(function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start': new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0], j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src= 'https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f); })(window,document,'script','dataLayer','GTM-KX9RFV5L');

Mister Beacon Episode #87

Bluetooth and Wi-Fi – The Industry Re-examined with Mist

February 26, 2019

This week we welcome Bob Friday, CTO of Mist, back to The Mr. Beacon Podcast! On this episode, Bob gives us an updated on how Mist are using machine learning in the realm of Wi-Fi and Bluetooth networks. We discuss some of the recent Mist deployments at Disney’s Swan and Dolphin Resort and The Veterans Department, and discuss the state of the Bluetooth ecosystem and where it is going. Find out why Bob thinks Bluetooth is going from a “nice to have” to a “must have”. If you missed the first episode with Mist, watch it here: https://youtu.be/AlNRGbTEMkc


  • Narration 0:07

    The Mr. Beacon podcast is sponsored by Wiliot, scaling IoT with battery free Bluetooth.

    Steve Statler 0:16

    Well, welcome back to the Mr. Beacon podcast this week, we're interviewing Bob Friday, who is co founder and CTO of mist. Bob, welcome back to the show.

    Bob Friday 0:28

    Thanks to be here, Steve, as always,

    Steve Statler 0:30

    yeah. So I think last time you were with us was a year and a half ago. So if anyone wants to kind of the baseline of what you're doing at MIS, then they can go back to that. And of course, before missed, you are at Cisco, you were CTO of enterprise mobility, which is, you know, basically the wife, what is enterprise mobility, it's Cisco.

    Bob Friday 0:51

    Back, when I left Cisco enterprise mobility was kind of the wireless switch routing group all was all put together in one group. And we were basically in the transition of converging wireless was becoming a mainstay and actually converging with all the switch router functions.

    Steve Statler 1:06

    Okay, then you went off and you set up missed and misters suddenly was working with Cisco last time we spoke. So here's the things that I think it'd be great to talk with you about in just a few minutes. So one is, let's get an update on how things are going missed and baseline what you guys are doing. You now have customers that are going into production. So I think it'd be really cool. You've got a couple that I would love to talk about if you can talk about them. You've done some work for Disney, the swan and dolphin resort out in in Florida, and also the Orlando VA. So it's always good to talk about real deployments lessons learned what you've done. So let's talk about that. And kind of the third topic I'd love to cover is your view of where the Bluetooth ecosystem is. There's been a lot of controversy and naysaying and optimism as well. So let's get your view on that. But first of all, for those people that didn't watch the original episode, can you explain to us what does mist do? Yeah,

    Bob Friday 2:11

    so this is really on like two missions right now, you know, the first mission is really around what I call Jeopardy wireless, wireless Jeopardy, right? And this is really around trying to build a solution that can answer questions on par with a wireless network expert. And for those who remember que, Watson in jeopardy, right, when they were they built a solution that can actually play Jeopardy, that was kind of one of the inspirations for Miss, if they can build software, they can play Jeopardy, we should be able to build software that can play wireless Jeopardy, right. And so that's one mission that missed his on. The other mission that we've been on as kind of is bringing indoor location on par with GPS. And this is a mission that dates back to my aerospace, Cisco days of always trying to get the industry. You know, can we agree on something that, you know, when we open up our phones, it works. It's like GPS. And that's what got us into the BOE? You know, because when I looked at one indoor location, there was always friction points for why people were having a hard time adopting it, right, whether it was the Overlay Network, right, you know, the indoor location kind of always required that. So I think we're solving that problem with a convergence at myths, right? We are converging Wi Fi and BLE together, Wi Fi is no longer a nice to have, it's a must have. So we don't argue people don't argue about deploying Wi Fi anymore.

    Steve Statler 3:32

    Yeah, that seems to be like now standard operating procedure is the big guys, at least you buy a Wi Fi router, you get Bluetooth radio in there. And typically, it's kind of fairly basic radio, what do you guys do that's different from just simply having a Bluetooth radio with a Wi Fi access point.

    Bob Friday 3:51

    So I think, you know, when it comes to BLE, I was a miss this kind of best of breed, I mean, we've done two major things. One is we've kind of virtualized this whole BLE beacon. And that's what kind of what we were known for. That was one of our first patents of really, taking the whole battery beacon problem make it easier to use. Probably the second big thing we've done is really around leveraging machine learning, to actually learning the path loss model, right to get better accuracy, right. And this is really something I've always wanted to do since my aerospace days, and never had the opportunity to do it. But now with all the Compute and Storage power we get at AWS, there's really no limit beyond, you know, beyond my Amazon bill, of what we can do now. So that problem is actually solved now where we can actually learn this path loss model for every AP and every device. So that's probably the other big thing we brought into the table.

    Steve Statler 4:41

    So the so what does path loss model mean is that they're kind of the decay of the signal over space. So you kind of get a better correlation between the signal strength and the position is it

    Bob Friday 4:55

    I mean, it did you remember you remember the the equation for a line

    Steve Statler 4:59

    you're I, I've got it in the book somewhere, but you'll do a better job of explaining.

    Bob Friday 5:05

    That's your classic y equals mx plus b. Right? Yeah, you know, what you're trying to learn is how fast that our society is dropping with distance.

    Steve Statler 5:14

    Yes. And so it's not a straight line. That's the key thing that we talked about in the in the in the beacon technology book. And so, so it's not a question of just understanding a standard curve, it varies from place to place. Is that the Is that why you need AI in there?

    Bob Friday 5:30

    Yeah. Because, you know, for anyone who's worked on, you know, artists, I base location, you know, that curve changes, you know, depending on where you are, and it changes for every device, right? There's a big, there's a difference between an android and iphone, especially on the intercept, right, how well they actually hear the BLE signals. Okay. So that it varies across all your devices.

    Steve Statler 5:48

    So you've got machine learning there, but it sounds like is that machine learning elsewhere in your offering? You're talking about kind of the Jeopardy Watson model? Is it just purely on the location piece? Or is it more than that? Yeah, you know, if

    Bob Friday 6:02

    you look at on the Wi Fi side, the reason that Wi Fi and BLE go better together as Wi Fi is kind of become this industry standard for connectivity, right. And you look at BLE BLE is starting to become the industry standard for your location, use cases. And we look at our customers who are actually putting these use cases on top of their wireless network. These are critical services now that kind of require better visibility on the Indian user experience, right? You know, so it's one thing I tell you, it's one thing back in the old days, when we had unhappy employees on your Wi Fi network, is a whole different thing, we have an unhappy customer. And so that's where the wireless for Jeopardy Marvis AI piece comes in, is really helping customers get complete end to end visibility on a customer's connectivity experience. Right? You know, because if you're gonna put a location based app on someone's phone, it's not gonna be a very good experience, if they don't have good connectivity. And you have to have good connectivity and good location to get that whole experience to work. Right?

    Steve Statler 7:04

    How can I help improve the connectivity? I can see why it could help understand this path loss. But it actually helps make sure the pipe is flowing better.

    Bob Friday 7:16

    Yeah, basically, you know, I would say the paradigm shift is, you know, back in my aerospace days, you know, 15 years ago, that was all about really helping enterprise, it, manage a whole bunch of access points, where AI is helping is really starting to provide that end to end visibility from the actual user device, all the way to the internet. Okay, so it's no longer good enough when someone says, Hey, why is your experience bad? Usually, it is struggling with trying to figure out is it on the client? Is it on EP? Is it on the network? Or is it in the Internet somewhere? Right? So AI is starting to bring all that information together? You know, and it's something that you know, a regular person does manually, right. But it usually will take an IT guy, maybe hours, if not days, depending on how complicated the problem is. And that is we're really, we're leveraging the same way they did for you know, Watson, for Jeopardy. We're really going through terabytes of data very quickly. And that's what computers are good for, right? Going through terabytes of data, looking for correlations and similarities. And that's where AI is starting to make a difference in the networking world.

    Steve Statler 8:21

    So you're, you're pausing through, you're sifting through huge amounts of data that gives you a sense of where the performance problems are, and, and how do you resolve those issues that you spot?

    Bob Friday 8:35

    Yeah. So it depends on what type of problem you actually see. If it's an RF problem, we can actually adjust those ourselves, right? If it's some sort of channel power interference problem. If it turns out it's a DNS or DHCP problem. In some cases, we can correct it, we have access to the servers. And we're actually working with some of our partners, right in what we call AI for it, you know, so we identify a problem in a firewall or a router or a switch, you know, can we actually start sending messages back to the router switch, they get things corrected? A typical use case would be something like a config, right? Someone changes the config in a switch. And really, the action you want to take is you want to rollback, right, you've identified that the upgrade has gone bad, and you basically want to roll back to the previous one.

    Steve Statler 9:19

    Okay, that makes sense. So I mentioned that we spoke a year and a half ago, what have you guys been doing for the last 18 months?

    Bob Friday 9:28

    Yeah, I think you know, if you're what I would say if I look at my BOE effort, and I've always called this kind of moving from a nice to have to must have, we've been getting increasingly more and more use cases, you know, from our customers. Right. And so I think last year, we talked about the Orlando what we're seeing in hospitality. So we're working with companies like turnout. Now. These are the guys who put the little tag if you ever go to the conference, and you see the little BLE tag on the back of your thing. Yeah. That's almost become common now. It's hard to go to Congress. is now where they don't put a BLE tag on the back of your tag to actually track you through to the conference in the event. So we're working closely with that, that started to emerge out there. The wayfinding use cases are starting to emerge where we're starting to see more and more people build mobile apps that want to have a more immersive wayfinding experience. We just won a large airline customer last year, you know, so they're starting to look at use cases such as you know, hey, if you're in the VIP lounge and you want to get the, you know, you're late for your plane, you want to basically do the Uber help you want to get the party from the lounge back to the things that we're starting to see that become more route reality. I think one of the general themes we're seeing across all the verticals, right now is kind of this concept of Uber help, right? Whether it's in retail healthcare, the concept of, Hey, someone needs help in this in particular, what we're doing down in the VA for Vets right now is, you know, if a vet is in a wheelchair, or on crushes any help, you know, how do we make it easier for them to get help when they emerge? You know, when they get to the hospital?

    Steve Statler 11:07

    Yeah, because hospitals are starting to be judged based on their performance and the user experience at the hospital. And I think we've all been to places where they need help to help us better. So that's great. So you basically can ask for assistance. And because you have, you have better antenna technology, you're using this AI you can spot where the person is in situations where there's more ambiguity. What are you guys doing with the Disney Disney Resort? They're the swan and dolphin.

    Bob Friday 11:41

    Yeah, so the swan dolphin is where we're doing the event planner notes that are mentioned around now, right, that is where, you know, they're looking to try to monetize location based services. So that's where we're leveraging our virtual beacon technology. And the machine learning right to to help better locate both the asset right that tag that beacon on the back of your tag. And that's where we're starting to leverage the phone. You know, getting people from A to be in a large, large conference room, I would say in the VA, the healthcare in general, for people who are familiar with healthcare, healthcare is always an early adopter of these types of technologies. There's a big trend in healthcare right now, where they're looking to BLE as kind of a standards base horizontal technology, similar similar to Wi Fi to really complement all those existing proprietary IR ultrasound technologies. And that's really, it's really opening up a whole nother set of use cases in the healthcare space. And I think the bigger thing is, you know, ble and location is kind of classically really resonates with the b2c, you know, anybody who has a business that's trying to serve as a customer. But I think probably the more important thing for me is I'm starting to see is even the carpeted enterprise. Customers are starting to show an interest in BLE, right? These are like your mainstay financial customers, corporate enterprise corporate campuses, where they're starting to look at Bally's kind of their facilities management, right? You know, as they move to open space, they want better visibility on how their employees are actually using the space, right? You know, because once you go to open space, people naturally select where they want to be. And so you start to see strange behavior where certain parts of the building don't get used for some reason, and sort that out.

    Steve Statler 13:29

    Well, so you have this sophisticated antenna structure for BLE, you're applying this extra analytical layer. But it seems like there's kind of two ways BLE can be used with your product. One is you have tags that you're spotting, they're attached to assets. And the other one is you're kind of generating, you're you flip it on its head, and there's an iBeacon event that has been sending that is being sent to someone's phone. So the first one, there may not be a phone involved at all. The second one, there's an app on someone's phone that you're triggering something based on. And the virtual beacon concept, as I understand it is you have a dashboard, which may be of a huge open space. And rather than having to send some guys out with step ladders to move stuff around, you basically just click on the map and say, here's where I want the triggers to be. Is that a reasonable summary of what the two sides are? Or feel free

    Bob Friday 14:29

    to? Yeah, I think you'd be if you look at the use cases, we're seeing, you know, there's the wayfinding use case of, you know, someone has an app on the phone. There's the asset where we actually listen to beacons trying to locate assets and abilities. There's probably a third use case emerging around IoT where BLE is starting to become kind of an IoT standard right? You know, we're a hospital like NYU may actually have hand washing dispensers that are all BLE enabled, right? And they're starting to use BLE to keep track of you know, when the battery when the liquid needs to be Replace. So those were you're starting to see Billy starting to become kind of your wayfinding your asset visibility and your IoT standard in enterprise business now?

    Steve Statler 15:11

    Yes. Well, I think this I'm hearing a lot of this asset tracking, I'm hearing a lot about this hand cleaning thing. It's so critical for medicals. But the thing that you guys are doing, which is very unusual, is this virtual beacon piece. Is that is that getting traction? Where do you Where have you found people need a virtual beacon verse versus asset tracking, which is, if you can do that more accurately, that's, that's great. But I'm interested in the virtual beacon bit.

    Bob Friday 15:42

    So the virtual beacon, you know, that actually started in the retail space, that is where, you know, you look at a retail store where they have to basically change up their marketing or their planet planning, you know, where they put the products. That's where we actually started the virtual beacon is, they found that they hadn't moved those beacons around every time they changed the store layout or the marketing program. That was the original use case for virtual beacons. Some of the more interesting ones we're seeing now is even in the theme parks, right? Especially around augmented reality, right, when you look at augmented reality, like in a grocery store, where they have visions of someone walking down the aisle and using their phone to actually augment what's on display shelf, there, you need enough virtual beacon to kind of understand where the person is in that store. So that's another use case where people are using virtual beacons to kind of roughly locate where someone is. And then I'll probably the in more interesting third ones I've had lately has been even the theme parks, you know, visualize a room with a mist infrastructure, where you can change the you know, find the Pokeyman, right? You basically you have different game formats where you could basically put virtual beacons anywhere in a warehouse with kind of an augmented reality experience, right. And you have to basically work your way through a maze. And you can basically change that maze up for different game formats. Fantastic. So those are examples where virtual beacons are actually starting to see traction now.

    Steve Statler 17:05

    Yeah. So you have a store reset, as they call it, where they move the shelves around. Last thing you want to do is to get the guys that deal with technology involved to on the ground, and so many examples where those spaces are dynamic. So okay, I think we've covered a lot. We'll wrap it up in just a minute. I guess actually, one things I should have done earlier is just disclose. We do have a common investor, Norwest in Wiliot, when the first time I spoke to you, I didn't work for Wiliot so my interest in what you do is got nothing to do with the fact that we have a common investor. But I do get to hear things are going very well for you guys. I think there's been it's really interesting with the beacon ecosystem, I kind of sense a little bit of negativity that things aren't going well. And that business is not good. And beacons were a fad. What? What's your view on that?

    Bob Friday 18:00

    So I'm still a big believer. And really, you know, for me, you know, when you look at indoor location, buildings, a key component kind of salt, removing the one the friction points. Because if you look at indoor location, like I said, the friction points have really been around the Overlay Network that typically it's in to some extent battery beacons or an overlay network. And that's why there's friction and resistance to that, you know, so we're solving that problem with converging Wi Fi and BLE together. The other problem with indoor location has always been the survey, it's always required a lot of survey to actually get it up and running. The machine learning piece of that has taken that piece of the friction out of point. And probably the third biggest friction we've had in the industry is really around, you know, what is Google apple? Where are mobile devices really going to support what indoor location technology are we going to see them adopt. For me BLE has become that Wi Fi. We looked at Wi Fi originally, right industry and Wi Fi location never really got there, right? It never really got to the performance level. And some of that's because the mobile device operators really never embraced Wi Fi as a location technology. I think we're seeing that with BLE I think Google Apple are starting to embrace BLE is kind of that location technology. Especially, I don't know if he kept track of the BLE 5.1 spec coming down the pipe. Right. Very interesting.

    Steve Statler 19:21

    It was so what's in that, that you think is of note now I think you know, of note is the AOA in the EOD, right angle, the rival angle of departure partially correct.

    Bob Friday 19:34

    Right and we're gonna see better location technology now I think we're we're all waiting to see what Apple does who knows Apple gonna embrace it. But you know, that is another Okay guys, we are going to see a much better location technology get into our phones here hopefully with ble 5.1 So I think that's another data point on you know, are we gonna see BLE go from a nice to have the must have?

    Steve Statler 19:58

    Yeah, so the thought of having angle of arrival support on a phone is kind of interesting. So what would that look like? Can you? I mean, why would you do that?

    Bob Friday 20:13

    Well, I think we will do it for better accuracy, right? I think, you know, you know, different during my days of location, right? Whether it's time angle, just the more information you have, the easier it is to make an estimate, right. And so if we have timing or phase information, all that is going to help on making the user experience better.

    Steve Statler 20:33

    Yes, yes. So so this is a feature that could add some a new feature for a new version of the iPhone, that would allow you to kind of use it as a sniffing finding device, can you because at the moment, you have a beacon on a shelf, and yeah, you know that you're getting warmer or colder. But if you had a phone that actually knew the angle, that the beacon was, you could start navigating through, you could pass through the big cold medicine shelf and locate the specific item that you're looking for and solve that problem, which would be amazing. It would be also very good for people that sell Bluetooth tags, which so I'm not averse to that as a use case.

    Bob Friday 21:13

    So I think I agree I'm, I'm still a, I still firmly believe that we're going to see Billy go from a nice to have the must have, I think, you know, we're seeing data points, right. You know, on my side, I'm seeing customer use cases emerge that are gonna embrace it. I think on the technology side, we're seeing the BLE SIG, no 5.1 Ai, all that is pointing in the right direction. I think in the tag, you know, the tag ecosystem, yourself included, right? We're starting to see a whole new generation of energy harvesting tags come down the pipe, all that goodness. And there was diversity starting to see the indoor map ecosystem evolve, right? We're starting to watch indoor mapping involved, which is kind of a key part for all the wayfinding use cases.

    Steve Statler 21:59

    So what's happening in the mapping space from your perspective? Now the I

    Bob Friday 22:03

    think we're seeing two things, right? I mean, we're seeing a ecosystem of map vendors come to the market, right? These are your jibes stream locus, lab map, map wise. Those guys were also watching Google and Apple kind of work out their strategy around indoor maps, right? Google and Apple are starting to kind of sort out how they're going to handle indoor mapping. So I think, I think it's still up in the air, whether you know, as indoor mapping gonna be solved by an ecosystem of players like here, you know, here just acquired Marcelo, you know, how's that gonna evolve over the next couple of years?

    Steve Statler 22:38

    Fascinating. Well, this has been a great update conversation. Bob, Bob Friday, CTO, co founder of Mr. Great insights, love hearing about your new technology where you see the market evolving. So let's check in. Let's check in with you again next year.

    Bob Friday 22:55

    Okay. Thanks, Steve. And yeah, hope to see you in the Next Conference.

    Steve Statler 22:59

    All right. All the best bet you have three songs before on your trip to Mars, you get another three. Which ones would you choose? Do you think?

    Bob Friday 23:13

    Do I take it easy - Eagles.

    Steve Statler 23:15

    Well, you have that one last time. So yeah. Yeah. The Eagles in the Grateful Dead. I can't take it.

    Bob Friday 23:24

    This time we would have to be Pink Floyd. Wish you were here and brain damage. Okay. Creedence Clearwater Have you ever seen the sun? Or ever since ever seen the rain? And then pirates is the other one I'm working on is was it band Morrison brown eyed girl?

    Steve Statler 23:41

    Fantastic. Well, those are great songs. Any particular reason why you would choose those ones?

    Bob Friday 23:46

    Because those are the ones I know the best now. I have plenty of time there all the way to Mars. They'd be doing me with you.

    Steve Statler 23:54

    Yeah, you might your avatar would probably expand significantly, but at least you'd be able to listen to those and you'd be able to model your performance on those. So why I'm full of admiration. I have tried to learn the guitar. People who've watched the podcast from my home have probably seen the guitar that is accumulating dust in the background. So the fact that you're actually doing it and you've stuck with it for a year and a half is pretty impressive.

    Bob Friday 24:19

    Now, I will tell you, Steve, if I learn to play guitar, anybody can learn to play guitar. That's the challenge I got when I got into it. One of my friends told me if I could do it anyway, good day.