Get Connected Podcast Episode 6: Power Cloud Automation

0:00 [Music]

0:03 get connected

0:06 you're listening to get connected

0:08 [Music]

0:09 thanks for joining us

0:13 in this series we discuss the various

0:16 issues shaping your industry how the

0:18 changes in supplier ecosystems impact

0:20 the way you work on a daily basis how

0:23 your customers consume the services you

0:25 create and how the industry as a whole

0:27 steps up to the demands of the digital

0:29 marketplace let's get connected

0:36 hello listener let's get connected

0:38 [Music]

0:41 my name is chris lewis

0:43 by day i'm an independent industry

0:46 analyst by night as you should know by

0:48 now i'm a podcaster

0:53 welcome to get connected where we delve

0:56 into the issues shaping the networking

0:58 industry and the experiences we all

1:01 benefit from by being better connected

1:05 in this episode we're talking about

1:07 power cloud automation how to bring

1:10 powerful automation to your edge telco

1:14 and i.t data center clouds

1:16 to enable simplified value-driven

1:19 operations

1:20 in the hot seat alongside me today is

1:23 the senior director of product

1:25 management from juniper's cloud ready

1:27 data center group mr james kelly welcome

1:31 james tell us a bit about your role

1:33 great to be here chris so i've been a

1:35 juniper for a long time 15 years now and

1:38 all kinds of different roles and as you

1:39 said i'm i'm currently in the the cloud

1:42 and data center business group but when

1:44 i started off my journey back in 2006 i

1:47 started off much more technically

1:48 focused as a developer advocate

1:51 and went through roles

1:53 as various specialists focus on cloud

1:55 and devops

1:56 kind of growing from an engineer

1:59 to do more business development i worked

2:01 on such things as the contrail

2:03 acquisition and more recently worked on

2:06 the abstract acquisition but through all

2:08 that time yeah i've spent time in the

2:09 marketing teams

2:11 and uh it's been it's been a long but

2:14 very exciting journey of lots of

2:16 different positions and lots of

2:17 different exciting technology right

2:19 things are always changing in technology

2:20 that's one of the reasons why i love

2:23 working in this field and that's

2:25 certainly true of

2:26 of cloud and automation today

2:29 and it's great isn't it because the the

2:31 issue of of all these topics coming

2:33 together we often talk about now the

2:34 collision of technologies coming from

2:36 the data center the enterprise and the

2:38 service provider you know having had

2:40 that experience in all those different

2:42 if you like more basic engineering

2:43 levels and then marketing and having a

2:45 taste of all the different areas really

2:47 helps you shape your thinking of

2:49 bringing what is a very diverse set of

2:51 services and technologies to bear in

2:53 this emerging cloud environment yes

2:55 hopefully that's the case

2:57 so james cloud has evolved from a dc

3:01 technology through network and

3:03 infrastructure service how are your

3:05 service providers deploying cloud today

3:08 yeah so in terms of juniper's cloud

3:10 business we sell to cloud providers we

3:12 sell to enterprises and we sell to these

3:14 network and communications service

3:16 providers we just call them service

3:18 providers most of the time for short but

3:20 you know your cable close your telcos

3:23 of course you know over time we know

3:25 that data centers have you know trended

3:28 from more centralized to more

3:29 distributed of course today the level of

3:32 automation in data centers has

3:34 skyrocketed

3:36 [Music]

3:37 things like moving from virtualized to

3:39 containerized

3:41 i guess there's a confluence of the the

3:43 application teams and the operations

3:45 teams underneath it

3:47 and in the operations teams

3:49 among the infrastructure silos those

3:51 things kind of coming together this they

3:53 can be orchestrated together

3:55 and in the context of our service

3:57 provider customers that you asked about

4:00 the biggest thing that's happening in so

4:02 far as the the data center space for sps

4:04 today

4:05 is you know 5g is driving a lot of

4:09 modernization of their telco cloud

4:11 initiatives and of course these telco

4:13 clouds are also distributed um all the

4:16 way out to to the radio networks frankly

4:18 i think as you rightly point out there

4:20 you know your your customers are moving

4:22 from from cloud to multi-cloud to

4:24 virtualized you know into the world of

4:26 containers and kubernetes

4:28 how do you contrast what's happening uh

4:30 with your enterprise customers because

4:32 juniper doesn't just cover the sp as you

4:35 said it covers the enterprise side it

4:36 covers the data center so on how do you

4:38 contrast what's happening with the

4:40 enterprises and the service providers

4:42 where that transition within the service

4:44 provider in particular

4:46 is

4:47 certainly perceived to be more slowly

4:49 developing than in the enterprise market

4:51 so the

4:53 the thing that i mentioned i guess i get

4:55 virtualized to containerized we've seen

4:57 that

4:59 that evolution happened more quickly

5:01 in the enterprise space

5:04 so that transition is absolutely being

5:06 driven by 5g but if you go back you know

5:08 it's already seven years now that we've

5:10 got these open source projects like

5:12 kubernetes that are really dramatically

5:14 changing the way applications run in

5:17 data centers and let's you know admit

5:19 for a second that the name of the game

5:21 and the whole purpose of data centers is

5:23 to run applications

5:24 right enterprises have all different

5:26 kinds of applications

5:28 but primarily their endpoint you know

5:31 like server type of applications and

5:33 those have been refactored into

5:35 containers and into different micro

5:38 services of course and the systems like

5:40 kubernetes and its rise and prevalence

5:43 have facilitated all of that

5:46 and along with that there's various you

5:47 know infrastructure changes software

5:49 defined infrastructure changes

5:50 underneath kubernetes that support it as

5:53 well as operational changes on top of it

5:56 in the devops space but if you look at

5:58 how that happened for enterprise

6:01 they were able to adopt that quite

6:03 quickly

6:04 modernizing you know breaking off some

6:06 of their applications and modernizing

6:08 those into that cloud native uh so to

6:10 speak environment first and foremost we

6:12 use cloud native and kubernetes centric

6:15 almost uh synonymously of course because

6:17 the cloud native computing foundation

6:19 was really centered around kubernetes

6:22 and when you look at how service

6:24 providers have adopted that of course

6:27 they didn't have the opportunity to

6:28 easily break off chunks of their

6:30 applications like enterprises did

6:32 and start to modernize them right away

6:34 right service providers kind of

6:37 coalesce around standards bodies and

6:40 organizations that you know talk

6:42 together and figure out what we're going

6:44 to do in the 3gpp space what are we

6:46 going to do for

6:47 for 5g what are we going to do for our

6:50 iran initiatives

6:52 for these next generations of

6:53 infrastructures so they tend to be more

6:56 on this cadence of you know 3g 4g 5g etc

7:00 and those present natural steps for them

7:03 to modernize

7:05 and so

7:06 um with the rise of 5g and kubernetes

7:08 having already been on the scene this

7:10 was the natural next step for service

7:13 providers right to move from virtualized

7:16 to containerized from

7:18 things like virtual machine management

7:20 systems that had been used to

7:23 orchestrate the virtual network

7:24 functions in the 4g era to things like

7:27 kubernetes to manage the containerized

7:30 network functions of the 5g era and that

7:33 transition is still very much going on

7:36 um so they're they're straddling both

7:38 worlds as as enterprises are too legacy

7:40 doesn't ever just

7:42 you know go away but the the other

7:45 challenge that service providers face is

7:47 they don't necessarily have client

7:48 server types of applications that are

7:50 all endpoint micro services as they're

7:53 refactoring their their network and

7:55 their security functions these things

7:56 are often of course waypoint services

7:59 right that are processing traffic um for

8:03 for mobile subscribers for broadband

8:05 subscribers and for business subscribers

8:07 to do all sorts of things but because

8:09 they're they're waypoint services

8:11 because they're also packet processing

8:13 in many cases they also have high

8:15 performance high availability

8:19 all sorts of additional

8:21 requirements

8:23 and a standard of excellence they must

8:26 meet in terms of scale performance and

8:28 reliability

8:29 that uh that don't always need to be

8:31 engineered into uh enterprise systems

8:34 the same way no sure what is really

8:35 interesting about this is that you know

8:37 the the way they had control over the

8:39 environment in the past seems to have

8:41 gone

8:42 you know and they now have to take it

8:43 allowance of what's going on in the data

8:45 center and of course we used to have

8:46 these very centralized data centers but

8:48 now it's been dispersed and distributed

8:50 out to the edge and 5g as you rightly

8:52 say is a major change in that respect

8:54 because it's making the service

8:56 providers the telcos think about not

8:57 only the traditional endpoints and as

8:59 you said they're getting services out to

9:01 those endpoints but the increasing the

9:02 billions of endpoints that you know many

9:05 people at juniper point out are evolving

9:07 as we move into the worlds of iot so

9:09 we've gone from connecting buildings and

9:11 people now to the things and devices and

9:14 therefore i think the point that it that

9:16 came clear from what you just said

9:18 is that the traffic becomes really

9:19 important and in many ways the telecom

9:21 industry is to focus on its inner

9:23 workings and not on what it actually

9:25 delivers for its for its end customers

9:27 or in this case it's partners because

9:29 it's not just about the telcos

9:30 themselves it's about who sits within

9:32 that ecosystem and i think that i think

9:34 that was is that how you see 5g really

9:36 changing the vision from the the service

9:39 provider side yeah um i would agree with

9:41 all of that i mean i'd add also that uh

9:44 you know with 5g and with this

9:46 more multi-cloud era enterprises moving

9:48 to multi-cloud i think people understand

9:50 what the term implies surprise hybrid

9:52 cloud and the use of multiple clouds

9:55 well with service providers you know

9:57 this is all still true certainly in the

9:59 sense of their i.t clouds

10:01 but in their sense of their their telco

10:03 clouds there's a wider distribution

10:06 of multiple compute clusters right

10:08 they'll have much more of this spread

10:11 out across many remote sites many more i

10:14 guess you could say geographically

10:16 centralized or main sites and some of

10:18 these will be running on their own

10:20 on-premises owned infrastructure

10:22 as well as some of the hyperscaler you

10:25 know cloud infrastructure so there's

10:26 hybrid sense in that in that

10:29 sense of transformation of how things

10:30 are working too exactly is that that

10:32 ecosystem change of course is as you

10:34 rightly say it's not just about the

10:36 telcos it's about partnering with as

10:38 those hyperscalers bring things towards

10:40 the edge as you as a as a provider of

10:43 infrastructure and services to them

10:44 bring services to the table so the

10:46 management of it and imagine all of

10:48 these moving parts and let's face it as

10:50 we have

10:51 up to 100 billion moving parts in the

10:53 future or

10:54 building connected things

10:56 hit in the market then we've just got to

10:58 make sure this this efficiency is there

11:00 to deliver

11:02 i'd like to change directly a little bit

11:04 and

11:05 you mentioned that you were involved i

11:06 think in your introduction both in the

11:08 acquisition of comtrail and indeed

11:10 abstract so can we turn to abstra

11:12 juniper added apps to the portfolio of

11:14 products

11:16 what problem does abstract solve and why

11:18 is it important to the customers good

11:20 question and uh happy to pivot to to

11:22 talk about some juniper-centric stuff

11:24 and and tie it back to those

11:26 introductory remarks so

11:28 to answer the question you know quickly

11:30 and simply

11:32 abstra is a

11:34 a fabric management system

11:36 over

11:37 switching fabrics inside of the data

11:39 center now that's sort of a generic

11:41 description of it you know appstore kind

11:43 of made its name and the reason why it

11:45 was very famous

11:47 was

11:48 they the the abstract company when it

11:50 was founded in 2015 they drove a trend

11:52 towards intent-based networking and they

11:55 also had a maniacal focus on usability

11:58 across multiple vendors with a common

12:01 experience so the workflows are

12:03 abstracted across different vendors and

12:06 the fact that they're providing this

12:08 multi-vendor fabric management system

12:10 for different vendors of switches right

12:12 whether it be juniper cisco arista and

12:15 so on

12:16 under one operational umbrella for our

12:20 users our customers

12:22 that's really how they they made their

12:23 names fake and that intent-based

12:25 networking is really the term that

12:27 points to this

12:28 abstraction this facilitation

12:31 of these workflows um to an intent based

12:34 level so less prescriptive step-by-step

12:37 less oriented towards the actual

12:39 configuration and cli

12:41 of the individual devices right the

12:44 fleet of switches that it's managing

12:47 and it does this on a fabric by fabric

12:48 basis

12:50 and

12:50 you're right i said i worked on the

12:52 abstract acquisition

12:53 but one of the reasons why i had the

12:55 opportunity to do that is because

12:57 insofar as product managing contrail

13:01 i i said i mentioned worked on that

13:03 acquisition too

13:04 uh contrail when we first acquired it

13:07 and when it started out its life in 2013

13:10 it was really focused on

13:11 software-defined networking

13:13 for virtual machine management

13:15 orchestration systems namely openstack

13:18 and

13:19 we open sourced that i mean juniper open

13:21 source that product it became very very

13:25 uh well known

13:27 and well adopted

13:29 in the openstack space as a software

13:31 defined networking solution and as

13:33 kubernetes came on the scene we

13:35 introduced it for that as well we looked

13:38 at and dabbled in supporting you know

13:40 vsphere

13:41 but over time as contrail became a point

13:44 of control

13:45 in the data center for our customers we

13:48 realized that we wanted to

13:50 you know add in support for switching

13:53 fabric automation and management

13:56 and so we built that

13:57 as a company into contrail and what we

14:00 realized over time was that

14:03 we would

14:04 be better off for our own internal

14:06 endeavors like from a standpoint of

14:08 engineering testing code quality and

14:11 velocity if we had broken that apart and

14:14 didn't build it in contrail we also

14:16 found that the switching management even

14:18 from a user's perspective for service

14:20 providers and enterprises was best left

14:23 outside of contrail in a separate

14:25 uncoupled piece of software where these

14:28 things can work better together but

14:29 don't necessarily need to be integrated

14:32 there's an interesting point of contrast

14:34 here to like cisco aci for example

14:38 the way that you know cisco has built

14:41 aci their their flagship piece of

14:43 software for data center operations

14:46 they really do have everything under one

14:48 roof in terms of managing multiple

14:50 fabrics you know open stack

14:53 kubernetes and whatnot and we found that

14:56 in taking that approach it was a bridge

14:58 too far for customers and like i said

15:00 even internally it's just it's better to

15:03 have

15:03 separate solutions that the customers

15:05 can choose

15:07 based on their own merits and have a

15:10 company you know that has these

15:12 different pieces of software as products

15:14 but allows customers to use them without

15:17 having to present this monolithic

15:19 solution right without having to force

15:21 the customer to eat all of it

15:23 so with abstra and contrail right we we

15:26 deprecated all of the fabric management

15:28 stuff in contrail we were looking how to

15:30 rebuild that separately

15:32 and

15:33 the acquisition of appstore just

15:35 dramatically accelerated that strategy

15:37 of course and at the same time we're

15:39 able to do things like multi-vendor

15:41 which is something that juniper's

15:43 actually been pretty committed to in a

15:44 lot of ways

15:45 um so it was a really exciting

15:47 acquisition to work on i bet it was and

15:49 i think you touched on several really

15:51 important points there which is that you

15:53 know from a vendor point of view in the

15:55 let's say perhaps in the more

15:56 hardware-centric view of the world in

15:57 the past you of course tried to tie

15:59 people in and it was easier to tie

16:01 people in but actually as we move into

16:03 this more software-centric uh

16:05 nature of the bit of the beast shall we

16:07 say that actually bringing the right

16:09 pieces into it and managing it

16:10 orchestrating it all together

16:12 and thinking about it from the point of

16:14 view of the customer whether that's an

16:15 enterprise or a service provider or a

16:18 data center provider and actually

16:19 allowing them to build on what they

16:21 already have and perhaps even their

16:22 internal expertise you know we spend a

16:24 lot of time talking about the

16:25 technologies and the different vendors

16:27 in there but of course the actual people

16:29 that do that run these things within the

16:30 organizations often will drive the way

16:33 that this evolves so yeah very very

16:34 important point to bring out that um

16:37 that whole point about the the

16:38 multi-vendor nature of it uh and

16:40 actually giving the customer the choice

16:41 and actually managing these moving parts

16:43 as we go forward

16:46 so appstra manages the the fabric

16:50 and contrail manages the overlay in the

16:52 virtual networking

16:54 i think you touched upon this already

16:55 but are they standalone products or are

16:58 they

16:59 working together or are they moving

17:00 towards each other how would you

17:01 describe those two things moving

17:04 together yeah so

17:06 uh a great question and point of clarity

17:08 there you know in so far as evolving

17:11 contrail i talked a little bit about its

17:13 history just so i could tee up um the

17:15 logic behind us acquiring appstra

17:18 these products are today completely

17:21 independent in the sense that there's no

17:23 dependency of one on the other

17:26 and in fact there's no dependency

17:28 of using these products not just with

17:30 each other but also on the juniper

17:31 hardware

17:33 to reemphasize a point that you just

17:35 made

17:36 when you look at what we've done to to

17:39 decouple these products there's also you

17:41 know reasons and

17:43 areas where we have the opportunity to

17:45 reintegrate them

17:47 and that actually is an interesting

17:48 place to talk about for

17:51 for service providers and some of the

17:52 use cases that it benefits to have these

17:54 things integrated

17:56 but

17:57 perhaps what i'll mention first is

18:00 with contrail

18:02 having deprecated the fabric management

18:04 piece one of the things i mentioned is

18:06 that when we built contrail from the

18:08 get-go control networking we focused on

18:11 openstack first

18:13 well when we were looking to rationalize

18:16 contrail and to optimize it what we also

18:19 knew around this time is that our

18:21 enterprise customers and our service

18:23 provides customers were on this

18:24 transformative journey towards

18:26 kubernetes and cloud native and that we

18:28 needed a software-defined networking

18:30 product that was kubernetes first and

18:33 openstack second instead of the way that

18:35 we had kind of developed contrail was

18:37 very open stack centric so openstack

18:39 first so to speak in terms of its use

18:41 cases that it addressed in kubernetes

18:42 second so we flipped that on its head

18:45 and we spent the last 18 months at

18:48 juniper

18:49 redeveloping contrail from the ground up

18:52 in terms of its control plane to be all

18:54 kubernetes native so one of the exciting

18:56 things about this is we've

18:58 gotten deep into kubernetes and we've

19:01 made contrail so much more accessible in

19:03 doing this too one of the

19:05 challenges that our contrail users have

19:08 always had is that if you were looking

19:11 to use it for openstack you needed to

19:13 have your own hardware you know

19:14 openstack is an infrastructure as a

19:16 service system whereas kubernetes you

19:18 know you can run it on your laptop like

19:20 minicube you can run it on the public

19:22 cloud where you have you know on-demand

19:24 elastic and disposable

19:26 infrastructure

19:28 and so now you have the ability to

19:31 to run control with kubernetes and you

19:32 have the ease of deploying it on your

19:35 laptop or on the public cloud obviously

19:36 you can play it on bare metal too

19:38 but um we've rebuilt contrail as an

19:41 extension of kubernetes

19:43 we're getting ready to release that

19:46 version it's a version 22.1 um short for

19:50 2022 first quarter that's how our

19:52 versioning works

19:54 and and yeah so we've been on this

19:56 really exciting journey with contrail

19:58 and completely revamped it at the same

20:01 time as when we we did this abstract

20:03 acquisition so we're now set up for a

20:06 really exciting future where we've got

20:08 a

20:09 fabulously modern lightweight

20:12 purpose-built contrail

20:14 for kubernetes networking to still

20:16 support openstack networking of course

20:18 by the way and we've got appstra which

20:20 is you know

20:22 just killing it in terms of winning

20:23 business right now for juniper switching

20:25 and for juniper fabric operations

20:28 because of the things i talked about the

20:30 fact that it's really easy it's intent

20:32 driven it's multi-vendor so you know

20:35 certainly our customers often do have

20:37 switches from other vendors the ability

20:39 to have one

20:41 point of control and one piece of

20:42 software that they can train on and

20:44 learn

20:45 for those different switching fabrics is

20:47 is very exciting to them and unique in

20:49 the industry

20:51 and then you know to close the loop and

20:54 come all the way back to your question

20:55 about how are abstract and contrail

20:58 evolving together in a sense

21:01 well for service providers

21:03 one of the things that they're doing in

21:05 their kubernetes clusters is they've

21:07 also got

21:09 workloads that need

21:11 sriov

21:13 that's single root i o virtualization

21:16 high performance networking needs which

21:18 means that they cannot use the contrail

21:22 v router which is our forwarding plane

21:23 they effectively bypass it in favor of

21:26 additional

21:27 performance out of their their network

21:29 interface card

21:30 and when they do that

21:33 the virtual network of that server of

21:35 course needs to be plumbed into the

21:38 switching fabric right just like a bare

21:40 metal server would if you have a bare

21:42 metal server and it's got some workload

21:43 on it and it needs to take part in the

21:45 same virtual networks as all of your

21:48 containers and your virtual machines

21:50 that contrail is is networking

21:53 well you know

21:54 appstra is really the system that

21:57 manages those virtual networks in the

21:58 switching fabric and so it's also doing

22:01 that for the s or iov connected

22:03 workloads whether they're virtual

22:05 machines or whether they're containers

22:06 for example and this is a natural point

22:09 of integration between contrail and

22:11 abstra

22:12 to effectively manage the

22:14 the overlay networks that are in the

22:16 switching fabric or evpn vxlan

22:19 overlays with the contrail managed

22:22 overlays in the world of a kubernetes

22:24 cluster and it's created the whole the

22:27 whole environment especially if we move

22:29 it to a service provider perspective you

22:31 know the those flows and putting we talk

22:34 about slices in the in the 5g world

22:36 where we're dedicating infrastructure to

22:38 particular tasks but actually this all

22:40 of this both the the the controller and

22:42 app store then also needs to feed up

22:44 into the other systems into my my

22:46 favorite area of bss oss and itms and of

22:49 course devops in that changing

22:50 environment how are you seeing your

22:52 customers

22:53 look at that effectively bringing

22:55 everything we've talked about together

22:57 to deliver this more powerful this what

22:59 we talked at the very top this cloud

23:01 automation the power of cloud automation

23:03 yeah that's a really exciting topic too

23:06 when you talk about higher order

23:07 automation systems like those

23:09 b slash oss systems when you talk about

23:11 devops and the need for higher order

23:13 automation you think about starts to put

23:15 together the picture of the full stack

23:17 right you you refer to it as like

23:19 underlay an overlay of like abstra

23:21 managing the switching fabric and it's

23:23 the fabric

23:25 automation insofar as the hardware

23:27 switching fabric and then contrail

23:30 managing the workload orchestration

23:32 system on top but of course yeah there's

23:34 all of these higher order operational

23:36 tools for service providers and

23:38 enterprises on top of that

23:41 and you mentioned devops one of the

23:42 interesting trends that both enterprises

23:44 and service providers are also getting

23:47 pulled into i would say more on the

23:49 service provider front but led by

23:51 enterprise in the same way that

23:53 enterprises led the kubernetes adoption

23:56 and service providers kind of saw all of

23:58 the the transition from openstack to

24:00 kubernetes from virtualized to

24:01 containerize

24:03 the service providers are also seeing

24:04 this transition in the enterprise from

24:07 server administration and more manual

24:09 based administration to automated

24:12 operations

24:13 and devops and with that you know one of

24:16 the most enabling technology

24:20 technology underpinnings i guess you

24:21 could say for devops is infrastructure

24:23 is code

24:24 and kubernetes lends itself very well to

24:27 doing this what we've been focused on

24:29 with appstra and contrail is that when

24:31 you're actually automating your

24:33 switching fabric or when you're

24:34 automating your networking for your

24:36 kubernetes cluster that you also have

24:38 the ability to take these

24:40 software-defined networking systems

24:43 and drive them as code right because

24:45 they are software defined infrastructure

24:47 and if you've got some other

24:48 infrastructure like your servers or your

24:50 kubernetes cluster itself

24:53 and you're setting that up and you're

24:54 setting up your applications

24:56 as code then you want to do that in sync

24:59 with all of your infrastructure as well

25:02 and of course

25:04 that gets into the benefits of all of

25:06 the site reliability

25:08 engineering the sre benefits

25:11 that accrue from infrastructure as code

25:14 so

25:14 what this means in simple terms is that

25:17 you know things like abstract and things

25:19 like contrail

25:21 they are intent driven they are

25:23 declarative

25:24 in their configuration and they do lend

25:27 themselves very well to infrastructure

25:28 as code in fact with contrail being

25:31 refactored to be kubernetes native

25:34 because kubernetes

25:35 is often you know consumed through the

25:38 kube ctl in its first forms of people

25:41 getting used to it and then configured

25:44 all through code

25:46 we actually have this adjacent product

25:48 called contrail pipelines i shouldn't

25:50 really call it a product it's a it's a

25:52 feature add-on of control networking

25:54 that allows it to be driven completely

25:56 through a

25:57 continuous integration continuous

25:58 deployment a ci cd pipeline and totally

26:01 as code so yes juniper is absolutely

26:04 doing things in this space we're

26:06 absolutely integrating through the

26:08 product apis with higher order systems

26:11 whether those things are chat up systems

26:14 incident management sorts of systems to

26:17 to reference your itsm systems and

26:19 working with system integrators and oss

26:23 bss vendors to build the integration

26:26 between our network automation systems

26:28 and their

26:29 wider purview of automating the entire

26:32 you know 5g solution let's say for a

26:34 service provider

26:36 james i think that is a great way to

26:38 draw together the little journey that

26:39 we've just been on from from what this

26:41 cloud automation is looking at appstore

26:43 and contrail and then how 5g is changing

26:46 it and bringing it together under that

26:48 story of how we need to be thinking more

26:50 about the way code is driven and the way

26:52 we develop things and not be obsessed

26:54 with some of the former if you are if

26:56 you like compartmentalization

26:58 of different technologies so james it's

27:00 been an absolute pleasure having you on

27:01 get connected thanks for joining us

27:03 great to be here chris

27:06 so there you have it listener the vast

27:08 experience that james brings to the

27:09 table from working across engineering in

27:12 different areas within the juniper

27:13 environment up in the acquisition of

27:15 abstract and comtrail but most

27:17 importantly bringing the story together

27:20 how the service provider environment is

27:22 changing

27:23 to bring together the benefits of

27:25 automation at all those levels from the

27:27 underlying fabric the underlay to the

27:29 overlay and then upwards and outwards

27:32 into the operational systems

27:34 environments the bss environments and of

27:36 course that new area of devops which is

27:38 helping to shape the future of the

27:40 service provider industry

27:44 now all the remains is a few thank yous

27:46 thank you to juniper for the opportunity

27:49 to dig into this subject of power cloud

27:51 automation and thank you to you too for

27:54 listening i look forward to talking to

27:56 you again soon

27:59 you've been listening to get connected

28:02 you've heard from us now we'd love to

28:04 hear from you

28:05 tell us what you think via twitter using

28:08 the handle at juniper networks

28:11 and if you like what you've heard

28:12 remember to tell your colleagues and

28:14 friends thanks for listening to get

28:17 connected