the data center

20:29 adequately to support the business

20:32 and if that's the case then you have to

20:33 do more than you know merely provision

20:35 things so that they so that connectivity

20:37 is in place you've got to be able to

20:40 um assure that that things are working

20:42 continuously the way they're supposed to

20:44 you've got to make sure that the

20:45 applications have um you know not just

20:47 the the connectivity but they've got you

20:49 know all the things that the drive

20:50 overall experience um so

20:52 for instance what does latency look like

20:54 within the application um you know are

20:56 you seeing intermittent drops are you

20:58 seeing

20:59 um you know congestion that's leading to

21:02 to changes in application side behavior

21:04 um if you look northbound into the

21:06 application space you know are we seeing

21:08 things that are happening you know at

21:09 the application side or maybe within

21:10 some of the the vmware you know suite of

21:12 tools and is that leading to the

21:14 decisions that need to be made or

21:16 information that needs to be sourced

21:17 from the underlying network this is like

21:19 it's it's so much broader than just

21:21 merely connecting things

21:22 um the whole role of appstra is

21:24 essentially saying look first let's

21:26 let's make it as simple as possible to

21:28 get the network up and running it's

21:30 table stakes let's do it well but let's

21:32 do it quickly so you don't have to spend

21:34 time on that

21:35 second you know the whole idea of

21:36 reliability is not that you know hey

21:38 i've pushed some configuration out and

21:40 it should be working it's like can i

21:42 guarantee can i demonstrate that it's

21:44 working

21:45 and that extending those workflows

21:47 beyond merely pushing the configuration

21:48 but to looking at sort of the assurance

21:50 and reliability side that's a big push

21:52 towards um as you start thinking about

21:55 you know what are the types of things

21:56 that impact experience we talked about

21:58 the policy overlap okay so you've you've

22:00 set your policy does that mean you're

22:02 good

22:03 well maybe but if somebody else has set

22:05 conflicting policy like isn't the

22:07 outcome that you want that you've got to

22:08 secure um that you have a secure posture

22:11 for a particular set of users whether

22:12 they're applications or tenants or

22:14 whoever it's more than merely pushing

22:16 configuration out it's got to be this

22:18 closed loop validation that's what apps

22:20 are trying to do and if we can do that

22:22 in an environment that's multivenger if

22:24 we can do an environment that's

22:26 multi-domain if we can do that in an

22:27 environment that's multi-technology

22:29 multi-topology like everything becomes

22:31 multi over time if abstract can be that

22:34 unified operations platform

22:36 over the top of all of those multis

22:39 like that's what we're trying to drive

22:40 here at juniper

22:42 all right well thanks for that mike and

22:44 uh you know after leading you to a

22:46 multi-everything world i guess a

22:48 multi-multi-world so uh you know with

22:50 that we're gonna wrap up i think that

22:51 was a great overview of your

22:52 announcements and you know clearly the

22:54 data center

22:55 has never been more important right

22:57 everything is moving to some kind of

22:59 data center be it you know a centralized

23:01 one public cloud or in an edge so uh

23:03 again uh thanks mike luxon from juniper

23:05 network for joining me i'm zeus caraval

23:07 from zcast signing off don't forget to

23:09 click to subscribe and i'll see you next

23:11 time

23:13 [Music]