A Better Way to Build Autonomous Transport Networks

0:01 foreign

0:05 [Music]

0:06 the biggest challenge I see is the time

0:09 to First outcomes because automation is

0:12 all about delivering business outcomes

0:14 and the reason why we see that problem

0:17 is the DIY approach where much of these

0:21 automation

0:23 Frameworks that our customers are

0:26 building in in-house while they are

0:28 elegant and holistic it actually takes a

0:31 lot of time to get them off the ground

0:33 because it's like bringing all the

0:36 different components from multiple

0:37 vendors applying a layer of own

0:40 automation on top this takes years to

0:44 get off the ground

0:45 it requires a lot of skill stuff that

0:48 you know many of The Operators are

0:50 running into a shortage of this actually

0:53 require it costs a lot more because

0:56 there are many many overheads to manage

0:58 and above all you know it's about the

1:01 ROI of your investment so when you see

1:04 the first outcomes it allows you to

1:06 invest further into the next use case

1:08 however this DIY approach doesn't lend

1:11 itself to seeing those first outcomes

1:14 quickly enough we believe there is a

1:17 need for a much more pragmatic Approach

1:18 at Juniper we think there is a better

1:20 way

1:21 we think that the future of Van

1:23 automation is cloud delivered

1:25 and that not only makes business sense

1:28 but technical sense as well

1:30 from a business perspective it's

1:32 actually 40 percent cheaper than any

1:35 kind of DIY approach

1:36 and in terms of time to First outcomes

1:40 it cuts the time by half

1:42 even better once you have got the system

1:45 rolling you're able to introduce

1:47 Innovation 70 faster than any kind of

1:50 DIY so the whole idea is to start small

1:53 but go fast go use Case by use case

1:56 deliver one show the ROI move on to the

2:00 next above all with SAS we can put an

2:04 easy button to it there is no overheads

2:07 it's much easier to operate by your

2:11 Workforce particularly in context of the

2:13 skilled Workforce shortage that many of

2:16 our customers are facing today so it

2:20 makes a lot of sense to start with SAS

2:23 delivered automation for bad Networks

2:25 from a technical perspective when we

2:28 think of AI Ops and that's very critical

2:30 as we proceed into networks that are

2:33 becoming far more complex

2:34 multi-technology multi-domain

2:37 AI Ops we have also learned you know

2:40 that cloud is the way to do AI Ops

2:43 because not only you need data sets at

2:47 massive scale and of course these are

2:48 anonymized data sets but in order to

2:51 build much better models and train them

2:54 around problems you actually need to

2:56 have that variety of data sources so

3:00 this is also about Collective Innovation

3:02 and speed because now you have

3:03 discovered one problem in North America

3:05 say around optical cable failures and

3:08 you can build a model in two weeks

3:10 around that and now it's available out

3:12 of the box to another customer in Asia

3:14 Pacific he doesn't even have to you know

3:17 he can spot that issue before it happens

3:19 and before it impacts experience so

3:22 putting it all together I believe that

3:24 starting small start with one use case

3:27 do a SAS approach for it but then go

3:31 fast from there it's really exciting to

3:34 see that automation is not an

3:36 afterthought we're actually seeing from

3:38 our customers that automation is part of

3:41 these rfps as and when they're building

3:43 new networks so that's in really

3:45 encouraging

3:46 in terms of the problems it solves and

3:49 the different use cases I broadly see

3:51 them in two categories the first one is

3:53 all around operational efficiencies

3:55 things that the customers have to do

3:57 every day which is where automation can

4:00 help make it faster simpler more

4:03 error-free and here the use cases

4:05 include things like you know software

4:07 upgrades activation of devices Services

4:10 links things like network maintenance

4:14 audit compliance inventory management

4:16 and so on

4:18 the second category of use cases is all

4:21 about improving Network and service

4:23 performance you need complex automation

4:25 systems to deliver it and examples

4:27 include things like service assurance

4:29 things like Network resiliency build

4:32 outs things like autonomous capacity

4:35 optimization things like traffic

4:37 engineering closed loop automation

4:39 Network analytics root cause analytics

4:43 going to Predictive Analytics and mother

4:46 of all use cases which is Network

4:47 slicing so it's actually absolutely

4:51 fantastic to see that automation is now

4:54 top of the agenda of all our operators

5:02 [Music]