Juniper Networks CTO Raj Yavatkar Discusses Beyond Labs

0:10 Hey, Raj. Welcome to another podcast.

0:13 Ray, it’s nice talking to you. So thank you for having me here.

0:17 Thank you. Well, I wouldn't say nice because we don't know what I'm gonna ask you yet.

0:20 So who knows if I'm gonna be nice or tough for you now?

0:23 But I actually appreciate you joining again.

0:26 So what I wanted to do for this,

0:28 podcast, I hear that from a Juniper perspective,

0:31 you continue to focus on innovation, and finding creative ways to add value,

0:36 you know, to your customers. One of the things I wanna

0:39 understand is I've heard about this Beyond Labs,

0:41 and I wanna find out a little bit about what it is,

0:45 right, and if you could tell our audience about why you're introducing this particular product,

0:50 at this time. So I think, as you know,

0:52 right, Juniper is rooted in innovation,

0:55 started, by taking out an incumbent by doing completely new stuff for routing area.

1:00 Now, last, few years we have evolved into multiple new areas such as enterprise data center,

1:07 AI/ML driven network automation and so on.

1:10 Juniper Beyond Labs is our effort to organize our innovation activities,

1:16 that span new products. We do partnerships.

1:20 We have partnerships with academia and research community,

1:23 and also we are beginning to do collaboration with other entities in the industry,

1:28 in the context of network as a service and so on.

1:30 So ideas to bring all those efforts under one umbrella so we can communicate with the customers.

1:36 When they come to our EBC, or we go talk to them or in other areas.

1:41 I like what innovation they're doing to bring networking industry forward

1:45 in new areas. That's really the goal.

1:48 Yeah. And the focus, is it more for service providers,

1:52 enterprise, or interoperability. I mean,

1:55 who's the main target audience that would use this?

1:58 That's a very good question. So I think, now,

2:00 focus is not on a particular sector it's really all of the,

2:04 segments we had.

2:05 Okay.

2:05 More focuses on making sure that we we try to push the

2:09 boundaries of innovation in networking in all areas.

2:12 So spanning, AI multiple network automation,

2:17 utilization, 5G, all the areas are open to that.

2:21 Yeah. Quantum computing. I think is one thing I am pursuing right now.

2:25 Yeah. I think this is great

2:26 because I think in a lot of cases where we go where the customers is,

2:31 you know, what problem are you trying to solve and address and stuff like that?

2:34 But what is the different technology areas that are gonna be you know,

2:40 that customers can leverage within the lab there.

2:46 That's, another good question because there are various areas we're looking at.

2:49 Some things are further out. This is the example of quantum networking.

2:53 It's about five years out. Right? The quantum computing has made a lot of progress.

2:57 So go that far out, but then also look at something out three years from now.

3:01 So generative AI has created a lot of interest.

3:04 It does how do we apply to networking products,

3:07 but as well as how we deliver customer service,

3:09 customer support, and so on. So that's another area.

3:12 Which is an extension of our AI driven enterprise thing we already doing

3:16 and which is in the Magic Quadrant.

3:19 of Gartner. Then other areas include,

3:22 stuff such as, the virtualization network as a service 2.0.

3:27 The way we understand network as a service is very different today.

3:30 More of a consumption based model where our vision in network has a service.

3:35 So various such areas we're going to try to access.

3:40 Yeah. No. I think this is good. I think one of the areas

3:43 when you talk about network as a service,

3:45 or even automation, and AI,

3:48 you know, I think some of the challenge that I hear from customers,

3:50 and you tell me if you agree or not, is where do I get started?

3:54 Right? What are the insertion points

3:58 and potentially you're saying this could help solve an issue

4:01 associated with that so there's more than just interoperability,

4:04 then also plan about where the future is going and,

4:08 and get a better direction from an innovative part,

4:10 and give them visibility. So does that help,

4:13 reduce the risk of, as part of the lab to help these customers say not only

4:18 it's about where these innovations are going.

4:20 But reduce the risk of innovation. Correct?

4:22 Is that a key part of it? That's right. That's it.

4:25 That's a very good point,

4:29 you're making. Because if you look at, I guess yeah.

4:31 Sorry. Go ahead. No. No. I was gonna say go I think when I look at innovation,

4:37 usually, what I hear from customers when I talk about,

4:40 whether it's AI, right, or

4:42 automation is I need to get some guidance on where my insertion points are.

4:48 Right, how to get started because in many cases,

4:52 I see way too many silos of automation.

4:55 Right? And people aren't maximizing their capabilities of how automation should be deployed.

5:03 So is this a key part of, the lab that addresses some of that?

5:07 That's right. So I'll give you an example. So let me step back.

5:10 Right? Our true north is experience first network.

5:14 We want to deliver to customers, which are network engineers and their customers,

5:18 which are the network users, really truly experience first network

5:22 and focus on making the experience as simplified as possible.

5:25 You brought a very good point. The automation is siloed from one domain network we came up

5:31 with this vision of end to end assurance across the domain.

5:35 End to end assurance means that your network should just work.

5:38 No matter, even if it's spanning from a data center or campus branch to,

5:43 edge to the transport network environment and to the cloud.

5:46 Right? So where Juniper Beyond Labs helps is that to get to customers to a comfort level,

5:52 we'll bring them into our different centers in Westford, Mass,

5:56 or Bangalore, or, Sunnyvale,

5:58 and we will show them the actual working prototypes of end-to-end assurance working at across,

6:03 and they can take some of this back to their install base and try to sell.

6:10 So we sort of raise the level of comfort,

6:14 respect to such a new idea. That's a purpose of it.

6:17 Now Raj, this is great. Actually, this is this is pretty exciting now.

6:20 You know, one of the things though is you can't do this alone.

6:23 Right? And customers aren't silo from a vendor perspective.

6:28 What what is the ecosystem environment

6:30 and what are some of the things that excite you about this Beyond Lab?

6:34 So that's an excellent point.

6:37 Right? We cannot do this alone, and customers cannot do it alone.

6:40 And there's a whole big ecosystem around us which we can take advantage of.

6:44 And we are very good at that. So Juniper Beyond Labs also

6:48 believes in building partnerships and collaboration with multiple ecosystem players.

6:53 For example, we are now sponsoring research and academia.

6:56 So some of the further out ideas can be brought into Juniper Beyond Labs,

7:00 and our engineers can start collaborating with universities.

7:03 So we have funded research at, for example, Stanford,

7:06 Purdue University, and so on.

7:08 Right? The other thing that we're doing is industry players.

7:11 There are much larger industry players in the ecosystem who are very complementary to our goals.

7:16 So for example, we have a strong partnership with Intel.

7:19 With Dell. And, the area of quantum networking,

7:21 for example, as I told you, we are also partnering with a startup in UK,

7:25 which was a pioneer in bringing quantum networking,

7:29 at this, entanglement concept to networking.

7:32 And we're using that in our quantum key distribution.

7:35 So ideas have to be the strong ecosystem of partnership.

7:38 So we emphasize collaboration across status.

7:41 I got it. Yeah. No. I got it. I think so it's more than just,

7:46 risk reduction. Right? It's also the opportunity there's some cost savings there as well.

7:51 Where the potential customer doesn't have to invest in this,

7:54 like, expensive type testing environment or innovative labs for that.

7:58 Right? Now, if we look at from,

8:01 from a Juniper perspective, there's a lot of other companies that have labs

8:06 and they've done testing for some time.

8:09 How is it that Juniper's different than everybody else in the market there?

8:15 That's good. So I think a lot of the times when we do larger testing,

8:19 it's about the current product portfolio. What is out there?

8:22 What is shipping on the truck? What we are saying is that's not,

8:25 you know, Juniper Beyond Labs is saying, let's take one step further.

8:29 Let's look at two years out, three years out,

8:31 five years out, say, what is possible?

8:33 And looking at the art of possible, try to convert that into actual prototypes

8:38 and demos that people can try to get their hands on.

8:41 I'll give one concrete example. Quantum computing is people say five years out.

8:45 Quantum networking, but we apply quantum computing is one of

8:49 the applications people are interested in.

8:51 What we did recently, and I'm very proud of that,

8:54 we basically looked at the problem of key distribution.

8:58 When you have IPSec like protocol, you have a symmetric key distribution.

9:01 And it's prone to man in the middle of attack.

9:05 But if you use quantum entanglement, you can do key distribution of

9:09 exchange without anybody coming in the middle.

9:12 That's a pretty secure thing. We have actually implemented that working with a tech,

9:16 start-up, which has a technology available for that.

9:19 And now, actually, you will not believe that we will be actually selling it with our IPSec,

9:24 MACSEC product. This quantum key distribution.

9:27 So it's not just looking further

9:28 but saying that how can we bring future here

9:31 and make it possible at least in incremental steps so customers will comfort it.

9:37 Oh, okay. Yeah. No. I I think this is good.

9:40 I mean, I think it's a step in the right direction.

9:42 I'm glad that Juniper’s thinking about its customers, reducing risk,

9:47 reducing the cost of testing this innovation.

9:50 Where do you see this going, though? Down the road,

9:52 because I know what part of your job is to think three,

9:56 five, ten years, and I know that's a difficult thing to do.

9:59 But this is a good step in addressing the two,

10:03 I think, important things, identifying where to implement innovation,

10:07 reducing the risk, actually a third one,

10:09 and reducing the cost of getting that information.

10:12 But where do you see this going next in that perspective?

10:17 So I believe two areas where we are, good.

10:19 I'm really excited. Almost, you know, you saw a lot of and you are the expert.

10:22 You can tell us, generative AI, has generated a lot of buzz,

10:25 Right? But it's not magic.

10:28 It's subject to hallucination, human beings in the room.

10:31 We have an advantage with AI driven operations we do for enterprises.

10:35 We are already a conversational interface.

10:37 We're already using machine learning to self resolve troubleshooting the case during self healing.

10:43 Now apply generative AI to that.

10:45 And I see a big inflection point with respect to taking a step up in terms of how you deliver

10:51 customer service. A customer experience.

10:54 So for example, if you want to configure,

10:57 BGP peering or inherit a circuit or do any of the,

11:02 complicated things that the WAN operator has to do.

11:05 We can simply have a Chat GPT interface now,

11:08 which says what is my intent? Express simply natural language.

11:13 And underneath, we implemented automatically.

11:16 That's something I am trying to push it right now very quickly.

11:19 That the field is evolving very quickly. Second one,

11:22 I mentioned in network as a service, 2.0,

11:24 I call it. Network as a service, 1.0 is out there.

11:27 It's mostly pay as you go model. But that's not enough.

11:30 Imagine that if you are a network engineer

11:33 and you want to do an end to end virtual network from on prem to the

11:37 public cloud to test out a new feature

11:39 or test out something,

11:43 new ideas without disturbing the production network.

11:46 Network as a service 2.0 could automatically set up

11:50 such a virtual network over a network

11:52 for you and make it available so you can start playing with it.

11:56 That sort of things we want to bring together by using Juniper Beyond Labs.

12:01 That will be good to experiment. We also get to show people what they can try out.

12:06 Right. Yeah. Basically, maybe close with is when we when we look at,

12:12 you know, customers and,

12:13 from that perspective is, are are you able to provide remote access and,

12:19 or is it has to be on premise type testing and stuff like that?

12:23 Oh, so the Juniper Beyond Labs is both about using public cloud

12:26 because that provides a quick way to provide things out.

12:29 So we want to provide both ability for me to take our innovation prototypes,

12:34 try them out either in the on prem, in a co-lo provider like Equinix or Megaport

12:39 or in a public cloud.

12:40 Because the innovation that you can do using public clouds and shortcuts.

12:45 So it's all three are possible. Good.

12:48 Awesome. Well, I think that addresses all my questions,

12:52 Raj. Thanks for joining the,

12:54 Deep Edge podcast. No. Thank you,

12:57 Ray. It's always a pleasure to talk to you. So.

12:59 And, hopefully, you could join again because I wanna hear how the progress of this is going.

13:04 Sure. Look forward to it. Alright. Till next time.

13:07 Thanks.