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

0:00 i'm Manila Shiz U we'll be talking about

0:03 location and analytics today so we we

0:06 have heard about all the marvelous and

0:10 AI and agendic and all the stories on uh

0:14 what Bob and uh team is doing we heard

0:18 about what west and the Wi-Fi 7 uh

0:22 access point is bringing to the table we

0:25 also heard from you know access

0:27 assurance how we are innovating so the

0:30 same infrastructure same cloud same LLM

0:33 uh we are we have the capability of

0:36 adding customers to give a better

0:39 experience with interolocation services

0:42 realtime location services and the data

0:46 what you collect from that is used for

0:48 analytics in different different ways so

0:51 it is with the Wi-Fi 7 we have an access

0:54 point where you introduce the U ultra

0:57 wideband with dual BLE radios so we can

1:02 solve more uh you know end customer

1:05 problems or give solutions to uh not

1:09 just one use case of say asset

1:13 visibility or user engagement or you

1:15 know if you have a customer who has a

1:18 electronic shelf label all those kind of

1:21 things

1:22 so that is

1:25 the there are three different use cases

1:28 you are familiar with most of the use

1:29 cases be a app used for wayfinding beat

1:34 virtual notification if you're going

1:36 into a store and getting a promotion

1:39 coupon while identifying that you are in

1:41 the um Delhi versus a grocery aisle all

1:45 these things are happening with the user

1:47 engagement and there are new things what

1:49 are learning from customers what will

1:52 solve identifying a contextual uh

1:56 marketing

1:57 campaign on the asset visibility and the

2:00 people visibility there is so much

2:02 innovation we have been able to uh do

2:05 with machine learning for asset tags uh

2:07 we were able to get large scale

2:10 consistent um accurate data point for

2:14 staff duress or employee safety or for

2:17 assets going between you know zones to

2:20 find out where is the utilization where

2:23 is the uh you know crowding of uh

2:26 employees in the in your corporate

2:29 facility last but least uh the analytics

2:32 part we'll talk about this so this is

2:36 basically from last MFT it's a look back

2:39 of what we were doing and I have

2:42 mentioned a u small bit of we have been

2:47 constantly trying to improvise all this

2:49 unsupervised machine learning what we

2:52 have been done on SDKs for app can be

2:56 extended to asset tag so lot of these

3:00 you know

3:01 uh healthcare workers do have like

3:04 address which is B tags we introduced uh

3:08 un unsupervised machine learning where

3:11 it computes the TLF which is a path loss

3:14 formula which gives you a better

3:16 accurate location of the tags as well as

3:19 uh the badges the other one we have done

3:23 is uh we improved the location SDK so

3:27 there are uh requirements of contextual

3:31 uh awareness for simple as I'll talk

3:34 about some customer use cases we solved

3:37 you are walking into a store they want

3:39 to know whether you are in this facility

3:41 so they can give you um you know uh your

3:44 local inventory of that store but you

3:47 don't want to do a way finding so there

3:49 are certain improvements we have done to

3:52 make it lightweight there are

3:53 improvements we have done to uh give a

3:57 proactive lightweight and fit into all

4:00 your

4:01 SDKs i think Marvis client is star by

4:05 think Sudir and Slav already mentioned

4:08 so Marvis client uh also is kind of a

4:12 super app which provides multiple

4:14 personas where one of the persona is to

4:16 give you a location where you get indoor

4:19 location of that particular client which

4:22 can be used

4:23 for each vertical has different use

4:25 cases uh you don't get the same thing

4:28 when you are uh in a in a warehouse or

4:32 if you are in uh nonGPS facility that's

4:37 where we are expanding that last but

4:39 least so this is a unique uh innovation

4:42 we are the only ones to came up with

4:45 auto zones where using AI um computer

4:50 vision or image segmenting uh technology

4:54 to take your flow plan and classify that

4:57 as different zones this is again none of

5:00 this is uh this comes from the customers

5:03 where some of us have bigger square feet

5:06 there is a customer who has like

5:09 around 2 million square footage of floor

5:12 plan they want to have zones without

5:16 zones we don't get the location of

5:18 people going between the zones they want

5:20 to know how many people are coming into

5:22 conference room what's the usage of the

5:24 conference room and those are the things

5:26 we did last and what is so so this is

5:29 again auto placement we had uh great

5:33 presentation last time uh what we have

5:36 seen is there is

5:37 uh 90%age of the times when you have the

5:41 guided uh principle for auto placement

5:44 we are able to place them in the flow

5:46 plan accurately and 7 to 8%age we are

5:50 actually able to solve misplaced APS uh

5:54 you know fat fingering the MAC address

5:56 will be looking similar but you place AP

5:59 which is supposed to be on a different

6:01 floor or a different entry what happens

6:04 there is a lot of the location

6:06 calculation as well as you know when

6:08 you're troubleshooting lot of things

6:10 goes away those can be eliminated this

6:13 is where you know this year we are

6:15 looking at making it location aware AI

6:18 ops where you just need to give a floor

6:22 plan we want to make your life easier

6:25 give you auto placement ment give you

6:27 auto orientation how the AP is placed

6:30 where the 16 antenna array is facing to

6:34 the north those were the things we were

6:37 able to do uh focusing on for this

6:41 year so this is basically a demo where

6:44 we are showing how easy is to create a

6:47 auto zone you basically create a flow

6:50 plan just drag and drop it uh add an

6:53 image and you know start auto zone

6:57 It will if you have any names like say

7:00 this is a

7:01 ballroom and the other room is also a

7:04 ballroom you will take that name from

7:06 the image and create the zone uh saying

7:09 this is Portland ballroom and so on and

7:11 so forth why it is important is the

7:14 conference rooms may be spread across

7:16 now you can actually label them and see

7:19 uh what are the utilization of my

7:21 conference rooms these are big uh

7:24 decision making information where they

7:26 want to know there are some places in

7:28 your real estate which is not utilized

7:31 versus some of them is

7:33 overutilized so when you speak about

7:36 location uh we all have been aware of

7:40 the apps which has been using the B

7:43 virtual B the 16 antenna array for you

7:47 can track the B app based we can have B

7:50 tags you can also have the passive B

7:53 which is the you know devices you

7:55 walking around in your store or in your

7:57 office we also do the Wi-Fi and uh

8:00 connected and unconnected now we

8:03 introduced is the ultra wideband uh

8:05 again the same philosophy all standards

8:08 based uh we are sticking with some of

8:11 the ultra wideband standards and that

8:14 way we

8:16 uh nothing proprietary we are doing

8:21 yeah go ahead the ultra wide band I know

8:24 like earlier you talked about like a

8:26 multiarray uh ultra wideband antenna is

8:28 that prevent you from installing like

8:30 additional sensors because I know wide

8:32 band you need here the densify so it's a

8:36 completely different radio okay so this

8:38 has nothing to do with the B radio it's

8:41 a new chipset new radio and that's a a

8:45 different So you know how we have the

8:47 BT11 to complement like the Bluetooth

8:50 yes do you have something similar for

8:52 the ultra wide band or uh not yet okay

8:55 so okay right now it's AP47 is the first

9:00 access point

9:02 you might lead into this but a big So so

9:05 if it's coming up that's

9:07 fine part of location is knowing where

9:10 the APs are mhm are you going to talk

9:12 about how miss knows where your own APs

9:15 are yeah that is the one first I was

9:17 talking about auto placement but I will

9:19 talk about in detail

9:21 uh in the next slide and tied into

9:23 standard power and GPS and Yeah yeah

9:26 okay okay i got you and another question

9:30 for the Wi-Fi connected users um are we

9:33 talking RSSI or are we talking like fine

9:35 time measurements so no these are all uh

9:38 so we are different a little bit from

9:42 uh triangulation versus probability

9:44 surface so we do

9:48 PLF and path loss formula machine

9:51 learning and we do based on probability

9:54 surface that's on the B so Wi-Fi is all

9:57 the connected based on the triangulation

10:00 of APS or RSSI okay is that what you

10:04 want to talk about Keith there you go

10:06 okay so auto placement I gave you a

10:09 first slide of how auto placement has

10:12 been used in the field how it has been

10:15 getting good results or acceptance so

10:18 one of the key things is auto placement

10:21 now we are using util uh ultra wideband

10:24 so what is the purpose of ultra wideband

10:28 is you get much accurate uh location

10:32 accuracy

10:34 and it is

10:37 the

10:40 the two things there one the accuracy of

10:43 the data and efficacy of the data so

10:47 these two things is significantly high

10:50 compared to the Wi-Fi FTM based uh auto

10:54 placement so we are also uh using

10:59 virtual B arrays which we have I believe

11:02 you have seen the access point around

11:04 here so we have the latest which is the

11:06 diagram there which has the uh the newer

11:09 array which gives you if you can ask

11:12 this question can the UWB auto placement

11:16 can be just

11:17 uh by itself no we are actually using

11:20 the directionality of the antennas to

11:24 supplement for passive listening as well

11:28 so accuracy efficacy and scale along

11:33 with that the other advantage is there

11:35 is no disruption of Wi-Fi so you can do

11:38 it whenever you want to do you can you

11:40 know even do um every day to check that

11:43 your access point is in place i'll give

11:45 you a sample data of uh C of a run so

11:49 that

11:53 way this is actually a auto placement

11:56 run with ultra wideband so the blue is

11:59 uh the map placement and the green is

12:03 actually

12:04 uh the algorithm prediction so you can

12:06 see it's pretty accurate in most of the

12:09 parts but there is one

12:13 uh I don't see

12:15 this yeah this is the only place where

12:18 there is a small uh you know deviation

12:22 from the actual deployment that is

12:24 actually the the algorithm corrected the

12:27 map and if you look at the red dots

12:29 these are the uh three anchors used this

12:32 is 43 access points

12:35 uh And this gives you like 0.5 m

12:40 accuracy 95%age of the time i guess it's

12:44 my question is tied back to those red

12:45 ones mhm uh is there a GPS in all the

12:49 APs or only the red ones or is that I

12:54 mean that's a lot of GPS's to not be

12:56 working so AP 47 do have GPS in all the

13:00 APS that doesn't guarantee you that

13:02 you'll have the GPS signal right so we

13:05 will use the appropriate uh you know

13:09 anchors which have the the GPS signal we

13:12 will also pick from you know what is the

13:14 right one to get the minimum number of

13:17 anchors you want to add W just to give a

13:19 little context so Keith for this

13:21 particular customer uh so the algorithm

13:23 is automatically determining the anchors

13:25 right there's not this isn't anything

13:27 you know no customer input uh but for

13:29 this particular customer um it's all

13:32 AP47s with GPS uh about uh 10 to 15% of

13:38 the APs actually are receiving GPS

13:40 signal it's a you know indoor high-rise

13:42 kind of building um and uh and so you

13:46 you know the the GPS you know especially

13:49 indoor is

13:50 you you won't be able to leverage

13:51 anywhere um this auto placement with UWB

13:55 using GPS um uh is actually a method

13:59 that we've now uh has been approved uh

14:01 by FCC for us uh to use as you know as

14:04 part of our standard power um and so

14:06 this will you know this is this auto

14:08 placement with UWB and and a few other

14:10 methods will be the basis for kind of

14:12 some of our standard power to augment

14:14 where GPS uh is not filling in indoors

14:17 if a customer desires to use and so for

14:19 the standard power part of the the FCC's

14:21 algorithm is how many hops away from the

14:24 lock you have and so you get fuzzier

14:28 location yeah you have to take the worst

14:30 case yeah so how how does that affect

14:32 when you I mean if they're all GPS and

14:34 you even get a little bit y could you

14:37 have more locks more anchors right so

14:39 you could Yeah if your if your error is

14:42 very small as you go more hops your you

14:45 know your worst case is doesn't grow

14:48 exponentially let's say but if you're

14:50 uncertain of your location then you

14:53 introduce more ads

14:56 so so this is a question I've had for a

14:59 while how are you guys

15:01 exposing because as customer customers

15:04 begin deploying Wi-Fi 7 especially if

15:07 they do one for one replacements in in

15:10 buildings with concrete walls and and

15:12 minimal windows right getting getting

15:15 good anchors could absolutely be an

15:17 issue how are you

15:20 exposing a lack of good anchors to the

15:24 customers in via dashboard via messaging

15:27 whatever so that they know okay maybe we

15:30 do need to add some additional APs near

15:33 external walls or or whatever wherever

15:35 we have to to get that TPS

15:39 and and you you mean in the in the

15:41 context of standard power yes okay i

15:44 sorry yeah go ahead go sorry

15:47 analash's intent was not to go uh not

15:50 specific for standard power but Keith

15:51 had asked about it so I brought it up uh

15:53 and so in the context of standard power

15:56 uh we will have some alerting of hey

15:58 we're you know you have standard power

16:00 configured but we're not able to um make

16:04 the requests right there you we'll have

16:06 it's not fully baked yet but we'll have

16:09 some mechanism of hey I don't have I

16:12 don't have appropriate geoloccation I

16:13 can't do GPS or or I can't I can't make

16:16 the request because I don't have my

16:17 geoloccation okay thank you but you know

16:20 the intent is um we want to build a a

16:24 mesh as you know it's you especially for

16:27 indoor deployments where you're going to

16:28 use standard power which may not be all

16:30 that common we'll see uh we have

16:33 multiple methods um of kind of building

16:37 this APAP mesh uh and

16:40 so it with the intention that it's

16:43 unlikely that you wouldn't at least have

16:45 one AP in the vicinity with with GPS now

16:48 it's you know anything's possible but um

16:51 yeah the intent is just use what you

16:53 have and permeate as you know if you

16:54 have more great if you have just one

16:56 okay we'll work with that okay thank you

17:00 all

17:01 right okay let's switch gears and talk

17:05 about

17:06 uh the cool kid in the

17:09 block so we have this Marvelous client

17:12 we spoke about this in multiple

17:15 uh context and this is the different

17:19 personas of Marvelous client right so

17:21 this is the one what the access

17:23 assurance use for uh Knack onboarding

17:26 have the certificate installed you have

17:29 a enterprise user having their Windows

17:31 machine and using this for you know

17:35 telemetry finding uh trouble tickets and

17:40 the last one is a warehouse where you

17:43 have uh zebra devices like you know u

17:47 their handhelds what they are trying to

17:49 do their job and if at all I can get

17:53 location as well as the telemetry that's

17:56 where all these different uh personas us

17:58 combined together and one thing I just

18:03 want to point to here is uh this

18:07 particular data set what we are

18:08 collecting be it on the poster uh

18:11 identifying what's the radio driver uh

18:13 what's the operating version all those

18:16 things combined with uh the location of

18:20 that particular person when a roaming

18:23 issue happened this eliminates a lot of

18:26 hassle for IT administrators to manage

18:29 those devices in the saying that you

18:31 know uh usually when you have a

18:33 warehouse user calling for a ticket

18:36 saying that I have issues in my

18:38 application he comes back at least uh 20

18:41 minutes later to make that ticket by the

18:44 time we lost what was happening so I had

18:47 a customer who have around uh 100,000 of

18:51 these devices and they're using this to

18:54 find out okay if I get a ticket they

18:56 look at the location of that uh that

18:59 particular person for last whenever this

19:01 issue happened they look back and play

19:04 what are the other tickets where you had

19:06 the same problem at that same time or is

19:09 that just oneoff things or is that a in

19:12 environment uh issue where they can

19:15 address that's one uh use case of it can

19:18 eliminate a lot of troubleshooting boots

19:20 on the ground and things like that uh

19:23 the last one I want to emphasize here is

19:27 as Bob and Sudir already mentioned lot

19:31 of this information of what is your

19:33 battery at that time when this happened

19:36 what's your CPU utilization all this

19:38 telemetry is being uh taken in for our

19:42 large language model as well as for the

19:45 user himself to see this is what is

19:48 happening from uh you know an IT admin

19:51 to troubleshoot their

19:53 issues oh the quick one then um

19:56 obviously the client's got to be running

19:58 for you to get that telemetry data

20:00 client has to be Yeah this is only when

20:02 the client is on action so we are not

20:05 expecting

20:07 uh this is when you are working when

20:10 your uh you know associate is in your

20:13 warehouse yeah it has to be run so so I

20:16 I guess I could have the client on my

20:19 device but unless I set a client you're

20:21 not going to get that telemetry data so

20:23 is there any ways of trying to encourage

20:26 people to actually have the client on so

20:30 the there are two audience we are

20:33 serving here right one is

20:36 uh the Oh

20:39 sorry okay

20:41 thanks so this is the BY use case this

20:45 is just for certificate onboarding they

20:47 they they're on their own but when it

20:50 comes to enterprise or you know u

20:53 warehouse users these are all managed

20:55 devices right the uh IT team decides

21:00 what you want to have on the

21:02 applications on their managed devices

21:04 right so this is the only case where you

21:06 have BYOD you can opt in if you want to

21:09 send telemetry or not yeah no I guess

21:11 what I'm thinking is okay it's an

21:13 enterprise device you push the client to

21:15 the device that's what the um

21:17 corporation is going to do but how do

21:20 you get the user to start the client oh

21:22 you it's just an agent oh so even on a

21:25 phone it will always be it's just an

21:27 agent sitting and sending tele you don't

21:29 have to do anything on Android on

21:32 Android oh not on iOS yeah I was going

21:34 to say on you the app that's what I was

21:37 going yeah so it's only on uh Windows

21:41 Mac OS and Android okay yeah

21:45 okay okay anything else i will jump to a

21:51 little bit of some of the success

21:53 stories uh or case

21:58 we have uh a customer who has around

22:02 more than hundreds of sides which has

22:04 been using our access point turning our

22:07 uh you know BL radio to manage the

22:11 electronic shelf labels across the their

22:14 brick and motor chain we now have around

22:18 8 million of ESL tags being managed or

22:23 using our infrastructure

22:25 uh completely eliminating any overlay uh

22:28 networks overlay hardwares and that is

22:32 where you know you get more out of it's

22:34 not just the Wi-Fi you can do u extra

22:38 value add for your

22:40 business I will quickly jump I'm trying

22:44 to be sensitive of time uh This one is

22:47 just a the way we think about the app

22:50 integration is always uh people think

22:52 about you know way finding u it's not

22:56 just wayfinding we have more than three

23:00 customers who is trying to do like

23:03 detecting whether you are in the store

23:05 so that you can actually get access to

23:08 the inventories in that particular store

23:11 and once you go out of the store it

23:13 automatically want to clear them from

23:15 the card so that they can use it for the

23:18 next potential customer coming to the

23:20 store and this can be accomplished with

23:23 the you know Juniper's SD location SDK

23:27 which is utilizing the same VBLE

23:30 technology

23:31 uh and last one I want to share you was

23:34 this is one cool uh this is again the

23:38 same use case uh there is AR VR

23:42 wayfinding nowadays in the apps and

23:45 there are custom customers using our

23:46 location SDK to do uh you know camera

23:50 based uh augmented reality uh way

23:53 finding and another important use case

23:56 what they I brought it up here is there

24:00 are this particular customer has some

24:03 classified areas where you're not

24:05 supposed to bring your phones and we

24:07 dictate the location and what happens is

24:11 like they send like an amber alert like

24:13 a bus sound to say you're here you're

24:16 not supposed to have a digital device in

24:18 this premise so the way of thinking uh

24:21 location location is not like a nice to

24:25 have it is more towards you know adding

24:28 value and it's getting to a a must-have

24:30 in this

24:33 world all

24:35 right 3 minutes okay all right uh so we

24:41 collect a lot of data um and this comes

24:44 to a analytics where we have a complete

24:48 full stack from access point to SD van

24:52 uh so product called premium analytics

24:55 premier analytics gives you the IT

24:58 administrator or IT persona you can take

25:02 that data uh using for your capacity

25:06 planning your you know resource

25:08 management all the regular things for it

25:11 but the same data can be utilized by

25:14 line of business in a different angle

25:17 for example if I give the occupancy

25:20 analytics what I was talking about

25:22 before what an IT person will be looking

25:25 at may not be the same angle as a

25:27 marketing person will be looking at or

25:30 maybe a real estate will be looking at a

25:32 different angle where it will give you

25:35 where I should be sending my janitorial

25:37 stuff or you know how I need to move my

25:41 uh you know office space into more phone

25:44 booths or things like that

25:46 so

25:48 okay there are uh by default it is 13

25:52 months there are a lot of customers uh

25:54 so I want to give you a simple example

25:56 so there was a customer who want to go

25:59 from u a typical phones to wipe phones

26:03 and they were using the uh trend of last

26:07 13 months of Wi-Fi data to find out

26:09 where is the roaming happening so that

26:11 they can plan to have a natural

26:14 migration when uh uh the schools were

26:16 coming back from break want to have the

26:19 new devices and they had a very

26:21 successful uh implementation another

26:23 customer used for uh the network POE

26:26 budgeting and planning for having new

26:29 stores what's the trend how what's the

26:31 peak time and uh there are different

26:35 ways of looking at it from it uh one of

26:38 the customer was looking for finding out

26:40 they have uh they were trying to buy

26:44 some lease lines or you know uh internet

26:48 links and they want to know what is my

26:50 utilization what time I need to have my

26:53 peak utilization when you can and this

26:55 literally happened where they were

26:57 showing this data to the vendor service

26:59 provider and said this is what I'm using

27:01 I don't want to use this data so it

27:04 helps you on that so asset insights is

27:06 basically Basically whatever we had on

27:09 the asset tags now we want to give you

27:12 breakdown of you know historic data of

27:15 how these asset was going across your

27:19 facility which zone to which zone uh

27:21 there are use cases when it comes to

27:23 healthcare where they have devices which

27:26 has been rendered and kept so those

27:29 devices can is been utilized or not uh

27:32 those are the things you can get from

27:34 asset visibility knack uh or the access

27:37 assurance is the same as what we have

27:40 talked about before with that um I will

27:44 give it back to the uh thank you

27:46 everybody I know the room has become

27:48 really hot it must be the all the hot

27:50 technology that came out or maybe not uh

27:52 uh but but really we stand tall on the

27:56 shoulders of a lot of our customers that

27:59 have brought us this far and this

28:01 community honestly I think uh mobility

28:04 field day and you know WLPC uh have been

28:07 uh you know really pivotal in in helping

28:10 us shape and stay keeping us grounded in

28:14 in the innovation we're going after so

28:17 we really appreciate every one of you

28:18 delegates in the room thank you for the

28:20 participation thank you for the

28:21 encouragement thank you for the

28:23 partnership uh um for everybody online

28:26 uh we appreciate you making time uh for

28:28 being part of the Juniper presentation

28:29 and for all of our customers and

28:31 partners uh a humble thank you we would

28:34 not be here today uh um uh you know at

28:37 this place literally leading the

28:39 innovation for the industry for wire and

28:41 wireless uh without all of your uh help

28:43 and support so thank you everybody and

28:45 we'll see you at the next MFD