Juniper Location & Analytics: UWB, Dual BLE, PMA-Meeting/Security insights Marvis Client

I'm man chais. Uh, we'll be talking about location and analytics today. So we, we have heard about all the Marvel and AI and agenda and all the stories on, uh, what Bob and uh, team is doing.

We have heard about what was, and in the wifi seven, uh, access point is bringing to the table. We also heard from, you know, access assurance, how we are innovating. So the same infrastructure, same cloud, same LLM, uh, we are, we have the capability of adding customers to give a better experience with indoor location services, real time location services, and the data, what you collect from that is used for analytics in different, different ways.

So it is with the wifi seven, we have an access point where you introduce the, you ultra wide band with dual BLE radios. So we can solve more, uh, you know, end customer problems where, or give solutions to, uh, not just, uh, one use case of say, asset visibility or user engagement, or, you know, if you have a customer who has a electronic shelf label, all those kind of things. So that is the, there are three different use cases you're familiar with.

Most of the use cases where be, uh, app used for way finding be, uh, uh, virtual notification if you're going into a store and getting a promotion coupon while identifying that you're in the, um, deli versus a grocery aisle. All these things are happening with the user engagement and there are new things. What are learning from customers?

What will solve identifying, uh, contextual, uh, marketing campaign on the asset visibility and the people visibility? There is so much innovation. We have been able to, uh, do with machine learning for asset tag.

Uh, we were able to get large scale consistent, um, accurate data point for staff dur or employee safety or for assets going between, you know, zones to find out where is the utilization, where is the, um, you know, crowding of, uh, employees in the, in your, uh, corporate facility. Last but least, uh, the analytics part, we'll talk about this. So this basically from last, uh, MFT, I'd say look back of what we were doing and I have mentioned, uh, uh, small bit of, we have been constantly trying to improvise all this unsupervised machine learning.

What we have been done on STKs for app can be extended to asset tag. So a lot of these, um, you know, uh, healthcare workers do have like badges, which is BLE tags. We introduced, uh, UN unsupervised machine learning where it computes A TLF, which is a path loss formula, which gives you a better accurate location of the tag as well as, uh, the badges.

The other one we have done is, uh, we include the location, SDK. So there are, uh, requirements of contextual, uh, awareness for simple as I'll talk about some customer use cases we solved. You are walking into a store, they want to know whether you are in this facility so they can give you, um, you know, uh, your local inventory of that store, but you don't want to do a way finding.

So there are certain improvements we have done to make it lightweight. There are improvements we have done to, uh, give a proactive, uh, lightweight and, uh, fit into all your SDKs. I think Marvis client is, uh, star by things to, and, uh, SLA already mentioned.

So Marvis client, uh, also is kind of a super app which provides multiple personas where one of the persona is to give you a location where you get indoor location of that particular client, which can be used for each vertical has different use cases. Uh, you don't get the same thing when you are, uh, in a, in a warehouse or if you're in, uh, non GPS facility. That's where we are expanding that last but least.

So this is a unique, uh, innovation. We are the only ones to came up with auto zones where using ai, um, computer vision or image segmenting, uh, technology to take your flow plan and classify that as different zones. This is, again, none of this is, uh, this comes from the customers where some of us have bigger square feet.

There is a customer who has like, uh, around 2 million square footage of floor plan. They won, want to have zones without zones. We don't get the, uh, location of people going between the zones.

They want to know how many people are coming into conference room, what's the usage of the conference room, and those are the things we did last. And what is so, so this is again, auto placement. We had, uh, great presentation last time.

Uh, what we have seen is there is, uh, 90 percentage of the times when you have the guided, uh, principle for auto placement. We are able to place them in the flow plan accurately and seven to eight percentage. We are actually able to solve misplaced aps, uh, you know, fat fingering.

The Mac address will be looking similar, but you place the ap, which is supposed to be on a different floor or a different entry. What happens there is a lot of, uh, the location calculation as well as, you know, when you're troubleshooting, a lot of things goes away. Those can be eliminated.

This is where, you know, this year we are looking at making it location aware AI ops where you just need to give a floor plan. We want to make your life easier, give you auto placement, give you auto orientation, how the AP is placed, where the 16 anti RAs facing to the north. Those were the things we were able to do, uh, focusing on for this year.

So this is basically a demo where we are showing how easy is to create a auto zone. We basically create a, a flow plan, just drag and drop it, uh, add a image and, you know, start order zone. It will, if you have any names, like say this is a ballroom, um, and other room is also a ballroom, you will take that name from the image and create the zone, uh, saying this is ballroom and so on and so forth.

Why it is important is the conference rooms may be spread across. Now you can actually label them and see, uh, what are the utilization of my conference rooms? These are big, uh, decision making information where they want to know.

There are some places in your real estate which is not utilized versus, uh, some of them is overutilized. So when you speak about location, uh, we all have been aware of the apps, which has been using the BLE virtual BLE, the 16 an array for, you can track the BLE app based. We can have BLE tags.

You can also have the passive BLE, uh, which is the, you know, devices you are walking around in your store or in your office. We also do the wifi and, uh, connected and unconnected. Now we introduced, uh, ultra wide band.

Uh, again, the same philosophy all standard spaced. Uh, we are sticking with some of the ultra wide band standards and that way we, uh, nothing proprietary we are doing. Yeah, go ahead.

The ultra wide band, I know like earlier you talked about like a multi array, uh, ultra wide band internet. Yeah. Does that prevent you from installing like additional sensors?

'cause I know a tri band, you need to densify The sensor. It's a completely different radio. Okay.

So this has nothing to do with the BLE radio. It's a new, it's an additional headset new radio, and that's a different, so you Know how we have the BT 11 to compliment like the Bluetooth? Yes.

Do you have something similar for the ultra wide band, or, Uh, not yet. Okay. So, okay.

Right now it, AP 47 is the first day access point. Keith, You might lead into this, but a big, so, so if it's coming up, that's fine. Part of location is knowing where the aps are.

Mm-hmm. Are you gonna talk about how Miss knows where your own aps are? Yeah, that is the one first I was talking about auto placement, but I will talk about in detail, uh, in the next slide and Tied into standard power and GPS and Yeah.

Yeah. Okay. Okay, I got you.

And Another question for the wifi connected users. Um, are we talking RSSI or are we talking like fine time measurements? So no, these are all, uh, so we are different little bit from, uh, triangulation versus probability surface.

So we do, uh, PLF and path loss formula machine learning, and we do based on, uh, probability surface that's on the BLE. So wifi is all the connected based on the triangulation of APS or RSSI. Okay.

Is that what you want to talk about, Keith? There you go. Okay.

So auto placement, uh, I give you a first slide of how auto placement has been used in the field, how it has been getting good results or acceptance. So one of the key things is auto placement. Now we are using util uh, ultra wideband.

So what is the purpose of ultra wideband? As you get much accurate, uh, location accuracy, and it is the, the two things there. One, the accuracy of the data and efficacy of the data.

So these two things is significantly high compared to the wifi FTM based, uh, uh, auto placement. So we are also, uh, using virtual BLE arrays, which we have, I believe you have seen the access point around here. So we have the latest, which is the diagram there, which has the, uh, the newer array, which gives you, if you can ask this question, can the UWB auto placement can be just, uh, by itself?

No, we are actually using the, uh, directionality of the antennas to supplement for passive listening as well. So accuracy, efficacy and scale Along with that. The other advantage is there is no disruption of wifi.

So you can do it whenever you want to do, you can, you know, even do, um, every day to check that your access point is in place. I'll give you a sample data of, uh, uh, cus offer run. So It's actually a, a auto placement done with the ultra wide band.

So the blue is, uh, the map placement and the green is actually, uh, the algorithm prediction. So you can see it's pretty accurate in most of the parts, but there is one, uh, I don't see this. Yeah, this is the only place where there is a small, uh, you know, deviation from the actual deployment.

That is actually the, the algorithm corrected the map. And if you look at the red dots, these are the, uh, three angles used. This is for 43 access points.

5 meters accuracy, 95 percentage of the time. I, I guess it's my question's tied back to those red ones. Mm-hmm.

Uh, is there GPS in all the aps or only the red ones, or is that, I mean, yeah, that's a lot of GPSs to not be working. So AP 47 do have GPS in all the AP aps. That doesn't guarantee you that we'll have the GPS signal, right?

So we will use the appropriate, uh, you know, anchors which have the, the GPS signal. We will also pick from, you know, what is the right one to get the minimum number of anchors when address. Uh, Just to give a little context.

So Keith, for this particular customer, uh, so the algorithm is automatically determining the anchors, right? There's not, this isn't anything, you know, no customer input. Uh, but for this particular customer, um, it's all AP 40 sevens with GPS, uh, about, uh, 10 to 15% of the aps actually are receiving GPS signal.

It's a, you know, indoor high rise kind of building. Um, and, uh, and so, you know, the, the GPS, you know, especially indoor is not you, you won't be able to leverage anywhere. Um, this auto placement with UWB using GPS, um, uh, is actually a method that we have now, uh, has been approved, uh, by FCC for us, uh, to use as, you know, as part of our standard power.

Um, and so this will, you know, this is this auto placement with UWB and, and a few other methods will be the basis for kind of some of our standard power to augment where GPS, uh, is not filling in indoors if a customer desires to use. And So for the standard power part of the, the fccs algorithm is how many hops away from the lock you have. And so you get fuzzier location.

Yeah. You have to take the worst case. Yeah.

So how, how does that affect when you, I mean, if they're all GPS and you even get a little bit Yep. Could you have more locks, more anchors, right? So you could, yeah.

If your, if your error is very small as you go more hops know your worst case is, doesn't grow exponentially, let's say. But if you're uncertain of your location, then you introduce more ads. Yeah.

Yes. So, so this is a question I've had for a while. How are you guys exposing?

Because as, as customer, customers begin deploying wifi seven, especially if they do one for one replacements and, and buildings with concrete walls and, and minimal windows, right? Getting, getting good anchors could absolutely be an issue. How are you exposing a lack of good anchors to the customers and via dashboard, via messaging, whatever, so that they know, okay, maybe we do need to add some additional AP near, near external walls or, or whatever, wherever we have to, to get that PS lock And, and you you mean in the, in the context of standard power?

Yes. Okay. I Sorry.

Yeah, Go ahead. Sorry. I lost, his intent was not to go, uh, it was not specific for standard power, but Keith had asked about it, so I brought it up.

Uh, and so in the context of standard power, uh, we will have some alerting of, Hey, we're you, you know, you have standard power configured, but we're not able to, um, uh, make the requests right there. We will have, it's not fully baked yet, but we'll have some mechanism of, Hey, I don't have, I don't have appropriate geolocation. I can't do GPS or, or I can't, I can't make the request 'cause I don't have my geolocation.

Okay, Thanks. But you know, the intent is, um, we want to build a, a mesh, as you know, it's, especially for indoor deployments where you're gonna use standard power, which may not be all that common. We'll see, uh, we have multiple methods, um, of kind of building this AP to AP mesh.

Uh, and so it, with the intention that it's unlikely that you wouldn't at least have one AP in the vicinity with, with GPS now it's, you know, anything's possible. But, um, yeah, the intent is just use what you have and permeate as you know, if you have more, great, if you have just one. Okay, we'll work with that.

Okay. Thank you. Alright.

Okay, let's switch gears and talk about, uh, the cool kid in the block. So we have this marvelous client. We spoke about this in multiple, uh, contexts, and this is a different personas of marvelous client, right?

So this is the one what the access assurance used for, uh, NAC onboarding have the certificate installed. You have a enterprise user having their Windows machine and using this for, you know, telemetry, finding, uh, trouble tickets. Uh, and the last one is, uh, warehouse where you have, uh, devices like, you know, uh, their handhelds, what they're trying to do their job.

And if at all, I can get location as well as the telemetry. That's where all these different, uh, personas combined together. And one thing I just want to point to here is, um, this particular dataset, what we are collecting and be on the poster, uh, identifying what's the radio driver, uh, what's the o operating version, all those things combined with, uh, the location of that particular person, when a roaming issue happen, this eliminates a lot of hassle for IT administrators to manage those devices.

In the saying that, you know, uh, usually when you have a warehouse user calling for a ticket saying that I have issues in my application, he comes back at least, uh, 20 minutes later to make that ticket by the time we lost what was happening. So I had a customer who have around a hundred thousand of these devices and they're using this to find out, okay, if I get a ticket, they look at the location of that, uh, that particular person for last, whenever this issue happened, they look back and play, what are the other tickets where you had the same problem at that same time? Or is that just one off things?

Or is that a environment, uh, issue where they can address, that's one, uh, use case of it can eliminate a lot of troubleshooting boots on the ground and things like that. Uh, the last one I want to emphasize here is, as Bob and already mentioned, a lot of this information of what is your battery at that time when this happened, what's your UP utilization? All this telemetry is being, uh, taken in for our large language model as well as for the user himself to see this is what is happening from, uh, you know, an IT admin to troubleshoot their issues.

Oh, quick one then, um, obviously the client's got to be running for you to get that telemetry data. Client has to be, yeah, this is only when the client is on action. So we are not expecting, uh, this is when you are working, when you are, uh, you know, associated in your warehouse.

Yeah. It has to be run. So, so I, I guess I could have the client on my device, but unless I start the client, you're not gonna get that telemetry data.

So is there any ways of trying to encourage people to actually have the client on? So the, there are two audience we are serving here, right? One is, uh, uh, the, oh, sorry.

Okay, thanks. So this is a BYD use case. This is just for certificate onboarding.

They, they're, they're on their own, but when it comes to enterprise or you know, um, warehouse users, these are all managed devices, right? The, uh, IT team decides what you want to have on the applications on their managed devices, right? This is the only case where you have BYOD.

You can opt in if you want to send telemetry or not. Yeah, no, I guess what I'm thinking is, okay, it's an enterprise device. You push the client to the device.

Mm-hmm. That's what the, um, corporation's gonna do. But how do you get the user to start the client?

Oh, you, It's just an agent. Oh, so even on a phone, it will always be, Yeah, it's just an agent sitting and sending to them. You don't have to do anything On Android.

On Android. Oh, Not on iOS. Yeah, I Was gonna say on not on iOS, the app.

That's what I was gonna, Yeah, so it's only on, uh, windows, macros and Android. Oh, okay. Yeah.

Okay. Okay. Any, anything else?

I will jump to a little bit of some of the success stories, uh, or case we have, uh, a customer who has around more than hundreds of sites, which has been using our access point, turning our, uh, you know, BLE radio to manage the electronic shelf labels across the, their brick and mortar search chain. We now have around 8 million of ESL tag being managed or using our infrastructure, uh, completely eliminating any overlay, uh, networks, overlay hardwares. And that is where, you know, you get more out of.

It's not just the wifi you can do, uh, extra value add for your business. I will quickly jump, I'm trying to be sensitive of time. Uh, this one is just, uh, the way we think about the app integration is always, uh, people think about, you know, way finding.

Um, it's not just way finding. We have more than three customers who is trying to do, like detecting whether you are in the store so that you can actually get access to the inventories in that particular store. And once you go out of the store to automatically want to clear them from the card so that they can use it for the next, uh, potential customer coming to the store.

And this can be accomplished with a, you know, Juniper's, uh, s st location SDK, which is utilizing the same VBL e technology. Uh, and last one I want to share, was this the one cool? Uh, this is again, the same use case.

Uh, there is ar VR way finding nowadays in the apps, and there are customers using our location, SDA to do, uh, you know, camera based, uh, augmented reality, uh, way finding, and another important use case. What they, I brought it up here is there are, this particular customer has some classified areas where you're not supposed to bring your phones and we dictate the location. And what happens is, like they send like an amber alert, like a bus sound to say you are here.

You're not supposed to have a digital device in this premise. So the way of thinking, uh, location, location is not like, uh, nice to have. It is more towards, you know, adding value and it's getting to a, a must have in this world.

Alright? Three minutes. Okay.

Alright. Uh, so We collect a lot of data, um, and this comes to a analytics where we have a complete full stack from Access point to sd-wan. Uh, so product called premium analytics.

Premier Analytics gives you the IT administrator or IT persona. You can take that data, uh, using for your capacity planning, your, you know, resource management, all the regular things for it. But the same data can be utilized by line of business in a different angle.

For example, I give the occupancy analytics, uh, what I was talking about before, what an IT person will be looking at may not be the same angle as a marketing person will be looking at, or maybe a real estate will be looking at a different angle where it'll give you, uh, where I should be sending my janitorial staff or you know, how I need to move my, uh, you know, office, uh, space into more phone booth or things like that. So, Okay, There are, uh, by default it is 30 months. There are a lot of customers.

Uh, so I want to give you a simple example. So there was a customer who want to go from, um, a typical phones to wipe phones. And they were using the, uh, trend of last 13 months of wifi data to find out where is the roaming happening so that they can plan to have a natural migration when, uh, the schools were coming back from break, want to have the new devices, and they had a very successful implementation.

Another customer used for the network, POE, budgeting and planning for having new stores. What's the trend? How, what's the peak time?

And, uh, there are different ways of looking at it from it. Uh, one of the customer was looking for finding out they have, uh, they were trying to buy some lease lines or, you know, uh, internet links. And they want to know what is my utilization?

What time I need to have my peak utilization when you can. And this literally happened where they were showing this data to the vendor service provider and said, this is what I'm using. I don't want to use this data.

So it helps you on that. So asset insights is basically whatever we had on the asset tax. Now we want to give you breakdown of, uh, you know, historic data of how this asset was going across your, uh, facility, which zone to which zone.

Uh, there are use cases when it comes to healthcare, where they have devices which has been rendered and kept. So those devices can, is being utilized or not. Uh, those are the things you can get from asset visibility, nac, uh, or the access assurance's the same as what we have talked about before.

With that, um, I will give it back to the, Uh, thank you everybody. I know the room has become really hot. It must be the, all the hot technology that came out, or maybe not, uh uh, but, but really we stand tall on the shoulders of, uh, a lot of our customers that have brought us this far and this community, honestly, I think, uh, mobility Field a and, you know, WLPC, uh, have been, uh, you know, really pivotal in, in helping us shape and stay, keeping us grounded in, in the innovation we're going after.

So we really appreciate every one of you delegates in the room. Thank you for the participation. Thank you for the encouragement.

Thank you for the partnership, uh, um, for everybody online. Uh, we appreciate you making time, uh, for being part of the Juniper presentation and for all of our customers and partners, uh, uh, humble thank you, we would not be here today. Uh, um, uh, you know, at, at this place literally leading the innovation for the industry for wire and wireless, uh, without all of your, uh, help and support.

So thank you everybody, and we'll see you at the next MFD.