Neil McRae, Juniper Networks | MWC25 Barcelona

0:00 Music

Discovering the Past and Future: AI and Barcelona

0:07 Good afternoon networking nerds

0:09 and welcome back to Barcelona, Spain.

0:11 We are here coming to the conclusion of day two

0:14 of our four days of live broadcast on theCUBE.

0:16 My name's Savannah Peterson here with my favorite co-

0:19 astronaut on the rocket ship, Dave Vellante.

0:21 Oh, are we going to talk space here? Is that right?

0:22 We're talking space on this segment. It's happening.

0:25 Did you know our next guest is named

0:26 after perhaps the most famous person in space ever?

0:30 Neil Armstrong. - Yes. Quite literally.

0:33 I watched him land on the moon.

0:35 You and a few other people. - I dye my hair gray.

0:40 Were you there with him next to him?

0:41 It was you and Buzz just hanging out having a margarita.

0:43 It was pretty remarkable. I know

0:45 people who say that was fake.

0:48 Yeah, let's not talk about those people today.

0:49 We've got a very tangible thing

0:50 and our next guest is very real.

0:52 Neil, thank you so much for taking the

0:53 time to join us today.

0:54 Thanks for having us. Thanks for having us.

0:56 This is going to be really exciting.

0:59 Juniper, definitely a leader in AI-native networking,

1:02 but you've just had an exciting announcement about AI-

1:05 native routing. Can you explain that to us?

1:08 Yeah, so look, we were in AI before it was trendy.

1:12 We've been working on AI-enabled networks

1:15 for over five years now, focused on the enterprise.

1:18 Then we went into the clouds

1:20 and now we're extending that into our wide area networks.

1:24 And what that allows our customers to do is

1:27 to focus on their business, not focus on running a network,

1:29 not focus on dealing with network issues,

1:31 because Marvis, our AI agent, takes care of all

1:35 of the network for them.

1:36 It allows them to do planning, it allows them to do fault-

1:38 finding, it allows them to do installations

1:40 and configurations.

1:41 But literally, Marvis takes the challenges out

1:44 of the network and allows our customers

1:46 to focus on their customers and their business.

1:49 Wow. And I can imagine there's a lot of complexity

1:51 to being able to do this.

1:53 How long have you been working on this?

1:55 I mean, you talk about... I'm actually,

1:56 no, I want to take this even deeper.

1:58 You say you were in AI before AI was cool.

2:00 Yeah. - How long ago was that

2:02 and do you think it's only been two years

2:04 that AI has been cool? Or what's your-

2:05 Well, yeah, I think

2:08 we were looking at AI as a solution

2:11 for enterprise networking.

2:13 You know what it's like, you come into the office,

2:14 you plug your laptop in, but you can't do any work.

2:17 And we found that there was a definite segment

2:20 for in the enterprise space, to enable enterprises

2:23 to take the pain out of networking.

2:25 As a network guy, I'd always get a phone call,

2:28 You guys in network, it's not working, it's your fault

2:31 And most of the time it wasn't actually.

2:33 But what the AI allows us to do is to really

2:39 get past those points

2:40 and be able to pinpoint where issues are

2:42 and move on with them.

2:44 AI in terms of ChatGPT

2:47 and generative AI, that's a more recent development.

2:50 It's only been two years,

2:51 but we were working on AI concepts, not gen AI,

2:55 but more AI concepts in the network to enable our customers

2:59 to have the network that they expect, which is one that

3:02 they actually don't even realize is there

3:03 because it's working so well.

3:05 And in some customers we've seen trouble tickets

3:08 reduced by almost 90%.

3:10 So they really see the benefits of deploying our Mist AI.

3:13 That is extraordinary. 90%- - 90%, yeah.

3:16 Reduction in complaints. - 90%.

3:17 I mean, I can say we probably want

3:19 that ubiquitously across the planet.

Optimizing Network Efficiency for Economic and Energy Gains

3:21 Absolutely. And look, and we're bringing it.

3:24 As I said, we started an enterprise.

3:25 We've been massively successful there.

3:27 We're now taking that into, we took that into the cloud

3:31 where we've got more and more investment in the cloud

3:34 because of AI, because of public cloud,

3:36 because of applications and digitalization.

3:39 And again, people just want that to work.

3:42 They need it to work, lives depend upon it working.

3:45 So having something that looks

3:46 after the network, pre-diagnoses faults, takes care of them

3:49 before you need to take care of them,

3:51 and then alerts the network operators to say, "Actually,

3:53 you might want to look at this in the future.

3:55 " Is something that we just got massive,

3:57 massive momentum on.

3:59 And then our service provider customers in particular

4:01 saying, "Hey, we'd like to see that in the

4:02 service provider network.

4:03 " So we're starting that journey

4:05 with our tool called Paragon, which takes a lot

4:07 of our experience

4:09 and most importantly, the data that we have.

4:11 So we've been running enterprise networks for some time,

4:15 and AI is all about teaching.

4:17 It's about teaching the AI what to do,

4:19 and we've got years of teaching experience

4:22 and being able to teach that AI

4:24 what is right in the network, what's wrong in the network,

4:26 and what it needs to do to fix the network.

4:27 So you guys, rightly so, you talk about

4:30 how you were doing AI before, it was cool.

4:33 You have a lot of experience in this space

4:35 and thinking about your service provider customers

4:38 and how they're applying AI, if you had

4:40 to break down the value pie, Neil, how would you say

4:45 the value pie splits between the part

4:47 that goes towards simplifying operations

4:50 and lowering costs versus energy efficiency?

4:54 Yeah, I mean,

4:55 I think those two things are actually more intertwined than

4:57 people realize.

4:58 The network is so complex today.

5:01 Back when I was building networks when I was younger,

5:04 you could run the network in your head.

5:06 Networks are so complex today that you just can't do that.

5:09 So you need tooling to help you manage that.

5:12 So as you're managing your network operations,

5:14 as you're running an efficient network,

5:16 that translates into power efficiency, that translates into

5:19 experience efficiency, having the best experience,

5:22 the network effectively drives itself.

5:25 And you're using the tools to ensure that you're doing

5:27 that in an energy efficient way

5:29 and in a way that delights customers.

5:31 Because at the end of the day, when we press

5:33 that YouTube button, we want it to ping up

5:35 and we want it to be punchy.

5:36 All of this AI helps us get to that end goal.

5:40 And our customers, they don't

5:42 know about it, but they love it.

5:43 And in terms of taking cost out, particularly on energy,

5:47 we're here in Europe of course,

5:48 but energy prices here have been quite challenging

5:52 and it's been a big focus for our enterprise customers,

5:56 but also we have a lot of regulation around CO2.

5:59 That's also a big driver for customers ensuring that

6:02 what they do is sustainable.

6:04 And with our WAN enabled AI and Paragon

6:09 and some of the other tools that we've got, not only can we

6:12 reduce the energy, but we can actually display it.

6:15 And a lot of service providers I talk to,

6:17 they don't actually know how much energy they're using,

6:20 so it's hard for them then to start to turn it off.

AI-Driven Routing and Operations

6:22 You can't solve what you can't see.

6:24 Exactly, 100%. 100%. - Yeah.

6:26 So you guys talk about AI native routing.

6:28 Yeah. - What exactly is that?

6:31 How do analytics fit in there?

6:33 What's the data model look like

6:35 and how does that affect solving problems?

6:38 Yeah, so I mean the data model is actually, we've had this

6:42 for some time, but we've never really taken the

6:44 best use of it.

6:46 The data model is all the routing information.

6:47 It's already in the network today,

6:49 but typically it's been hidden away from service providers.

6:52 What we are doing is we're exposing that data

6:55 and allowing the routing protocols with AI

6:58 to take either the best path on an experience point of view,

7:02 or taking the best path

7:03 that's actually the most energy efficient.

7:06 And again, trying to do that manually when you have a

7:08 network of thousands of devices in it, it's impossible.

7:11 This is where the AI really brings in its capability

7:14 and allows us to benefit the service providers.

7:17 And again, also simple things like ensuring

7:20 that you're doing what I call the network housekeeping,

7:23 making sure everything, the watering of the plants,

7:27 the feeding of thing is getting done often.

7:30 Most network issues actually happen when operators forget

7:34 to do something and we're human, we forget stuff.

7:37 So the AI takes that away from you

7:39 and it automates the initial configuration, the setup,

7:42 the policies, everything that a service provider needs

7:45 to run the network well

7:47 and also do that on a continuous and sustainable basis.

7:50 Yeah. So what about beyond routing?

7:52 Maybe you can talk about are there AI ops use cases

7:55 that are getting traction and what are those?

7:57 Yeah, so we see a number of use cases.

8:00 Again, this is kind of linked to

8:02 what we're doing in the enterprise,

8:03 but where you're doing SD-WAN,

8:05 where you've got a large network

8:07 and you perhaps have a problem

8:09 with your factory in some part of the world,

8:12 what the AI allows us to do is actually say, "Actually

8:16 this is having an impact on this process.

8:18 " So businesses have a real impact when there's a network

8:21 challenge or when we have to do something, they know

8:23 how it's affecting them and they can take action.

8:26 And most importantly, they know how, one

8:29 of the biggest benefits I hear from customers is,

8:31 "We're able to pinpoint the issue within seconds.

8:34 " Historically, that could take hours.

8:36 "There's a network problem, where is it?"

8:38 Depending on the environment, it could take days.

8:40 Yeah, absolutely. And you might have to coordinate

8:42 with multiple providers or multiple service agents.

8:45 You've got multiple stakeholders.

8:47 They just want to know what is

8:48 it and when's it going to be fixed.

8:49 So around exposing that data, exposing that capability,

8:52 and using Marvis

8:54 and Paragon, our AI tools to zero in on that really takes

8:59 so much pain out of service providers' day-to-

9:02 day operations,

9:03 but also allows them

9:05 to reduce cost on things like service level payments

9:08 where you've had an outage, you may have

9:10 to pay the customer a service level payment

9:13 because you haven't met the service level.

9:15 And of course, that's something that service providers want

9:17 to avoid, but every service provider wants

9:21 to put on the best service that they can

9:23 because they know how much the world is

9:25 depending upon the network today.

9:27 I can imagine that's been well received in your community.

9:30 This announcement's been live for 10 days,

9:31 I'm sure you've been engaging with folks.

9:33 How has the feedback been so far?

AI-Driven Transformation in Telecommunications

9:35 It's been very positive.

9:36 And there's a number of use cases that we're focusing on.

9:39 One of them is around what we call cell site gateway.

9:42 So connecting all the cell towers back to the core network,

9:45 you've got thousands of cell towers, millions of them.

9:49 It's a high volume process.

9:51 And when one cell tower goes down,

9:53 you're affecting thousands of users.

9:55 So you want to be able to address that quickly.

9:58 You also want to make sure that you're able

9:59 to plan the network for the future.

10:01 You're adding the bandwidth from the capacity that you need.

10:03 And again, with our ACX7020, that's kind of built into

10:06 that platform from the ground up.

10:08 And certainly the feedback from the service providers is

10:10 hugely positive and they're excited about it.

10:13 I love to hear it. Congrats.

10:15 So you have all these forces.

10:16 You got telco transformation we've been talking about

10:18 for a few years now.

10:20 You got this AI awakening, not new

10:22 to you, new to a lot of people.

10:24 Even of course, the gen AI piece is new to everybody.

10:28 You've got 5G monetization headwinds.

10:31 We've certainly heard that at this show,

10:34 but you've also got an imperative to modernize the network.

10:38 So how's that working with telcos?

10:41 Are they going through a network rationalization,

10:42 their portfolio rationalization exercise?

10:45 Is that a first step to modernization?

10:47 What do they have to do to modernize their networks?

10:49 Yeah, so look, we work with telcos in that space

10:52 along four or five pillars.

10:55 The first pillar is about the monetization.

10:57 How can you deploy more services to sell to customers?

11:01 Because ultimately that's the name of the game.

11:03 That's pillar number one. Pillar number two is security.

11:07 How do you ensure you're building a network that's secure

11:09 for today, for tomorrow, for the future?

11:12 And security is a challenge that just never goes away.

11:15 Every day we learn something in security.

11:17 Number three, sustainability.

11:19 And when I say sustainability, of course, I mean energy

11:21 and carbon, but I also mean a network

11:23 that's got a sustainable platform that's going to take you

11:26 for the next 10 years.

11:28 Often we see organizations build a network

11:30 and it's right for one year,

11:31 but then they're having to re-engineer it.

11:33 We want service providers

11:34 to have certainty about the network future.

11:37 And then finally, about automation and AI ops.

11:41 And actually you can't do any of those pillars without that.

11:44 So we've been working with service providers on

11:46 what we call network as a service, the ability

11:48 to use the network as a piece of code

11:51 or through an API,

11:53 and it allows service providers to build a network

11:56 for their customers on demand when they need it.

11:59 And one example of that is what we call a sovereign network.

12:02 So building a network in a country

12:03 where the data's never going to leave that country,

12:06 historically, that's been really difficult to do,

12:09 but with this new network modernization,

12:12 we're able to do it very easy.

12:13 And that's a capability

12:15 that many customers will pay more money for.

12:17 Especially in this industry.

12:18 Pillar one helps pay for all this.

12:20 Absolutely. You've got to, look,

12:23 the networking business is all about return.

12:25 If you're not making a return, you're not in a

12:26 network business for very long.

12:28 And I've been in this industry 30 years

12:30 and my starting point,

12:34 and it's one of the reasons why I love working at Juniper,

12:35 our starting point is actually our customer's customer,

12:38 which is the end users of the network.

12:40 And if we delight them, they delight the service provider

12:43 and the service provider works with us.

12:45 And that's how we think about it.

12:47 And we think about it through the lens of experience.

12:49 How is the experience for the end user?

12:52 And if that's great, the monetization comes.

12:56 Well, we're demanding, no doubt.

Network Complexity and Service

12:58 We- - Absolutely. And we're consumptive.

13:00 And the price per bit has been dropping like a rock,

13:04 and our consumption's been going through the roof

13:06 and many service providers have had

13:09 challenges sort of keeping up.

13:10 Having said that, it's still a pretty stable business

13:14 and they do great on their connectivity

13:17 for their infrastructure.

13:18 They have aspirations.

13:19 We've seen some service providers, actually

13:24 Eand is one good example, but they're few and far between.

13:28 But at the end of the day,

13:29 if you're providing better efficiency,

13:34 that drops money to the bottom line.

13:36 It allows you to have that, they say future-proof,

13:39 but the sustainable network.

13:42 Yeah, and absolutely. I'll just build on that.

13:45 I said earlier how complex the network is,

13:48 and we've been running that with a, the way we've managed

13:51 that is by doing fewer changes.

13:54 But actually we need to evolve the network.

13:55 We need to take the network into places we've

13:57 never taken it before.

14:00 How frustrating is it when you're in a car park

14:02 and you can't get service, you

14:03 don't know where you're supposed to go.

14:05 Or a tunnel and it says take a left and there's no left.

14:07 Oh yeah, even New York City .

14:08 As our service provider, our customers deal

14:12 with those issues that enables more

14:15 and more use of the network, more dependence of the network.

14:18 And we really believe that our goal-

14:21 And trust.

14:22 And trust, absolutely. And our goal is

14:24 to help our service providers deploy new networks quickly,

14:28 efficiently, that has the lower cost bit.

14:30 So we've got our new PTX core router,

14:33 fastest router on the planet.

14:34 It is from a cost per bit point of view, is one

14:37 of the strongest platforms out there.

14:40 And we are focused on delivering all of those innovations

14:43 that drive more bandwidth at a lower cost,

14:47 at a lower energy usage

14:48 and at a lower COT usage. And our customers love it.

14:50 I love it. You're not selling ports, you're selling value.

14:53 Absolutely. And experience.

14:56 It's that experience that the customer's value, 100%.

14:59 Yeah, and looking at the community holistically,

15:02 the cell phone user to the service provider,

15:05 to whatever might be going on there.

15:08 I have to close this segment with one

15:09 of our favorite topics.

15:12 As someone named after Neil Armstrong,

15:14 and I noticed your space pin,

15:15 your Apollo 17 pin when you walked up here.

15:18 Thank you. - You mentioned that you do a lot

15:19 of research around space.

15:21 Can you tell us a little bit about that?

15:22 Yeah. So I'm a space fan.

15:26 I've met nine of the 12 people who walked on the moon.

15:30 They're kind of my heroes.

15:32 Wow. - That's so cool.

Inspiration from Space

15:35 I just got goosebumps. - And actually as a kid we talk

15:36 about, this is why what you guys do is so important

15:41 because you put a face on what we do in telecommunications.

15:47 We spoke about this earlier, but Neil Armstrong just

15:49 stepping off the land on the moon.

15:52 That's what got me into engineering, no question about it.

15:55 And it was a wind in my sail of going into engineering

15:59 and learning, how did they do that?

16:01 So what I study is actually

16:03 how did they do all the communications?

16:05 And what you have to remember about those days is,

16:08 today we take cell phones and microphones

16:10 and all that stuff for granted.

16:11 Then they have to imagine that.

16:14 How does the astronaut talk back to the

16:19 earth, but also talk to the guy in orbiter?

16:24 And they started from a blank sheet of paper

16:26 and they had to imagine all this equipment.

16:28 They had to make it modular

16:29 so it worked across different platforms.

16:32 Actually, they had to build the first truly global network,

16:34 because you're going around the earth constantly.

16:37 So I take great inspiration

16:41 and experience on how they solve some of those problems.

16:43 And they were running on a computer that our iPhones today

16:48 almost make, it's almost laughable,

16:50 but they achieved some great stuff.

16:52 And our industry is, a lot of it's based off that.

16:55 So I tried to learn about how do things get built?

16:59 How do they do it? And then I try

17:00 and share that information to show

17:01 that actually in telecommunications we make such a

17:04 difference in the world.

17:06 Everything is dependent upon us.

17:08 It's a great place to have a career.

17:10 It's a great place to learn.

17:11 It's a great place to build new technology.

17:13 And that's really important in this industry

17:15 that we have the next generation of experts

17:20 and engineers come into it

17:22 And that they can get inspired by people like you

17:24 and us having this conversation.

17:25 Hopefully. That's the goal. That's the goal.

17:27 That's awesome. - I mean, this genuinely-

17:29 Hey, just real quick, do you want to plug this

17:31 sustainability report?

17:33 Yeah, sure. Sorry. Yeah, absolutely.

17:34 So on our website, juniper.

17:36 net, we've been working with SDK on

17:42 how we make changes in networks

17:46 and what service providers need to do

17:48 to be more sustainable.

17:49 I mean, I think many of them are on this journey,

17:52 but we've got some experience

17:53 and we've kind of taken that in across all

17:56 of our customer base.

17:57 And this just helps our customers

18:00 start thinking about a plan.

18:02 And it's not just a plan for sustainability,

18:04 that's absolutely important.

18:06 It's a plan for the whole network modernization.

18:09 And we at Juniper, we've done a number of these.

18:11 We're working on many more with many customers.

18:13 We're really keen to help.

18:14 And by sharing this information with the community, we want

18:18 to be a responsible network organization that's making the

18:22 future sustainable for all of us.

18:24 That's a beautiful thing. I can tell

18:26 that you're a community-focused man just like I am.

18:28 I got to bring it back to space for one more quick second

18:30 because you were giving me goosebumps thinking about the

18:32 astronauts, and I don't even think you know this about me.

18:35 So when I was a young girl, my mother didn't want me

18:39 to be a nerd because she was afraid that I was-

18:41 Actually, me too.

18:42 Really? Oh my gosh. Something we have in common.

18:44 Yeah, so she was always a-

18:46 When you were a young girl.

18:47 When I'm... Oh, thanks, Dave.

18:49 What are you talking about?

18:50 Love you for that. Love you for that.

18:51 I'm like, "Wait, wait, where are you going this?"

18:53 I'm like seven. I'm not seven anymore.

18:54 Okay. - I mean, elementary school, middle school.

18:57 And I always identified as a nerd,

18:59 I've always been a nerd. I'm a deep nerd.

19:00 And proud. - And exactly, and proud.

Inspiring Perspectives: A Forward-Looking Conclusion

19:03 And my mom, I think she thought I would be uncool or maybe.

19:06 I'm not throwing you under the bus,

19:07 mom, I love you very much.

19:09 And she knows this story, obviously that involves her.

19:11 But I will never forget, I was at the GOES-

19:15 S launch of our weather satellite.

19:18 Well, she's no longer giving us weather data

19:19 anymore, but back in the day.

19:21 And so I'm there filming it, I was there

19:23 for the NASA social program,

19:25 and my mother is watching the launch.

19:27 Oh gosh, it makes me emotional to say this.

19:29 And I remember the phone call after she watched that launch

19:33 and knew that I was there

19:36 and all she said into the phone at the top of her lungs was,

19:39 "Go nerds.""

19:42 That's great.

19:43 And we were both just, I mean,

19:43 I genuinely just feel it in my soul,

19:46 and I felt so seen in that moment,

19:48 and I really appreciate what you just said

19:50 because there's some little space nerds out there right now

19:53 who are wondering all the different ways

19:56 that they can get involved in that industry or play a role.

19:59 And look, we're talking about satellite

20:02 to device here at MWC.

20:03 The space has come into our industry again, which

20:07 for me is super exciting.

20:09 It's almost like doing my hobby for life.

20:13 So it's very relevant in terms of space

20:15 and what's happening in space in terms of the mobile

20:17 and telecommunications industry.

20:20 And again, it's interesting to watch you and your emotion

20:24 because often when I'm talking about the space program, I

20:26 want to cry because-

20:27 I'm holding them back right now.

20:29 Straight up holding them back right now.

20:30 I think this is the closest I've gotten

20:32 to crying on the show ever.

20:33 I love you mom, by the way. And also I love all the nerds

20:35 out there in this world.

20:37 On that note, I'm going to wrap this, Neil,

20:38 because that was just absolutely beautiful and spectacular.

20:41 Well, thank you for having us.

20:42 Thank you. Thank you so much for coming on.

20:43 Anytime. - I'm glad we got to talk about space,

20:45 and I look forward to talking to you about some

20:48 of these stories and customer examples

20:49 as this rolls out and scales up.

20:55 Absolutely. Our customers are loving it and we're excited

20:55 about it. Thank you for having us.

20:55 Awesome. Yes, absolutely. And thank you Dave for humoring

20:56 me when I go down a personal rabbit hole

20:58 and you never know where I'm going to go.

20:59 Oh, I loved it. It was great. Hey, very FOMO.

21:00 Yeah. Hey,

21:04 I mean, space is cool.

21:05 We're going to have to start covering some launches.

21:07 Would love to hear what you all think about space in the

21:09 comments, wherever you might be tuning in from this rock

21:11 or from outer orbit,

21:13 or who knows where you might be watching this broadcast.

21:15 My name's Savannah Peterson. We're here in Barcelona, Spain.

21:18 You're watching theCUBE, the leading source

21:20 for enterprise tech news.