Greg Yelas: 0:00
Bob makes a good analogy. He’s a big wine connoisseur and he says to make his good wine in his good quality grapes Fair enough, yeah, to make great AI and machine learning, you need great quality data, and there was nothing he could see in the market, and so that’s why he started MIST.
Michael van Rooyen: 0:15
In a world where every device is communicating, we’re no longer concerned only with connection, but protection. Welcome to Securely Connected Everything, your gateway to understanding the intertwined worlds of connectivity and security. We have a great conversation today, so stick around and we’ll jump right in MUSIC. Today we’re having a discussion with Greg Yelas, the regional sales leader at Juniper, looking after the go-to market for the MIST portfolio. Today we’re going to talk about Enhanced Access, which covers wireless and switching. Welcome, greg.
Greg Yelas: 0:53
Hey, thank you Michael. I’ve been in the game and networking for 25 plus years and started technical. You know I’ve started technical and then realised that I really loved customer conversations, communications, helping outcomes with customers, moved into the sales around routing and switching business for Cisco for many years and then I moved into account management positions and then, lastly, a client exactly looking after the largest retailer. And then I guess what came across is you know, someone told me about MIST and they hit me on the shoulder to lead the region for Australia and New Zealand and it must be on. I’d heard of MIST, I knew Cisco was an early investor, but then, the more I went through and started to learn about the MIST portfolio and what MIST stood for which we’ll hear more today I couldn’t say no and I got so excited because I felt we could really take the customer transformation, not just a step change but a game change.
Michael van Rooyen: 1:39
Yeah, great, and so MIST is an acquisition for those. So Juniper, a big network company as you just touched on, was an acquisition. We’ll dive a bit into that in a sec. But if you look at the positioning for MIST within the Juniper portfolio, it really fits in what area for that portfolio.
Greg Yelas: 1:58
Yeah, so I guess the product segment or the branding within Juniper it’s called the AI, artificial intelligence driven enterprise, so I guess it says in the name it really fits in the enterprise segment. So anything campus, complex manufacturing, retail, health care, education they’re really the verticals where MIST portfolio fits in really really nicely.
Michael van Rooyen: 2:20
Great and we talked around and, kind of you and I’ve been bouncing the name around for a while Enhanced Access, which is really covering wireless switching and certainly SD-WAN in that portfolio. Why has it become such a focal point for customers and businesses today?
Greg Yelas: 2:38
Well, I think, since mobility has been around for quite some time since the advent of the laptop, smart phones, tablets but, however, it’s way more prominent today in the workforce.
Greg Yelas: 2:48
But more importantly, I think, actually COVID took it to next level too, because everyone started working from home and obviously technology at home became so critical and there was no competing for bandwidth for you at home, maybe with your wife and your kids, but not in a corporate, and so people got used to the flexibility and the performance at home and they really wanted to see that in the workplace. And people are way more mobile now. If you look at how office spaces are made, there’s collaborations, there’s huddle zones, there’s little whiteboarding spaces. So people need to be mobile and, more importantly, they need to have great connectivity and seamless connectivity to be able to use applications like voice and video. I see sometimes I go to a customer site and there’s someone with a laptop on a stool while they’re in a whiteboard huddle zone and they’re videoing from there, and they’ve got local people there on smartphones, tablets, laptops. Everyone needs to collaborate and communicate and you can’t have a flawed experience.
Michael van Rooyen: 3:39
Of course and I mean on the point there, I go to visit customers as well and we’ve seen a little bit internally. Everyone went through that phase of returning from work expecting the service that we know and being uplifted over on services during COVID. Now people return to work, investments into the environment at work weren’t as high, for obvious reasons. You know no point spending money on systems that aren’t being used. So there’s a little bit of a lag on that. But the expectations there right. So users are expecting that server service and I see customers, and I’m sure you do as well you go into their offices and they’re actually all sitting on teams’ calls. Right, they’re all at night at home. They’re actually sitting together, you know, on team calls, even sometimes talking to the same people in the office.
Michael van Rooyen: 4:20
However, the importance there is, you know there’s a continuous drive to better video, better quality, better collaboration, sharing, whiteboarding, etc. So it’s a good point. So really nailing that access connectivity, you know, is some of the driver. So from a missed Juniper point of view, you know, have the user’s expectations and technology enhancements or advancements, I should say really shaped the development of the missed portfolio? Like, how did it start, what were the seeding reasons and what is the outcome it gives to customers today.
Greg Yelas: 4:52
Yeah, I mean, look, I think it’s stats with the founders Like, leadership is so important, right? And so the founders of missed were two guys a guy called Bob Friday and a guy called Sujay Hijjala. And if you look at Bob’s pedigree you know back he created the first wireless land control. I say he innovated that with airspace. He sold that business to Cisco, which became the Aeronet brand, and he ran that business, which was the most successful Wi-Fi at the planet. And so, come around the end of the 2010s, around that zone, bob started getting feedback from customers saying, look, this was a great first gen Wi-Fi experience.
Greg Yelas: 5:21
However, with innovations of devices, like the plethora of devices and innovation at the device end, the network needs to change, to adapt. It was very hard to take these controllers in and out of service, update them, patch them, as costly. So Bob went out and scoured the market and said, well, do we build or do we acquire? And there was another startup a few years earlier called Meraki and it was, you know, cloud based. And so he acquired Meraki into Cisco. He’s part of the leadership team there. However, then he knew the future was really around that user experience, and so what he did is he went back to the board at Cisco and said you know what we’re going to create, a startup called Mist.
Greg Yelas: 5:56
Cisco was an investor. Google Venture Capitalists was an investor as well, and they had to do it as a blank canvas and knew when I talk about this all the time. The reason was there was nothing in the market you could buy or acquire even though they bought Meraki that could lead the future into real AI, machine learning, the telemetry, the devices, all the stuff they could get. And Bob makes a good analogy he’s a big wine connoisseur and he says to make his good wine, he needs good quality grapes. To make great AI and machine learning, you need great quality data, and there was nothing he could see in the market, and so that’s why he started Mist. And so Mist was built in 2014 with a blank canvas, and it was about architecture and architecture a microservices cloud because there’s nothing about it and there still isn’t today.
Greg Yelas: 6:39
Mist is the only microservices architecture for the network end to end, from Wi-Fi switching and SD-WAN, and it was to get the user experience data. All that data 150 states change when you mobile on the device. There’s 150 state changes that happen in seconds. Right To capture all of that on every device. It’s a person, a camera, an IoT sensor. It’s a lot of data. So you need a platform and architecture that can take that data and make meaningful use out of that data to provide a better experience automatically for the user.
Michael van Rooyen: 7:11
You talked about Bob coming out of Cisco building Aeronette, which was a legendary product at the time. Obviously Meraki, a lot of people saw, and then Juniper. So why was Juniper not part of the Cisco portfolio? What? Was there some reasoning for that? Speaking through when?
Greg Yelas: 7:26
I was through the interview process speaking to the founders, you know and I’ve been here now going on three years they mentioned to me it was around vision. They’re very about vision and future and I guess Juniper’s reasoning for acquiring Mist was the Mist AI cloud. They believed in that architecture, the data they were getting, and also Juniper had a gap. Juniper didn’t have Wi-Fi, had switching yes, had some SRX as an SDWAN, and so they saw this as a way to bring all of that portfolio in. And what we’re seeing now within the Juniper stack is we’re extending our AI engine through to the Dallas and security portfolios as well, as well as things like cloud NAC network access control, which is very exciting.
Greg Yelas: 8:05
So these innovations are really starting to broaden across the portfolio, whereas with Cisco they still had the internet stuff that he had built and sold to them, still had Meraki that he had helped acquire. I guess he was just concerned, maybe, that they still had competing product portfolios.
Michael van Rooyen: 8:20
Whereas with.
Greg Yelas: 8:20
Juniper was. They were all in. Right 100% one direction from client to cloud.
Michael van Rooyen: 8:24
Yeah, fair enough, fair enough. And then you talked a lot about the AI enterprise, ai driven enterprise. You know that’s really the go to market bill 2014,. You know, in 10 years, next year, in that space, can you describe? You know what the differentiator that the AI driven Juniper model brings versus the competitors at a high level? People don’t understand. There’s a lot of people on the who might be listening to this might think while this is, while switching, switching you put up a few your APs, you get some switching it and off you go. But can you really explain? You know the differentiators?
Greg Yelas: 8:56
Sure, there’s probably four key differentiators, the first being because we monitor every client connection you know in real time and can capture all that in the cloud. We can, for the first time, give a real service level expectation of what the users and devices should be, should be meeting, and if they’re not meeting them, why they’re not meeting them, and give the raw data automatically if it’s not meeting that service level. And that’s the first ever seen in the industry and that’s what we do. The second thing is is we take all that raw data, we wash it in our AI engine, which we call Marvis. So a lot of people, if you’re Marvel comic fans out there Iron man, you speak to Jarvis, yes, we speak to Marvis, yes, and so Marvis has two flavors. That has a proactive side where it’ll do dynamic baselining learning the individual site, location and customer and do trend analysis so that will automatically generate tickets to resolve persistent known issues. Right, and I’ve seen customer environments where we plug in one AP, it’s picked up networking issues that that customer hasn’t seen and didn’t know there were issues automatically, like within seconds, and the second part of that is a natural language interface.
Greg Yelas: 10:01
A lot of customers are struggling to retain talent. It’s a big issue out there. People are moving. The talent pool is not as big as it would like with STEM and all the folks there. So Marvis is a natural language. So think about chat GPT. Chat GPT is trained on internet data. Marvis is trained on network data. So you can ask Marvis a question. I can say why did MVR’s Microsoft Teams call, fail or have dodgy quality video? Yesterday too? Marvis will come back and very lame in terms of say, here’s what the issue is, whether recommendation to fix it or even sometimes go hey, would you like me to fix it for you? So it allows that level one help desking and people may be coming through graduates to do a lot more power on the network.
Greg Yelas: 10:39
So that’s the second thing. Thirdly, it’s back to that Microsoft services cloud. This is only possible because of the architecture that underpins this, to be able to capture all that data and in real time do AI machine learning algorithms to be able to provide an outcome to the user. And then the last thing, which is unique, and it’s back to the return to work force.
Greg Yelas: 10:57
We had the seventh biggest bank in the world. They had challenges bringing staff back and facilities were trying to provide value added services to be able to track and trace people, things. Find the guy with the bar fridge on a Friday, you want the beers right, and so to do that, they needed to put a whole what’s called Billy battery-based beacons around every office. The cost of that for this bank globally was way more than just replacing the network. And when they realised that we’ve missed, we’ve got some patented tech that allows there’s 16 Bluetooth arrays and an access point, allows beamforming down and up and AI ML behind it to do one meter accuracy without the need of another network. And so when facilities and workplace saw this in this global large bank, they replaced the whole network and went mist purely on the return to workforce, which was interesting.
Michael van Rooyen: 11:47
From an ROI point of view. I just want to backtrack a little bit. You talked around operations, 150 points that are checking connections. So really from a operate point of view so we talk about, you know, wireless side or connectivity side the operations is really the big driver here. How do we reduce noise from a ticket point of view, how to reduce noise from a troubleshooting point of view One of the real premises, right, user experience first, but also how to reduce the operates right? So rather than just get a flood of tickets and troubleshoot we always know that wireless troubleshootings can be very, very difficult You’ve got to coordinate with somebody. You might miss it, but Marvis is doing a lot of that work for them. So would it be fair to say that in many instances Marvis could replace kind of level one, level two help desk I mean, you’re not going to get rid of them holistically, but really a big reduction Would that be a fair statement?
Greg Yelas: 12:36
I’d say augment it right. So I think, you know, one thing it’s going to do is they’re going to get way less more tickets, right?
Greg Yelas: 12:41
We generally see our customers. You know anywhere from 70 to 80, even 90 plus percent, like ServiceNow, which is a massive service management company. Their own stats they’ve deployed a Juniper Fullstack, you know WAN, sd, wan switching and Wi-Fi globally and they reduced their VP of technology 90 plus percent of tickets by hitting the help desk. So I guess the level one people are going to get way less tickets. And then the second is actually a guided assistance. I think, like you know, a Tesla self drives. Marvis will self drive that user to go hey, you’re level one, ask a question, so you can just practically go. Show me why I’m happy users today in the office. Marvis will show the users.
Greg Yelas: 13:19
You click on that user, I’ll say it’s Michael Van Rooyen, clicks on it and says here’s a recommendation to resolve it. So it helps the level one person as they’re coming up to speed, trying to a lot more up the chain too, without having to triage it to level two, level three, and, as you said, because we do everything dynamically, no longer do we need to send people to sites, or that’s often because usually with Wi-Fi, as you said, try and replicate the issue, try and roam around and cause an issue, get someone really skilled with a packet sniffer on a wireless device with multiple wireless cards roaming in office, try and replicate the issue. That’s all gone with our platform we go from here’s one Michael had an issue all the way down to the data forensics that says here’s a packet capture on exactly what happened in Michaels and we do that historically so we can send less. So the gap retail globally their stats are. When they replaced their legacy network with MIST, they reduced site visits by 85% which is huge, and they’re a massive distributed, global retail chain.
Greg Yelas: 14:19
So there’s something. So I wouldn’t say level one goes away. I just think we get to make a level one way more powerful, more like a level two-ish, three-ish, because with the data Marvis can help them and we give them way less noise as well.
Michael van Rooyen: 14:30
With the convergence of wide and wireless. So what are you seeing in relation to that? What I mean by that is wireless versus become the go-to. Are you seeing that really transition where people are installing less switching, relying much more on wireless, or is there still pretty much on par? And then, if you talk about the full stack, how is MIST? Is MIST started as a wireless product? How’s it now going across the stack? What’s the plans for a Juniper point of view to extend that? Past wireless and switching and now SD-WAN?
Greg Yelas: 15:06
Yeah, so the first one. So I think we the biggest challenge customers have with wireless first strategies was the visibility, because if something goes wrong with wireless, as you said before, it used to be a black art. Yes, now with the MIST platform, with Juniper, we have solved that black art because we have all the visibility there. So, Steve Day, and, as the public soccer mentioned, there’s the CTO of NAB, national Australia Bank. He mentioned with the Juniper platform with wireless and switching being so seamless, and now all of his corporate offices. He’s gone wireless first, which means far less switch ports.
Greg Yelas: 15:35
Another one of our customers, origin Energy. Again, they didn’t have confidence in their legacy provider to go wireless first and now that they’ve rolled out switching a Wi-Fi with Juniper nationally, all of their offices are Wi-Fi first. And they’re head of IT operations, oliver, he goes to me, greg, wi-fi used to be a pain in my butt, like every day. Everyone can play on a Wi-Fi, he goes. Now it’s my golden child and I get no noise and he goes. That’s the best thing, right? And then I guess, lastly, juniper. So that was one of the cool things about the Juniper acquisition, because Juniper and Junos was always an open standard and MIST was an open architecture, open.
Michael van Rooyen: 16:10
API.
Greg Yelas: 16:11
The integration. So one of the cool things is when they acquired MIST into Juniper, guess how many days of talk NBR for them to onboard the switching into our MIST cloud.
Michael van Rooyen: 16:21
Or generally when that sort of acquisition happens, could be months of work or, if not longer, years potentially. They were able to onboard the switching because of the openness of the Juniper platform and opens by cloud in 11 days 11 days to take traditional switching portfolio and bring it into MIST To provide telemetry into the cloud, yeah, so obviously now it’s advanced over time because of the architecture.
Greg Yelas: 16:43
So one of the great things about the microservices is that every fortnight we call it Christmas and MIST. There’s new features and functionality that in production the customers get. So if you buy a five-year subscription through Oro under a managed service, they’re gonna get over 100 innovations in that lifestyle. And if you liken that before, if to take control of it in and out, you might only get one major innovation every year, 18 months. So it’s really step change. The agility. And so where’s Juniper’s going is really spreading that AI our Marvis MIST AI cloud across more of the portfolio. So you know, started with Wi-Fi, then switching onboard it quickly, then now SDWAN with our session smart routers and our SRXs, and so now we’re moving into managed access and as well as data centre and security. So we’re bringing some of that stuff all in together.
Michael van Rooyen: 17:30
But what are some of the upcoming features and enhancements that MIST is bringing to board that customers should get excited about, whether they’re customers are not. But what are some of the innovations that you guys are driving that we should be looking forward to?
Greg Yelas: 17:45
Yeah, so I think, for our local region here. So I’ll just speak locally. So one of the key things I think we just launched last week is our local cloud. So it used to be all AWS public and GCP public in America and Europe. So we now, as of last week, have our local microservices cloud here in Sydney, which is great. So just for those customers that want to have something local, on country, right, fair enough. Do they necessarily need it? No, I don’t think so, but some customers want that. But it’s here, right?
Michael van Rooyen: 18:09
That’s one thing.
Greg Yelas: 18:11
The second is what we call our cloud access assurance. So think of it as cloud NAC. You know again, back in the Cisco days when I was there, you know I did partner with a bunch of customers around identity services, engine, ice, and I know customers that have done ClearPass and others. Now those again, those products fit a purpose at the time but things evolve so we’ve effectively mystified that. So rather than having a lot of boxes on premise, a lot of key architecture, latency considerations, complexity around policy creation, we have what I call mystified that We’ve now got a microservices access assurance in the clouds, the first native cloud access assurance. We can set all the policies there and it really simplifies the process and provides all the visibility. So now when I ask, why can’t MBR connect to the Wi-Fi? If it was a certificate or an 802.1x issue, we’ll have all the events and signatures to point, Marvis will go actually it’s not a Wi-Fi issue, it’s a certificate issue on an MBR device all in one place, which is really powerful, because other vendors have to go to multiple different pockets and places and takes a long time to find that. We can do that within seconds.
Greg Yelas: 19:15
But I also want to say we’re really important where I already come in, because whilst we’re mystified to make it simple, taking the customer on a journey on what they’ve done today with their ICES and clear pass and policy may be different in the new world, and so you need a partner like yourselves that understand both worlds. You better take that transition across. So that’s the only thing I’d preface to customers While we make it. We’ve made it really simple. You still need some thought to go. Well, how you did it yesterday may not be the way you do it tomorrow, and a partnership like Orro is really important to consult on that piece.
Michael van Rooyen: 19:45
When people classify wireless, they start thinking about IoT. They also start thinking about 5G. Are you able to talk about what you guys are perceiving happening in that space, what you’re doing maybe in that space because it’s important for people to differentiate but also how Juniper is working in those areas?
Greg Yelas: 20:00
Yes, I’ll talk about the IoT one first. So with IoT, one of the challenges is often people call them headless, right? Like how do you actually secure them, get on the network easily, when you can’t actually put like software subtletons on those devices and things like that? And so we have a subscription, which is part of our AI bundle, called IoT Assurance. It’s actually the same subscription you get with IoT Assurance is the same as the CloudNAC, it’s the same price and you get both of these functions, which I think. I actually think we should start charging more because customers get way too much for a little, I think.
Greg Yelas: 20:27
But anyway, this IoT Assurance. What it allows you to do is to seamlessly lifecycle and securely, through private, pre-shared keys, these IoT devices onto the network. So it means that you can get them seamlessly up to 50,000 devices in one hit. Get them on the network, whether they’re cameras, door sensors, hfacs, whatever it may be. Bring them onto the network securely in a segmented, private, pre-shared key so that maybe all those CCTVs, they’re all on that one VLAN as an example but then we can lifecycle them out. So we can actually lifecycle them out and re-change those in a real centralised fashion, which was before was very difficult. You’d have devices on the network from IoT that, from a security point of view, provide a real security risk because you may not have changed. You’ve got one password on the network file and leave it. We can now have a security around it with being kind of a dumb device.
Greg Yelas: 21:15
Yes, yes and so that’s called IoT Assurance, which I think is really exciting. And the second point is with our partnership with Cradlepoint, with Ericsson, yes, so Cradlepoint, we can now onboard a lot of the Cradlepoint SIM Management 5G devices into our Session Smart Routers and our SD-WAN portfolio. So through the Ms Cloud, through the partnership we can onboard, we can now get telemetry. So again, we’re all about operations. It’s simple, because if you go, hey, why is Michael having a bad experience? And it was a WAN related issue. Well, wan’s one thing we want to be able to give the user to go. Actually it was an LTE, it was a 5G issue and this is the latency and packet loss we’re seeing from the Cradlepoint device. So if that’s an issue, we want to be able to provide all of that in one place. So we’ve partnered with Cradlepoint for that, and so we can do all of the onboarding, all of the management, all of the operations within the Ms dashboard, with the Cradlepoint API integrations and, obviously, hardware, you know, as the side of the devices.
Michael van Rooyen: 22:10
So for customers who want customers, listeners anyone curious you know. Obviously Orro group is one place to start with us. From a managed service provider point of view, obviously we cover architecture, design, implement and the full gamut in conjunction with ourselves. But are there any other key areas that customers could go and you know, look into Juniper or Mist Any further? Is it just the Juniper website? Is there any particular trial before you buys? Is there any of those sorts of scenarios that customers could look into?
Greg Yelas: 22:40
Yes, I think. Go to juniper. net and look at our AI driven enterprise, mist. com, as well. Still there from the south face, but it’s MIST right.
Greg Yelas: 22:48
MISTcom you can also request we do. Every Wednesday we do a wireless Wednesday webinar. It’s an hour we go through an overview of the platform and a demonstration and then, lastly, we’ve got a couple of different options. One we’ve got an AI on us partnership bundle with Orro, where we can up to 40 access points. We have a co-investment model, you know, to be able to then prove what we’re saying here. We have a try and buy program, absolutely, and then we also got demo kit with our distributors and partners as well.
Greg Yelas: 23:19
So there’s many different ways, but I think it starts with education and awareness. Yes, other do it yourself. And if you need third party validation, which I know some customers do, if you look at the Gartner Magic Quadrant, yes, you know where the top right in execution and vision for wider wireless Juniper is the top furthest right, and actually, Gartner themselves were telling me that it’s the fastest in history to go from a visionary as a startup in 2014, 2015, which really got in the Gartner Quadrant about 2016 to being the leader of the last three years in a row. So, yes, if you need that third party validation, you can look at Gartner as well.
Michael van Rooyen: 23:54
Yeah, great, greg, thanks for your time. Thanks for coming and having a discussion about a mist and the play that Juniper is doing. Again, we’re very excited about it. I think customers that I’ve even spoken to may not be our customers today. You know, collaboration, the feedback has been fairly phenomenal. I must say, from what we’re seeing, the delight to see this happy Wi-Fi happy Li-Fi, you know, taking place has been phenomenal. So thanks for your time.
Greg Yelas: 24:23
No, I appreciate the part of you.
Michael van Rooyen: 24:24
Thanks.
