Light Reading Webinar: The Future of Network Automation for Tier 2/3 Service Providers

0:00 in today's webinar the future of network automation for tier 2 3 service providers sponsored by Juniper Networks

0:07 before we begin I will cover a few housekeeping items on the left hand side of the screen is the Q a if you have any

0:14 questions during the webcast please type your question into the Q a box and submit your questions to our speakers

0:20 all questions will be saved so if you don't get to answer you we may we may follow up by email

0:26 at the bottom of your audience consult our multiple application widgets you can use if you have any technical

0:32 difficulties please click on the yellow help widget here you can find answers to common questions

0:37 a copy of today's slide deck is available for downloading the green resource list widget near the end of today's presentation

0:44 please take one minute's complete to save it as open your feedback is extremely helpful

0:50 an on-demand version of the webcast will be available about one day after the event and can be accessed using the same

0:57 audience link that was sent to you earlier today I'd now like to turn the event over to heavy read and Senior principal analyst

1:03 Sterling Perrin thanks hello and welcome everybody to

1:08 today's webinar sponsored by Juniper Networks a future of network automation for tier 2 tier three service providers

1:15 really an exciting webinar that is the culmination of a large project that we embarked on with

1:22 juniper several months back and uh it was a survey based project actually one of the largest surveys

1:28 we've ever done and we'll do a readout today and I'll be joined in helping me with the presentation and also a great

1:34 help throughout the survey project Kevin Landry is joining me he's senior product marketing manager with Juniper Networks

1:41 hi Kevin um welcome and I will let you give a little bit of an overview of the project

1:47 itself but welcome to today's webinar sure thanks so much Sterling yeah so you

1:53 know who am I I just wanted to mention that and Sterling I were just talking about this I've been in the industry for

1:59 over 20 years and I started my career always in network management and network

2:05 automation so I started as a software designer designing these systems and

2:10 finally went into deployment type of uh for OSS roles and then move myself into

2:17 partnership integration roles and slowly went into marketing where you know I've

2:22 been there since the evolution of network network management towards sdn for the

2:28 when as well as intent based networking towards Network Automation and that's where we are today and I'm excited to

2:35 tell you that I'm uh heading up a new role at Juniper Networks and product marketing focusing specifically on

2:42 targeting our tier two and three CSP audience and that's why I'm involved in

2:47 this survey and let me me just move on to the next slide so what we're doing at

2:53 Juniper is we're actively working with tier two and three service providers so

2:59 we understand the changes that are needed to drive Network automation adoption

3:05 um and make it you know more of a compelling value prop for our tier two

3:10 and three customers so what we did is we took a look at this problem and we

3:17 wanted to confirm that the conversations that we're continuing to have with our customers are indeed representative of

3:24 the whole market in terms of the tier twos and tier threes we wanted to understand their unique challenges and

3:32 to see if any of their pain points were felt more acutely so really when we asked Sterling and heavy reading to do

3:40 this independent survey it was really to confirm a lot of the conversations that

3:45 we had to see if we really understood our customers and those challenges is and how to resolve those with our

3:52 Solutions um so the the survey could be downloaded through the webinar tool and I'd

3:59 encourage you to do that and it consists of you know the results of 217

4:05 respondents and every single one of those respondents was working in the

4:11 ipmpls OR IP integrated Optical transport networks and involved in

4:16 either deploying managing or purchasing a solution for Network automation

4:22 Engineering Management operations or assurance in addition if you want to learn more

4:29 about the Juniper portfolio for Network automation our Paragon automation as a

4:34 service there's a couple of videos as well that we've included in the resource section for you to view

4:41 now if you take a look at the types of outcomes that we've been discussing with

4:46 our customers that we wanted to confirm uh really they rest in three different groups so the first is every you know

4:55 customer every csb wants to be competitive it's crucial to gain competitive advantage to gain market

5:03 share and to create customer value you want to be able to do this uh in a way

5:09 that is fast and that's where automation plays a role so you're able with automation to innovate faster so that

5:16 you could bring those differentiated services to Market faster and by differentiated Services it could be

5:22 something like improving the net promoter score that reduces churn by delighting your customers with a

5:29 superior quality of experience for instance as well it's important from a business

5:36 perspective to be able to contain costs and lower the capex and Opex expenses as

5:42 well as lower the bar in terms of what is needed for expertise so that more can

5:48 actually get actively involved in uh in running the operations and using

5:54 automation secondly in terms of the people to be

5:59 able to be more productive is so important you know how many times do we

6:04 feel like we're doing repetitive tasks again and again in network operations and you almost feel like you're turning

6:09 into a robot well that's where mistakes can happen we can introduce errors when

6:15 you know we're not totally able to focus because we're being pushed to do too much so this is where automation can

6:23 help us do more or do better with less we're able to simplify and speed up those repetitive tasks by making them

6:30 consistent and standardizing the processes to make them repeatable and reliable and with that we reduce the

6:37 human errors and also we're able to scale operations so that as you know

6:42 there's more demand you're not necessarily going to have to increase your Staffing base you're going to also

6:49 be able to have more Junior resources people working on alongside senior

6:55 resources is and lastly I talked about you know the idea of differentiation and a juniper

7:02 our true north is the experience first networking so what we're doing with this

7:08 in terms of automating that is we're giving the ability to not only improve the experience for you as an operator

7:14 but to also improve the experience for end users by increasing up time and

7:20 reliability we're able to offer that you know high quality of experience that's

7:25 differentiated um and we're able to do complex tasks using Technologies like AI that humans

7:32 can possibly do and that's going to be able to allow you to reduce your mean

7:37 time to know and mean time to resolve um and you know do Predictive Analytics

7:43 and that stuff so that you could optimize capacity on demand and make sure that your network is absolutely a

7:49 reliable and Performing as best it can so when you think about these three

7:55 pillars and Sterling is now I'm going to hand it back to him to get into the amazing results that you've seen through

8:03 the survey uh keep in mind these these um

8:08 and we're going to be able to see if this survey actually matches what we had originally thought and with that I'm

8:15 going to uh pass it this morning all right thanks thanks Kevin

8:22 for that introduction so here is the agenda for the next 50 minutes

8:28 um we'll go through the survey data I'll do that taking a look at General automation Trends timelines

8:35 um then getting specifically into Cloud delivered automation then I'll hand things back to Kevin to cover the future

8:41 of network automation for tier two tier 3s um getting into some of Juniper's learnings and recommendations for the

8:49 market I'll wrap up with conclusions and hopefully we'll have about 15 minutes 10 to 15 minutes at the end for your

8:56 questions so the survey itself

9:01 um so I'll present very quickly here the demographics um you'll get the full

9:06 report you'll get the full demographic so the one thing well two things I'll draw your attention to here

9:12 um as Kevin said 217 operators globally is is quite a big

9:17 number um you can see the geographic mix we cover all the regions in pretty good size to get a good sample from from

9:24 everywhere um the other important point is revenue you know you can Define tier two tier

9:30 three lots of different ways we Define it by Revenue so we said under a billion dollars in annual

9:36 revenue is a tier two tier three company over a billion dollars annual revenue we

9:42 excluded from the survey so this is different for most of our surveys in that there's no operators of over a

9:48 billion dollars um of the under a billion you can see the mix here and we're able to parse the data and get some insights by different

9:55 sizes within the tier two tier three Community but to give it a concrete metric we said uh you know by Revenue a

10:02 billion uh also on demographics that the types of providers this is a basically a

10:08 telecom survey so you know cloud is not really represented we didn't want it really in this satellite as well this is

10:15 what what you'd expect from a Telco Market the mix of um of types of operators and then you

10:20 can see the job functions there let me let me get into some of the the data and

10:27 some of the key key findings that that we came across so we asked right right up front uh to what extent do you agree

10:33 with a set of statements um and and we ask them these statements and they could agree on a spectrum what

10:40 I did here is aggregated the combination of strongly agree and somewhat agree which are the two top top options and uh

10:47 looked at what are the most what are the statements that resonated most with the audience 81 percent

10:54 of of csps I believe automation is needed to remain competitive and

11:00 increase customer value that was the strongest statement uh of the mix so quite powerful and endorsement for

11:06 automation 77 percent uh believe automation is needed to cope with an increase in scale of operational

11:13 demand so a big focus on operations I called those two out here but there's also several others resonated quite

11:20 strongly actually you know statistically tied but but in percentage terms one

11:25 percent higher 78 percent uh saying that artificial intelligence for it or or AI

11:31 or and or artificial intelligence is essential for the future of automation so very strong resonance

11:39 um among the audience for using AI tools to improve and help their Automation and

11:45 also a statement which comes through elsewhere in the survey as well as increasing complexity of the network is

11:52 is driving the need for automation uh where are network operators or this

11:58 tier two tier three audience where are they in terms of uh um automation today uh we would say

12:05 they're at a fairly early stage 82 percent of respondents are still

12:10 operating with either a low level of of automation or or no essentially no

12:16 Automation in the network 52 as you can see on the right uh saying they have

12:21 some network uh management tasks that have been semi-automated so bits of Automation in the network 30 percent uh

12:29 which is a fairly large percentage almost a third of the audience are basically manual operations today

12:34 command line interface and other Legacy software tools really kind of manually uh punching things into the network to

12:41 to manage it uh so you know we we expect this to a large degree we thought it

12:46 would be fairly early stage uh one other point I'll make is we do just kind of scanning through the audience we also

12:52 have a number of large operators that tune into these webinars um

12:58 a lot of this data Willa will apply and Kevin will agree will apply to to large operators as well we do find that large

13:04 operators are also at a fairly early stage of network automation as I go through I'll kind of try to point out

13:10 where some of the similarities and differences are between this audience and the in the larger audience or more

13:16 General audience but you know there's a lot of similarities in in the data and also some very specific

13:22 differences where tier two and tier 3s just have unique needs so I'll try to point those out as we come across them

13:28 um primary drivers for Network Automation in this survey improving productivity stood out as the top driver

13:34 for automation 42 selecting that option about the highest on the list of the

13:41 many options they were given other top ones include improving customer experience increasing reliability and uptime

13:48 enhancing innovation a couple points here I would make um

13:54 increasing reliability is very important across our surveys we've seen that not just for automation but we've seen that

13:59 Rising as as an issue for Network operators so it's not a surprise really that they're going to look to automation

14:04 to help with that kind of overarching mandate for their networks going forward what we find in automation generally is

14:11 is there are kind of two two major buckets that come to the top in terms of motivations and it shows up here as well

14:18 you've got the internal or network focused things like improving productivity or improving Network

14:24 reliability and then you have the more external customer Centric types of benefits things like

14:31 um you know getting Services out faster or in this case um uh and ability to to innovate and get

14:39 you know basically get get Innovations out to Market faster so it really covers a broad spectrum

14:44 um and it's not one or the other and it's not kind of one bucket really more than the other it's really across across the board

14:51 in terms of surprises uh from the survey I was surprised a bit by mitigating

14:59 skilled labor shortage resonated less than expected that's down toward the

15:04 bottom of the the list we hear so much about skilled labor shortages

15:10 um and it's actually fairly consistent in in survey so I don't think it's an outlier for automation but it it doesn't it's

15:18 not a driver as as maybe some people would think it factors in and we'll talk in a bit but it's not the primary driver

15:23 the other one that was interesting particularly um in light of what I just talked about with AI Ops aiops itself or AI as a

15:32 driver is not is not a driver for automation

15:37 it is actually not inconsistent uh with with what I just showed a couple slides back though what we're seeing is that

15:43 operators reviewing AI as a tool to achieve goals not as an end itself and I

15:49 think this is quite consistent with that view so operators reviewing AI is a tool it's critical for automation we just

15:55 showed that automation itself is critical to improve productivity and customer experience

16:01 Etc but operators are not going to invest in AI for the sake of AI itself there's a lot of hyper course in AI but

16:06 this is really I think the way the market really needs to look at how these are tools that play into other goals

16:15 um the last one I'll present here before I hand things back to Kevin for a moment is challenges to network automation uh

16:23 type of question we ask a lot in surveys we asked here what are the biggest challenges to achieving your network

16:29 automation strategy and the top two on the list for the tier two tier threes

16:36 are high costs or lack of budget and complexity of integrating

16:41 um of integrating into existing systems these are are quite strong challenges I

16:48 would say also in the greater audience I think much of this would apply to to the tier ones as well

16:53 other specific factors that that come to light I think that are very tier two three specific are are a lot of things

17:01 around Staffing internal resistance to change uh was quite interesting uh to have but also insufficient number of

17:08 staff so lack of staff is preventing them from achieving their automating automation goals as well as lack of

17:15 internal expertise so these are all things related to staff and actually one of the things we did is we segmented the

17:22 data a bit further and looked at what about the tier threes and the smaller tier twos how did they answer

17:28 this question and we Define the smaller tier twos as 200 million dollars in

17:33 annual revenue or less and when we did that the lack of uh staff insufficient

17:40 staff numbers actually Rose to number two on the list only behind lack of budget so the smaller the operators go

17:47 the less budget they have and the less staff they have to achieve what they want to achieve with automation but even

17:55 in sort of resource heavy companies the the um the complexity is always there

18:02 across the board we're finding uh with that let me hand things back to Kevin for a moment to provide some additional

18:07 insights and then we'll we'll move on to the cloud portion yeah thanks uh Sterling so I just wanted to comment

18:14 what's what's exciting about the survey is that really if you look at uh you

18:19 know the the content like what we're taking a look here specifically for tier two and tier three is there's never

18:24 really been a survey in the industry that is focused on this segment so firstly that's really exciting and just

18:32 to comment on a couple of the things that you mentioned as drivers so you did mention the uh you know the whole idea

18:39 of the job labor shortage is not surprisingly not uh ranking high and to

18:45 me I think that might be because strategically you know maybe management doesn't really identify it that is the

18:51 primary or highest primary driver to invest in automation but the people that

18:56 are doing the work actually know that it is a barrier that's blocking the adoption of automation so I'd like to

19:03 highlight that you know when you take a look at some of the findings where you said that there was insufficient number

19:08 of Staff or a lack of expertise and that's significant around 26 percent that's something to think about as well

19:16 you know automation is coming with tons of challenges as we've seen the idea of

19:21 higher costs and complexity and the the ability for automation to be able to

19:27 tackle that challenges I think Paramount but what we're seeing is Eve because of

19:34 these challenges even with successful deployments there's still a lot of

19:40 delays it's taking way too long 68 of csbs think that automation is not

19:46 deployed fast enough and when the automation is successful at being

19:51 deployed a lot of times um you know this is only 40 percent of

19:57 the case because 60 of do-it-yourself automation projects have been seen to

20:02 fail and that's from another third party research that we've done a year ago

20:08 uh something else of interest is that you know we talk about it not being deployed fast enough well in certain

20:15 circumstances for do-it-yourself automation it takes them over three months we're talking about many several

20:22 months of work to be able to get these projects rolled out uh 92 percent of

20:28 respondents actually said that and 51 said that uh the third party automation

20:34 projects that they take on take even longer like double the time six months or longer so you know these are the kind

20:41 of challenges that are really preventing I guess tier twos and tier threes to be

20:47 able to adopt Automation and deploy it and take advantage of this to gain that competitive Advantage speed the

20:53 productivity and everything that we've seen in terms of differentiating the quality of experience so you know with

21:00 that we wonder is there a solution for that and that's what we're here to talk to you about in the rest of the webinar

21:07 a solution that's going to be able to make things faster for you and be able to do things that meet your requirements

21:13 at your pace and with that I'm going to switch over to Sterling and let's cover the next uh chapter because I think

21:20 it'll be very enlightening yeah yeah we'll get into the cloud uh just a couple comments on on your your

21:27 insights here which which of course I agree with uh wholeheartedly um yeah I also found it very interesting

21:34 um the third party automation projects take longer than do it yourself uh so you

21:41 know you think of motivations for for using a third party to speed things up it's that was some motivation doesn't

21:47 seem to be working the other the other point on labor is uh yeah we do these surveys and you get things that seem to

21:54 be contradictory and really what you got to do is spend a lot of time figuring out what's what's going on there it's

21:59 kind of easy to dismiss the data is bad but it tends not to be the case in this case what the data has shown is

22:07 labor shortage is not a driver to automate your network it's not the primary driver but when The Operators

22:13 have made the decision to automate what they're quickly finding out is that they

22:19 don't have enough people so labor is becoming the problem when they make the decision to automate and that that showed through in the survey

22:25 uh on on several levels so yeah I would not dismiss that Labor's not an issue in

22:30 the market it's just where where it's um where it's showing up and here it's operators and even as they get quite far

22:37 down the the path of automation they seem to be finding even more that the complexity and the lack of staff is is a

22:44 problem uh let's let's dive into the cloud here which which was quite one of the really interesting aspects of the

22:50 survey and Juniper was without a doubt one of the early movers and seeing this opportunity so

22:57 um kudos to you guys for doing that we've expanded out quite a bit in the survey based on some of the earlier findings in Cloud

23:03 uh we we asked operators here how interested is your organization adopting at least some automation

23:10 applications delivered from the public cloud and we did specify public Cloud here um to to to highlight that

23:18 um and this was really I would say one of maybe maybe the most surprising find from from the the study 93 of operators

23:27 surveyed are ready to adopt automation via the public cloud meaning if you look at the pie chart

23:34 either six either likely to adopt the 67 or 26 who said There definitely will

23:40 adopt that's the 93 uh really um surprising a huge endorsement

23:45 um I think for the the potential for software as a service for for the tier

23:51 two three uh operator audience we haven't asked this question specifically I think to larger tier ones but so I

23:58 don't have a good comparison but clearly for for this audience um very interested in in public Cloud

24:04 Model not just Cloud but but public so quite strong and um

24:09 there are challenges uh and this is what we expected would weigh in if you think

24:14 about the different challenges that people raise to cloud-based delivery of services

24:20 all of them you know resonated here when we asked them what what would block adoption

24:26 concerns about customer data housed in the public Cloud interesting also though

24:32 that they're also concerned about customer data housed in a private Cloud both of those were on the list network data traversing

24:39 the internet without full encryption so kind of getting into areas of the network where the operator is not in

24:45 control of the data fully encryption would factor in there data sovereignty

24:50 as well um leaving country or region or just areas where the the data is not managed so we would

24:57 expect these things to to weigh in and they did but I go back again to the fact

25:04 that even though operators are fully aware of these issues 93 said they're at

25:10 least likely to adopt public clouds so they're confident I would say that these challenges can be

25:16 addressed um they'll look to suppliers they'll look to their own internal um

25:22 but they are clearly want to move forward in in a way that they're aware of of what's out there right I don't

25:28 think it's a naive response that um 93 say they they want to invest

25:35 different factors that would support moving to um cloud-based automation so we asked

25:43 what are these different factors would support um the use case to adopt Network

25:48 automation tools cloud-based and high availability guarantees versus

25:54 um compared if delivered with higher availability guarantees compared to on-premise 63 so looking to Cloud to

26:02 deliver availability uh reliability essentially in excess of

26:07 what they can do in a in an on-premise solution I already mentioned up earlier that this reliability availability issue

26:14 is really Rising across across the board and surveys so they're seeing Automation and cloud-based specifically could play

26:20 a role there we put some other things in there with very specific metric numbers uh reducing Opex cost by 50 or more

26:28 which is quite a big number 53 saying yeah we would we would uh invest uh for

26:33 that type of benefit total cost of ownership by 25 or more would invest

26:39 um as well as as greater security so lots of reasons would support moves to

26:45 cloud-based automation if if the solutions could could do you know what what stated here on on the slide so

26:52 quite quite strong benefits that they look to achieve

26:57 uh the final bit uh before I hand things over to Kevin to kind of move into the

27:03 the solutions and and um kind of where we go from here based on the data we

27:08 asked about AI a couple questions on on AI clearly it's a hot Trend in the industry and we asked a very specific

27:15 aspect of AI would you consider an AI enabled software as a service solution also trained on other leading networks

27:23 we know that when data is trained more broadly it can have better outcomes more

27:29 accuracy our performance quicker to to come to the right decision there's

27:34 clearly um you know risks to doing that and so we we asked specifically about AI

27:40 trained on other other networks and uh we really found this was a I would call

27:45 it a qualified endorsement 83 saying yes uh it's not a strong a question as as we

27:52 asked about public AI so we we didn't I mean public Cloud we didn't say are you planning to we just said would you

27:58 consider this this is really the first question we've asked along these lines in a survey 83 say they're at least

28:04 willing and open to considering it it doesn't mean they're absolutely committed but they're open to the concept again I'm not showing it on the

28:12 the slide but we asked about you know different challenges to to um

28:18 AI based on and trained on other networks including their own data and uncomfortable levels you know com how

28:25 comfortable they would be doing that um the uh

28:31 you know the percent that said they're against using their data to train other networks was not it resonated it was

28:37 about a third saying that's a concern but it wasn't the top by any means so it's in the mix but it's not a show

28:44 stopper same thing with lack of trust in AI it was at the bottom of the list so operators are not saying they don't

28:49 trust AI they're not saying they're unwilling to put their data into the mix and they're at least open to

28:56 um using AI That's trained on on other networks so quite um you know fairly strong given the

29:02 early stage I think this is an error that that we will certainly investigate more and some strong opportunities not

29:08 just SAS based but also bringing in AI as well with that let me hand things

29:14 back to Kevin to kind of talk through where we go from here

29:20 great thanks Sterling so what we've seen my takeaway from that is that you know

29:25 93 of csps are are saying that they're likely to adopt this or definitely will

29:32 adopt Cloud delivered automation that's really compelling and all those risks

29:37 the issues that they had in terms of cloud are pretty much non-issues because you know the the actual reward outweighs

29:44 the risk and a lot of the tier one Cloud providers like AWS Google Cloud you know

29:50 Microsoft right is there they have a lot of these issues in terms of availability security and so forth taking care of so

29:58 you know I think that as we go forward you know this is really the vision for

30:03 the future of network automation for tier two and tier three is and likely uh

30:08 similarly um AI is going to play a critical role in terms of moving the needle to be able

30:15 to do more with less and being able to do those things that are not humanly possible and giving you that edge for

30:22 competitive advantage and we believe that that technology needs to be delivered as the cloud to take advantage

30:28 of all the training of other leading networks as well so with that let me talk a little bit about the future of

30:35 network automation for tier two and tier 3 service providers this is where I talk about Juniper's vision and what I could

30:42 say is that our vision is here today here and now we have our Cloud delivered

30:50 Network automation offering called Juniper Paragon as a service and it's going to give you that better faster

30:56 simpler way to achieve what we deem as our true north experience first

31:02 networking and we do this through automation that is extremely intuitive your operators

31:09 are going to be able to gain access to a simple user interface that's use case

31:14 based so that they don't have to use a monolithic monolithic system and actually be inundated and overwhelmed

31:21 with all these capabilities and features that they would never use they're able to use exactly the use cases that they

31:28 need to um to achieve the results that they want uh in addition I mentioned the AI

31:34 enabled part so as we streamline the GUI we're also offering that incredible Insight that only artificial

31:40 intelligence could help in terms of you know reducing the number of alarms and all that noise and honing in on urgent

31:47 actions our Paragon automation solution is encapsulating a full breadth of use

31:55 cases our vision is to be able to tackle everything from assurance and

32:01 orchestration to network planning and optimization as well and our system is

32:07 deployed in the cloud and it's based on state-of-the-art intent-based networking

32:13 it's experience driven it's fully Cloud native and it's fully open with apis

32:21 that you could use and also integrated use cases where if you actually purchased multiple use cases you know

32:29 they don't rest in silos you're able to use those cross-functionally in order to

32:34 achieve closed loop Automation and that's all through our software as a service deployment

32:41 let me talk a little bit more about that model we're able to basically allow you

32:47 to innovate faster and take on some of that Financial Risk by allowing you to

32:52 pay as you grow it's a model that works in your favor so the idea is that we're

32:58 able to allow you to pay only for the use cases that you need and we're also

33:04 offering a free starter kit so we're going to bundle this free with our Hardware initially so that you're able

33:11 to just pay for what you're going to use and we're finding that this model is

33:16 very attractive for our customers uh more attractive than you know the Perpetual license model of on-prem

33:23 deployments because it's subscription based now that free offer starter kit

33:29 does come with certain terms and and uh conditions but I encourage you to actually talk to your Juniper

33:35 representative about that um I think in a nutshell Paragon as a

33:41 service delivers all those values that I talked about in terms of delivering you

33:46 know that continuous innovation faster Innovation being able to allow you to

33:52 provide value for your customers being able to give you more productivity being

33:57 able to allow you to differentiate your services through higher quality and you

34:03 know I mentioned it's use case based and I'm going to talk a little bit later on about specific use cases that we've

34:09 deployed that we have in our Cloud instance today I mentioned enabled and

34:15 how that is going to make the difference in terms of reducing the noise in the network and being able to predict things

34:22 before they happen so that you could proactively resolve them it's able for you to

34:28 um you know look at you know the Sea of data and be able to correlate down to just the Urgent actions that need to be

34:35 taken and it also provides you with a way to be more intelligent that you know

34:42 the problems that you prioritize it also Paragon gives you that Easy

34:47 Button as operations you know we have for field Engineers a phone app that

34:52 guides them through device onboarding and it's extremely intuitive and easy to use for an example and I encourage you

34:59 to take a look at the videos that I have uh added through heavy reading in the

35:04 resource panel the first one is on our Cloud delivered Automation and specifically that phone app device

35:11 onboarding use case that's quite compelling and secondly a full demo that I've given at Mobile World Congress on

35:17 all the use cases that we have available today now if you take a look at why Cloud well

35:25 firstly it lowers your cost we found through other third-party research that it yields a 40 percent cost reduction

35:33 because you don't have all the overheads you don't have to buy all the server

35:38 servers and all that Hardware that you would have to procure and put through your it Department all that red tape is

35:46 eliminated because we're taking that problem away from you you we we manage

35:51 and maintain all our servers and all the software licenses you know in our Cloud instance

35:57 so you're able to also gain agility you know you're able to actually lower costs

36:02 by being more agile to be able to not have all that you know continuous uh

36:07 development where you're going to have to validate that features work you know every so many months and stuff because

36:13 every uh week we introduce new upgrades and capabilities and features to your

36:19 existing use cases and that's going to help you lower the cost barrier because you're going to be able to have

36:24 immediate access and secondly we're able to accelerate

36:29 that time to Market we found through this third party research that 50

36:35 reductions in deployment times were seen for cloud delivered automation for the first time you deploy the platform as

36:43 well as 70 faster new use case deployment and that's because it's just

36:48 so simple uh you know you're basically having a subscription and in addition to accelerating time to deployment you know

36:55 it's just incredibly easy we're giving you that easy button so that operators are able to be more productive and

37:01 deliver results faster in the end of the day it's all about unleashing that operational efficiency

37:07 being able to do more with less being able to solve the right problems being

37:13 able to detect issues that humans Miss and reducing that mean time to know so

37:18 that you could diagnose root cause faster and improve and impact a positive

37:24 customer experience so this is in a nutshell what our Paragon automation as a service offering

37:31 has to to deliver to you let me talk a little bit more about

37:36 lowering the risks and AI so we've seen through the survey and I'll reiterate that 93 of csps are ready today to be

37:45 able to actually adopt Cloud delivered Network automation from the public Cloud

37:51 even given the risks that we had outlined um 83 percent would consider an AI

37:57 enabled solution which is trained on other leading networks that's Cloud delivered so we're really you know

38:04 raising the bar in terms of you know mitigating risks and delivering a solution where you could take advantage

38:10 immediately and we're doing so in a way just to remind you of what I said before that's

38:16 pay as you grow so you know you only pay for the use cases that you need you know

38:22 it's bundled for free with our Hardware so you have an easy way to get started and you know it's a cheaper solution

38:28 than the Perpetual on-prem deployments because you're going to only pay for

38:33 what you need to meet your requirements so it lowers your total cost of operations

38:38 now let me talk to you about what's here and now and available today and our initial release of paragon automation as

38:45 a service so there's three use cases that we've delivered and this is just the start uh the first is device

38:51 lifecycle management which includes onboarding and again I encourage you to watch the video in the resource panel

38:57 and that really will deliver devices extremely fast so that you can lower the

39:03 time that it takes to roll out your services with network Observer sorry with network

39:10 observability you're going to gain unprecedented in the sites across devices the network and services and

39:18 with network trust this is a security capability that's going to allow you to gain Integrity confidence in the network

39:25 so that you could ensure that you have integrity and you are compliant

39:31 so I'll tell you a little bit about why each one of these use cases matters now so you know we've talked to our

39:38 customers in the past about device onboarding and they've said to us many times that it's actually quite a

39:44 cumbersome process when they send field technicians out to you know do their jobs on site it takes maybe two to seven

39:52 days per device and these you know people have to be highly trained and

39:58 they use manuals there's a lot of work involved to plan these deployments out and even with that all that work and the

40:06 expertise there's still human errors that happen and they have to work and sometimes debug issues with the network

40:15 operations center and they have to use CLI which is a very manual cumbersome approach approach too that we know is

40:22 notoriously error prone in addition you know there's unsafe practices for security like using USBS

40:29 and you know writing down notes that are of IP addresses and stuff like that maybe on Post-it notes or you know it's

40:37 just not a very um I guess um you know secure process when we're

40:43 doing these type of activities and lastly you know we have features that automate such as ztp you know zero

40:51 touch provisioning but the problem is that it stops there it doesn't actually you know validate that the service uh is

40:58 working the way that was intended and that leads to more truck rolls so with

41:03 paragona as a service what we're doing is we found that we're reducing that

41:09 device deployment time down from those days to 10 to 30 minutes per device it's

41:16 a dramatic savings uh in addition everything is self-guided there's no need for training because it's

41:22 completely intuitive you have an application on your phone that is you know accessed in the cloud that's

41:29 completely in sync with what the network Operation Center sees in their Paragon desktop instance so everything's

41:36 coordinated you're able to basically get a self-guided way of which cables to plug in where and after everything is

41:43 done and everything's validated then you're safe to know that no you don't have to return the site because

41:49 everything's been tested it's fully automated there's no errors you know it's safe because you don't have to use

41:54 USBS so you know it is the way that we think that you know device onboarding

42:01 needs to be reimagined so the next use case that I wanted to talk to you about is network monitoring

42:08 so in the past I had mentioned this before operators are overwhelmed by a

42:13 sea of networking data events alarms and it's overwhelming a lot of these systems

42:18 that we've used in the past have you know thousands of critical alarms and bigger organizations and major alarms

42:25 how do you know you know what is truly the Urgent action to be taken you really don't and it takes long to figure that

42:31 out um you know and it's focused mainly on devices and not services in the network

42:37 and many times it's just information overload and when you do fix things you you know it rests upon scripts and

42:44 do-it-yourself techniques for monitoring which you really need to be trained for you have to have a high skill level to

42:50 do and even with this there's still a lot of issues that are being missed so where we think with Paragon as a

42:58 service you know we're going to change the model to have a brighter future uh for automation is by offering Network

43:04 observability and not only Network observability for devices but for the network and services as well it's going

43:12 to allow you to synthesize all that data and focus on the Urgent actions that you

43:18 need to take so instead of thousands of critical alarms you're going to know exactly the Few alarms that you're going

43:24 to need to uh to investigate further in addition we're reducing a lot of the noise through that Ai and the analytics

43:31 we're able to also predict things that might be happening proactively so that you could address capacity concerns and

43:38 and issues of misconfigurations ahead of time um and all of this is designed in such a

43:45 way that it you know even the most Junior Knox staff could be able to

43:50 participate because a low level of skill is required so in this model you know

43:56 we're going to make sure that you don't miss those issues that you know it took so long to to uh to investigate before

44:03 because you know we're giving you immediate insights uh into the problem so that you could address things quickly

44:09 and even proactively and lastly I wanted to talk a little bit

44:14 about network security and it's a big problem in the industry in terms of uh

44:20 especially as we move to Data Center interconnect and and so forth security is an issue we want to make sure that

44:27 our devices are not vulnerable to DDOS attacks and we want to make sure that they're compliant you know a lot of

44:34 times uh there's devices out there that we don't know that we could trust how do you know if you could trust it it could

44:40 be tampered with before it arrives on site there's been cases where there's been fake devices or fake components

44:47 sometimes the device might be shipped with the wrong firmware that is going to introduce vulnerabilities sometimes the

44:55 hardware is end of life or the software's end of life and that could introduce vulnerabilities as well so

45:02 it's really important to be able to actually get more compliant in order to prevent you know these uh these

45:08 malicious attacks from these you know suspect entities so that's why in Paragon as a service we

45:15 have fully integrated into all our use cases this network trust and compliance

45:21 use case and the first thing we we've created as part of our Hardware our Juno ask Evo is to introduce zero trust

45:29 security um as part of our TPM 2.0 chip and that chip cryptographically is signed and is

45:37 able to validate that this the hardware that you have is actually indeed genuine

45:43 Juniper Hardware in addition you know through Paragon as a service you're able to do secures ETP as well as you

45:51 understand your hardware and software to make sure that it is the right load you're able to know that it's not end of

45:58 life and confirm trust and we do this compliance checking on a on a continual

46:03 basis and we're benchmarking against you know industry standards such as the um

46:09 uh sorry the center for Internet Security uh and all in all this measures the

46:16 trust continuously and allows you to see the trend if it's improving so that also

46:21 you can gain recommendations on how to improve this trust and we don't stop at just ztp we're able to actually extend

46:28 that to be able to validate not only the integrity and Trust of your network but

46:34 also that it performs to what you had expected and with that those are the three use

46:39 cases that uh We've delivered so far and we have way more to come including

46:45 active Assurance intense based service provisioning and more so I encourage you

46:50 to take a look at our demos that we have in the resource panel it'll give you even more detail and a walk through

46:57 through the GUI so you can see exactly this product in action right from the cloud

47:02 and with that I'll pass it back to Sterling Perrin to talk a little bit about our concluding remarks

47:08 yes let's uh let me wrap up I'll wrap up quick we get a lot of questions in which is great

47:14 um and I don't want to lose out on those we'll get to as many as we can but yeah quick quick summary here

47:21 um just to summarize the the highlights from from the study tier two three operators as we've shown

47:28 are at an early stage of automation but I think the data shows that the interest level is is quite High uh driven by a

47:36 number of things when we look at the challenge it's it's the high costs for the tier two tier threes it's the high

47:42 costs of implementation as well as the complexity of integration remaining the top challenge not surprisingly we've got

47:49 at least two questions came in around integration so we'll get to those in a second

47:55 um staffing issues are are front and center not as a driver for automation but when you make a decision to automate

48:01 how to move forward staff quickly becomes an issue among the smaller operators as well as budgets so that

48:09 those tend to be quite strong we did show there's a lot of interest tremendous interest actually in public

48:15 cloud-based uh automation as an option and as Kevin outline can address a

48:21 number of issues that tier two three operators are are running into so a lot

48:26 of opportunity there and then as we touched on AI um and specifically AI trained on other

48:33 networks is a emerging opportunity that we certainly wouldn't discount at this

48:39 stage as The Operators are showing they certainly want to learn more about what what can be done there with that let me

48:45 jump right into the questions um if we don't get to all of them and I

48:50 don't think we will Juniper will be able to follow up one-on-one individual ghost show the list of the questions

48:56 or some of them are quite product specific so they'll I'm sure want to get in touch Uh Kevin the first one uh and a

49:03 question that I also have myself so we talk about public Cloud delivery of

49:09 of automation of applications uh somebody is asking about providing the

49:16 same solution for customers that want to deploy in their own private cloud and so

49:21 actually and I would add to that you know you have public Cloud you have private cloud and then you have traditional can you just kind of talk

49:27 through how the automation applications will cut across those three three major

49:33 options yeah definitely so would Juniper is developing with her Paragon automation

49:39 um so it's based on a common source for both our Cloud delivered and SAS options

49:45 and before that we also have uh on-prem applications that continue to be you

49:51 know um used in the field and that we're selling today so that's not going away

49:56 but where we're moving to where the future is is you know the cloud so what

50:02 we're doing is we're going to have that SAS Cloud delivered option as well as an on-prem self-hosted option where they're

50:09 going to have their own private cloud and um you know although the on-prem solution might differ for commercial

50:16 reasons we're trying to keep those in sync yeah and another thing I could say is

50:22 that uh you know in both cases they're fully the intention is to support multi-vendor networks not just Juniper

50:30 um and for Paragon you know it's just like the the applications we've had before that were multi-vendor right

50:36 we're continuing to do that we've just started with the Juniper Cloud Metro and we're going to be expanding to you know

50:42 other parts of the network as well um and you know we're gonna offer that

50:48 irrespective of whether it's delivered as the SAS option or the on Prem deployment option

50:55 yeah okay excellent uh yeah we're still getting questions in uh let me get right

51:02 to one of them uh it's a short question but it definitely gets into the the challenges

51:07 of integration so I wanted to at least start the discussion here uh somebody's asking specifically how do I integrate

51:13 my existing BSS system with your service

51:18 um and if you want to talk more broadly about integration that's fine but this is specific to living today yeah definitely I mean for

51:25 any BSS or OSS system we have a Northbound interface that's fully open

51:32 so it really is you know the traditional model uh even with on-prem we're able to

51:37 point to that and we know we have a full like uh Postman collection you're able

51:42 to actually issue you know like rest restful requests down uh and you're able

51:48 to basically get notifications it's exactly the same mechanisms that you know are standard in the industry today

51:54 you're going to be able to use those uh so you know it's not really

51:59 um something that is going to be outside of the norm for what you do today

52:04 okay let me kind of combine two because they get at uh kind of also around integration but

52:11 uh so a question one question is will Juniper provide

52:16 the same oh no that's the wrong one uh yeah here it is Will Juniper support

52:22 multi-vendor equipment for the SAS solution and then combined with that

52:28 somebody made some comments which might kind of inform the discussion here given the findings

52:34 that integration costs are one of the most significant barriers do you think the market is always going to lean

52:39 towards pre-integrated viewer slash single vendor solution so I'm just curious you talked about a multi-vendor

52:47 but how do you see multi-bender and do you see you know some of these challenges pushing to single vendor and

52:55 you know we've noticed uh specifically in tier two and three is that there's a

53:00 lot of homogenistic networks where they do go with single vendor a lot more scenarios than uh whereas that's the

53:08 case than in um tier ones where they kind of have a dual vendor approach uh but I do see

53:14 that multi-vendor you know we've seen it even in some of the largest networks where Paragon automation the

53:20 applications have been chosen to manage you know Cisco and Nokia and and other

53:26 vendors right so it's not uh you know um outside of the realm of possibilities

53:32 that you know this is the way that a lot of the big providers are going today to go with one single vendor for a solution

53:39 for automation that can manage multi-vendor now of course there's always going to be some kind of like

53:44 intricate type of you know configurations and things that uh every

53:50 vendor does and that's where really the intent based uh you know device modeling you need something that's model based to

53:57 be able to describe all the Yang models for that and so forth so you know these

54:02 you know single vendor systems are still able to work with multiple vendors and

54:07 where they can't do it for for instance for optical you know there's not a lot of standard standardization in terms of

54:13 uh you know the interfaces for that in terms of Yang and like you know these

54:18 kind of approaches so what we could do there is actually integrate with other

54:23 systems and that totally makes sense it continues to make sense going forward

54:30 yeah good point on some of the difference between the tier twos smaller versus the larger um the smaller are

54:37 just generally historically more implying to go single vendor but that's kind of as a

54:43 I don't know a lowest common denominator or just is it as just it's the only way for them to go forward I think as

54:49 networks become more open and we talk about other open apis it doesn't they're single better because they have to I

54:57 think they would much prefer to be open they just don't have the resources to do it it's that and it's also if you take a

55:02 look at you know the findings from the survey a lot of operators haven't actually adopted automation we saw that

55:09 massive statistic there around 50 right that are still using you know the kind

55:14 of uh you know semi-automated approach and like 30 something percent or manual so you know if they're just using CLI

55:22 it's a little bit different but when they do evolve the automation it totally makes sense to have you know a single

55:27 vendor platform that basically is offering multi-vendor for for all so that you can converge and have

55:33 consistency I would say yep yep absolutely um a question for me

55:40 um I'll take this one um and this related to the AI and uh

55:46 data trained on on other networks data points I showed early on somebody asked

55:53 um do you find operators are able to no that's the wrong one these things keep moving uh would would your audience also

56:01 contribute their data to this collective training um we did ask that I didn't show it but

56:08 we asked about the challenges on on the next um question after that one in the survey

56:13 and 60 well 35 said they were not willing to use their own data which

56:20 means that 65 percent were open to the idea of using their own data to train

56:25 other networks knowing that they're benefiting from the collective use of data from others so yeah there is a

56:32 majority uh in fact about two-thirds saying that they're willing to do that even though

56:38 um it could be a concern for sure um we might have term for just one or

56:44 two left um which one do I want to grab

56:51 um yeah this one is specific to Paragon automation uh there's actually quite a

56:56 few product specific ones maybe we'll close out on this one um is is the Paragon automation

57:02 requiring additional Professional Services customization or

57:07 is it ready to go out of the box that that's a good question you know how much customization is needed when you're

57:12 doing these cloud-based yeah I mean uh you know it's an interesting question

57:18 because in the past with on-deprem deployments we found that there was you know professional Activity The

57:23 Professional Services activities that were needed more so but with the cloud

57:28 instance we've designed things in a totally different model that's total you know it's we take away all the problems

57:34 in terms of you know you don't need to deploy uh servers you don't need to install you don't need to upgrade right

57:41 so I would say that for the most part all that deployment Professional Services is eliminated and then secondly

57:49 because it's so intuitive and you know cross-functionally end-to-end in terms of use cases I would say that you don't

57:56 really need a lot of education so where Professional Services comes in and offers you know Consulting and tutoring

58:03 to try to like onboard people into the system I I think that it really alleviates the need for all that expert

58:09 training as well where I do see Professional Services playing a role is

58:15 more on the integration front so definitely you know if you're going to be using rest apis and doing development

58:20 a lot of tier twos and threes don't have um uh organizations in their company

58:26 that are specifically for development they might want to go with a third party and so forth we could take that problem

58:32 from you and you could hire Juniper Professional Services to do that kind of activity as well

58:38 sorry I hope that answers the question good stuff I think it did Let's uh we're

58:43 at the top of the hour so we'll close out here um there are more questions uh Juniper I'm sure I'll be happy to get

58:49 back to everybody individually and address those I do want to thank everybody for tuning in today thanks for

58:56 your great questions and of course thanks Kevin uh for your participation in this webinar and help throughout this

59:02 project that I definitely agree this was some interesting data I'm glad we did it thanks a lot thanks so much Sterling

59:09 thanks for having us sure and thank you everyone for attending and I just want to remind

59:14 everyone to check out the links for the paper download that read it and also check out the videos there's a really

59:20 great demo where you could get a you know walk through of the GUI I think it's going to be very revealing

59:26 all right that will close it out thanks everybody have a great day thank you