AI-Enhanced Workflows with Cisco Catalyst Center

Alright, welcome everyone. Thanks for everyone being here.00:11This is another edition of our Cisco expert insight series, but this installment's gonna be AI enhanced workflows, right? And I'm joined here. My name is John mcdon I'm joined here with my friend and colleague May. We're gonna go through this presentation with you to take a look at some of these.00:15Nifty, you know, weapons that we have inside catalyst center when it comes down to AI enhanced workflows.00:36All right. And so as we start here, this has been a great series. I'm pretty sure that, you know, some of you guys that are here has, you know, had the pleasure to take a look at some of our series that we've had, like, you know, beyond deployments and simplifying wireless operations, and now we're here, right? So like, and a lot of times.00:44The way we.01:04Run these sessions here is that we try to make sure that each one of these sessions build off each other, right? So as we come here, you get to learn a little bit of each one of these sessions here and now we're getting into the I enhanced workflows so that we can, you know, build off those knowledges that we have there, ok?01:05And before we start, I just wanted to let you know that you can scan this QR code and I'm gonna get my mug out the way so you can do that. And you can ask any question that you have for our panelists, they'll answer any questions that you have there once you scan in. So that's where all your questions can be directed as we go along.01:25And so let's get started. We got a bunch to cover here and here's our agenda for today. All right, so what I'm gonna do is I'm gonna start out, we're gonna kind of split the baby here split the baby here. We're gonna, I'm gonna go all the way from the AI ops overview. We're gonna talk a little bit about issues and events.01:46Network reasoner and baseline, which is one of the main AIs that we're, we'll start with within that regard, but we also wanted to add into some of the historical data that you can get within Catalyst Center, you know, and getting some peer comparisons and site analytics and AIR and which we're gonna touch on too as well. We want to get a good.02:06Demo at the end of May two as well. We'll touch on some AI endpoint analytics too as well. All right, so this is our full agenda here, so.02:26Let's keep going here. All right, so what is Catalyst Center AI ops alright so that's one of our personas inside Catalyst Center. We have a few personas, I believe it's four. The AI ops is going to be part of our, you know, AI operations that we can use within these work.02:34And we're gonna go through each one of those here.02:54So, with our focus here inside Catalyst Center, we will show, you know, some of the workflows that we have here and what we'll do is that we'll take a look at the architecture that we have for the catalyst center. We can see some of the network analytics that I was talking about, the endpoint analytics and AI.02:57RM and event analytics. And as all of us know here, like.03:17We're feeding a lot of data from our network infrastructure into the catalyst center, and as it works within the back end of our, you know, controllers and, you know, for the whole entirety of the controller. What's need to be understood is that a lot of the data that we're sending into Catalyst Center, we all know that is your data of your network.03:23Work. It's going to be anonymized data that we send to the cloud to get, you know, AI working the way we want to, but it's also going to be encrypted, so just to let you know, you have your, your piis and your security processes so your data is safe inside the cloud when it's as it comes through. So it encrypts before it goes out.03:43Decrypts once it gets into cloud and it's it's secure, right? So I just wanted to let you know that based off the, the process travels for the data there too as well. So but it's correlating for this, these particular personas that we're gonna touch on here today.04:03Okay, and so with Cisco AI ops, there's a view of complete network statuses as a glance. That's what as we see, right? So, with each one of these here, the clients, network devices, infrastructure connect connectivity, applications, all these gather up, you know, so that you can.04:20Pretty much, you know, slow down the noise and all those false positives that you get for the data that that that's so massive inside catalyst center, right? So all this correlated data needs to be handled in the right way so that you're not bogged down in and being bothered by these processes there, right? So you know with.04:40Clients you have your wireless and wire clients you know a lot of data that you're getting that's needs to be reliable and needs to be understood and on a macro level and a micro level two as well.05:01Okay, and so with the network devices, of course you have your, you know, operations and domain information that you get there too as well. With the infrastructure connectivities, we have a lot of those services there that that's gonna be coming in. So we want to make sure that all of that is reliable and, and you get a lot of these issues that.05:12Is detected inside an organized way.05:32Okay, and some of you may have seen this slide before, but, you know, with the process of the device coming into Catalyst Center, you know, a couple of things need to happen for as a prereq, and that is enabling Netcom yang. And so you're.05:35Your devices there in order to get the the data correlated and get the telemetry going. Another one is making sure that the cert is downloaded into the device, and if that happens.05:55You know, even from, you know, VM standpoint, KBM on prem or AWS or Azure, you know, if we have these two things here, we're able to correlate the datas and make sure that the telemetry is running in and we can do a lot of things with that telemetry and.06:07Push scripts from for the telemetry process data and we're gonna talk a little more about how you can check and make sure that the telemetry is set and how you can make sure that the subscription and the certs are downloaded to as well.06:27Yeah. And again, like, like I mentioned, you know, from I was talking about the data being correlated into our widgets here, but this is more so getting into AI workflows. We still get the data from your infrastructure into Catalyst Center, but then the way I explain it is that catalyst center is going to be.06:39Your network, right? Because, you know, we all here has different networks and has personalities in it, changes and moves up and down throughout the day. And what Catalos Center is gonna do is gonna be friend your network and as you, as any friend you will learn it, you will learn the personality or the persona in acting.06:59Acting process of what your friend is and then.07:20When you have that baseline, you're able to understand, you know, what's, what's off, what's like you know what's off on our network. And that's what's gonna do, it's gonna learn your network, and it takes a few days to do that or say a few weeks to do that. Once you have that, that data, then you can start developing anomalies.07:23Right, you know, some issues that's happening and it gives you some good insights on what's what's going on in your network from an AI standpoint. Then you can compare it to different sites, you can compare it to different devices and and get a good predictive, you know, workflow on and solving problems faster, right? So that's the main thing. So you get the.07:43You get the detections of, you know, that's, that's running off that baseline. You get to get those trends and insight that you have there and then you can compare it to different sites and devices. That's the main path of of just using AI network analytics specifically.08:04But it gets deeper, but we're gonna do is gonna show you how to enable it. This is gonna be one of our calls to actions later on, and that's just enabling it here inside systems and settings. What you'll do is you'll just toggle in this for AI network analytics, but then also AIRM.08:21We'll talk more about that. And then you just make sure that the storage for your, your cloud is, is selected too as well Instead the bottom of the screen there too as well. So if anyone's logged into galleson, you can see it in the system settings here where you can select your region, ok?08:38And so this is an FYI slide, right? So, you know, if you, if you need to be air gapped, you can do so as well. Now you if you're using AI for, you know, if you're.08:54Want to use AI for any reason inside a air gap, you kinda don't get missed out on that, right? Because we have machine reasoning, right? And we're gonna talk more about that about how the machine reasoning use machine learning and just troubleshoot your network from a visibility standpoint, right? So we have the visibility standpoints from three 60 degree views.09:09Health dashboards, you know, the historical data that we get, but also, you know, the observability while will we get those integrations in for, you know, external type selections, right? But then we get into the insights and actions that I was talking about like intelligent captures, network reasoning, you know, you know, issue events and some AI driven EV.09:29Events there too as well. So we'll be able to, still.09:49Take advantage of some AI stuff, but you still can be, you know, air gap within, you know, a cloud tettered process there too as well. So we have that. All right. And so I wanted to touch based on a lot about health and historical data that we have there too as well because what that does is that it's going.09:53Gonna build this into what our baseline is coming from. How is it built into getting into a lot of the AI network analytics like trends and insights, heat maps, peer comparisons and and what have you, right? And AI RM too as well. So let's take a look at what we get from a 3603 view and what we need.10:13From a prereq here, right? So couple prereqs is making sure the device controllability is on here. This will be able to give you processes on this is not enabled, you know, by default, so we want to enable that too as well. And then also the ports that we need open in between the the certain devices.10:32In in catalyst center. So your wireless devices in between catalyst center, we need these ports open 8443162, some UDP ports and TCP ports that need to be open between your devices and Catalyst Center shows here too as well, right? So these are the prereqs for your your name visibility.10:53Okay, keep that in mind, and overall, you know, it's just our full dashboard about health here, right? And you, you get to pick different sites and that's important because we wanna talk about site analytics here. And when you.11:12Select this different site, obviously you're able to go beyond different time zones here for not to different time zones but different times of the day, different times of the week, different times, a 3 h span. But you also can go back 30 days here and get some historical data here too as well. And as you can see, you know, we have 8080 5 % health.11:31Health in this latest process for access points specifically, and we can drill down everything related to that if we wanted to. But then also I wanted to direct your attention to this site analytics.11:51Widget here, right? So we have this site analytics widget and, it's broken down into onboard attempts, onboarding durations, connection speeds, and enrollment attempts. You can get a good percentage of health of that on any type of location that you want. Right now we're doing, we see London here, we can see.12:04All of the health within that site analytic there that we that we select here based off how that looks. All right, so that's like the widget here that most of us are already comfortable with. But if we drill down we can take a look at some of the KPIs that's gonna take, you know, take part inside our AI analytics there.12:23And that's the the AP health details, right? And so we click into this wireless network, you can, you know, have the AP overall health that we see here and we can get into a AP or device three 60 for the access point and get a good look of what's going on here. So you can see this here just how you can get to the three 60.12:43Degree view page, but at a glance we get to see what's the CPU utilization, the summary of the radio utilization, you know, whether it's fair, whether it's, you know, whether it's worse, we can get a good trend of what the APs are running there too as well. So we we've seen this before with the timeline there that we see.13:03And as we all know, each one of these three 60 degree views have a timeline travel tool, right? And we can drill down into any type of timeline, but what I wanted to point out is in addition to the AP health scores, like the CPU utilization and air qualities and.13:22And the noise processes there. We also want to make sure this is where we can manage the telemetry status alright and we're gonna talk more about how you can check and see that your telemetry status is good so you can see it in the dashboard that it shows good here too as well. But if it didn't show good.13:42I'll show you how to, you know, check inside your devices to make sure that the telemetry is set up. So, so couple of show commands that you're able to do. But as we went through here, we get to see some of the Cisco the system resources, the noise dbmiss as well.14:00All the information here that we have. So each one of these parameters are gonna give you some insight on.14:20Different time zone or different time spaces inside the three 60 degree view and we have a client through three 60 degree view as well too as well, so we'll be able to see that and standalone APs, a lot of KPIs play a part inside getting these, this data correlated.14:26And so with this data we can do is that we can pretty much control the threshold, right? So it's configurable, right? So the thresholds for the KPIs were able to dial it down or dial it up based off, you know, the noise that that we care about, and make sure that it's relevant.14:45Data that we have there. So for air quality for 2.4 ghz we can manipulate what where we want that percentage to stay under, right? Or over.15:05Let's just say over for, you know, for good quality here, right? And that's where we have that set. And so from a network health standpoint overall, we're able to we're able to drill down into any of these time travel tools here and just manipulate this data here to see.15:15See exactly what happened to the specific time here, and we can take a look at everything relative to that, right? So, I don't know this is.15:35More common that you guys have seen, right? But, you know, from an access point standpoint, you can take a look at the utilization, what's poor, what's the noise quality there? You can take a look at the APs that's up and down, you know, each one of those that you can take take a look at the details for.15:44Okay, and of course from a wireless standpoint, you can take a look at the client's view, right? So we can drill down for whatever location that we want to. We can filter by SSID. We can take a look at some of the details in the connection statuses, right? So some of the fair connectivities that we see here, so we want to be looking at.16:03The others that's about 20 % that's and critical alright for a wireless standpoint, we can drill down from here.16:23And then the overall client health dashboard here, right? So each one of these widgets has that processes for having trends that we can take a look at so we could filter on the metrics for the for each one that's that's presenting here and we can go in more detail and seeing that trend there too as well. So our.16:33SNR, we have client roaming times. You can see that it flags whatever's over the threshold.16:54Again you can manipulate or configure that threshold so you can know that it's relevant for you. And we have connectivity for physical links there too as well. That was a question that one of my customers had just recently there too as well. This is a widget that's pretty good to see some errors inside the links that you.17:01That there too as well.17:21This is from a client standpoint here. All right, then moving down into the issues and events, and we're gonna touch on that so the issues and events and how to identify the problems inside your network is using the machine reasoning process that we have there, right? So unlike traditional systems.17:23So who can see only up and down events, we're able to see more than that, right? And.17:43These are some of the ways we can use AI here, right? So, but it can utilize AI to analyze the network environment, right? And again, it's learning your network environment, so it's it's gonna be able to correlate for you, the, you know, and prioritize different issues, right?17:50So what it does is that it takes many issues.18:10And bring it into our correlation of one single issue, right? And it the way you able to see that is from this toggle button, we can turn this toggle button on for AI driven issues that we have here. And so as long as the the AI cloud is, is set and enabled, we're able to get these.18:14Issues here, right? So from aP1P2 standpoint, you can see the issue type.18:34And then you also can see the the AI icon here and that's how you know that it's correlated from many issues using AI to a single issue, right? You can see what areas being impacted and you can see it's global here. You can see what type of site is being impacted, what type of P one issues that it had.18:39As you can see here, there's NO P one issues, but we have twelve that are open here that we can take a look at and we could just filter and say anyone the P1P2P3 issues here.18:59All right. These are what it's actually showing some of your assurance issue that help identifies it's identifying root causes, which we'll talk about how it looks reducing pretty much that, that noise, right? So, as you can imagine, you know, just being able to take many issues and correlate it into a.19:10Single cohesive issue really cleans up things a bit, right? So this toggle is pretty valuable here for our customers there. Alright, so moving on, this slide is a little busy, but it's just showing different widgets that we have or dashboards, let's just say, and we'll go through each one of these.19:30Some of these I'll have a single slide for, but here's the AP performance advisories that we have and a lot of what that shows is identifying any poor clients that any clients that's having poor experience inside the AP, you can take a look at it here. You can leverage AI to analyze within that four week.19:50Time span and take a group of APs and and and see, you know, what is correlating to those issues there too as well. And then you can easily get into a a root cause for that there too as well. Beyond that, you can do that by groups of radios too as well. So you have that information. We have.20:11Trend deviations for different locations or different buildings that have APs, so each one of these circles here I'll talk a little more about here, but it's gonna pretty much take a four week span there to a four week interval to see how the trend is as is airing or what type of systematic you know deviations that that that's happening. It's like.20:30Network there. And then overall you'll get to get into the root cause analysis that we'll talk a little more about the who, what, when, where, and how there too as well.20:50Jared, about the trend deviation. It is that for those of you guys that are into monitoring your network and trying to see, you know, what's the use of my, of my areas and do I need extra, extra APs, e.g..21:00This is a good tool because if you're in higher ed or if you're in building management, you might see that this big lecture hall is supposed to be full of students and it's empty. And all of a sudden an area that was offices.21:21Is now full of people, so the trend deviation can also be used to that extent. I just wanted to point it out cause you know my my bad sense went off when you said here's the utilization when, of, of the AP, so.21:39Back to you. Oh yeah, NO NO worries, and I I forget to to tell everybody May has, you know, a couple of decades into higher and wireless experience there too cause I have a decade in it, but, you know, she's my superior in that there too as well but yeah thank you. Appreciate that May. We'll jumping in on that.21:55Okay, and so let's jump in here too as well for a network reasoner and how that works, right? So, so we have network reasoner here and it's like, you know, workflow engine.22:15And, what we can use it for is we can create and build, you know, expert systems there and it builds a whole knowledge base that we can then inference into root causes that we have there and just get remedies of identifications of issues that that we're able to see here. So, that knowledge base.22:27This is pretty much fully, you know, your the best practices that we have validating inside this design. So just like when I mentioned, you know, may have in you know two decades of experience in in this. This is gonna be a knowledge base of.22:4740 years of experience there too as well so we can pretty much resolve issues a lot better, right? So, as we, you know, the more knowledge we have, the more, the faster we're able to conclude some of these issues here too as well, and we can find those consistencies and complex and details, right? So, that.23:05That's what the network reasoner is doing you can run this within let's just say subscription issues or assurance issues, you know, anything that within your network is having issues with that's why I touched on a lot of the health and three 60 degrees again because that's gonna be that.23:24Knowledge base there that got listen to learn from you.23:44Right, and so telemetry like right so let's just pick an issue, right? So telemetry status, right? So if we see a telemetry status showing that assurance telemetry status is poor, there's a.23:48This is where you can actually see it inside the the console, but it's also when we see it inside the controller, we can take a look at the device and get some couple show commands that we can make sure that the telemetry is active here and the way and the way that looks there, and the way that looks there is that we can.24:06Have it, we can run our machine reasoning. Let's just say, e.g., we'll run our machine reasoning engine to troubleshoot telemetry alright and the way we do that is go to tools and then network reasoning as you saw, we can launch the assurance telemetry.24:25Analysis so we have a a bunch of cards here. So our assurance telemetry analysis or CPU utilization, we're gonna do this one here, right? So this is low, when this is low. Once we see that we can take a look, we can let it run, right? And it says run it again because it's ran. So you can run it at you know multiple times to get a good look of.24:39A full knowledge base of your network, you can check the device activity issues as you see here checking telemetry subscriptions, right? So that's what we're pretty much running through and you can get details in there too as well from the conclusion process and let it troubleshoot for you and it it'll running.24:59Perform those processes based off the best practices for it. And you're just running this and seeing what are some of the relevant, you know, issues or activities that you need to drill down into.25:19And so let's just say like it is an issue with the telemetry right here. It just shows that, you know, you can run a couple of show commands, so we run the show command for telemetry connection, which is telemetry connection all. What we're looking at is this.25:31Pain here, right? And making sure we check the state of the telemetry. If it shows active, then we're all good. If it shows anything other than that, it's gonna be one or two things which is the certs that's not downloading to the devices or there's a firewall issue.25:47Alright so we can find out what, you know, firewall holes are not functioning for us there, you know, netconf, it could be, but if you see, you know, N A the telemetry config is missing, right? So that's just telling you what that process is. You can then.26:02You can then, you know, run a shell telemetry subscription summary. You can take a look at the data that is showing to see if the telemetry is is running if it's config configured. And if you see some of that data there shown there, right? So this is just from the address to the source address there for it, so that's the device.26:18To the catalyst center there as well.26:38Okay, and then if it's missing, let's say the telemetry is missing, what we can do is we can resolve the issue by going to the specific device. We can go to actions and telemetry and we can update to the telemetry sessions, right? So that's pretty much just selecting a checkbox there.26:43Making sure that all the other configurations that we need are are there too as well. But then, you know, going through to, you know, through this workflow here to actually force the telemetry into the into the device there too. You can do it within the lab to test it out, and once you, you know, you kick your tires on it, you make sure that it.27:05It does solve the issue that then you can.27:25Check this box and run through that workflow and of course the telemetry there and, you know, wait await a few to make sure that that data is live and and showing you showing you everything that you need to. Here's the suite of our, you know, root cause reasoners that we have here. We have a few of.27:29Right, so, you know, I'm trying to see if I see the telemetry there, but we have the client counts, we have the CPU utilization, we've had Lisp pub sub, we have ping devices we gotta hold.27:49Slew of root causes that we can run through the regional and get some information at you as well, right? The power supply failure layer two loops, this is a big one too as well. So this is all you know machine learning and AI processes that we're able to use for each one of these network cards here.28:01Here let me see. Okay.28:20All right. So now we're done with the historical or let's just say live there data there I say. We get to baselines where we actually learn in our network from here. That's our quote unquote manual way of seeing some of our data that's been correlated for us, so I'll say manual, but it's.28:24No, it's correlated for us with disappointing and clicking here, right? But now we get into the behavioral.28:44Processes and what we, what we get from AI working force here, right? So, or learning learning our network here. So with this toggle here, this is gonna be within the issues and events settings, so you'll have the issues and events and all the regular issues and events, but you'll have a toggle here that shows AI.28:53High driven.29:11Issues and you can toggle that and we had a screenshot on it of making sure that we see the issue type that we have here. And so again it's as dynamic now, right? So it's pretty much allowing you to see all the relevant anomalies as opposed to all the anomalies inside the US issues inside the issues and insights.29:12Okay, and again, so as I was talking about, you know, learning your network, the the green blob here, I'm gonna show a couple of slides on here as you can see, the green blob here is pretty much the personality of your network, if you will. And then.29:32The blue line is going to be, you know, the data that runs and the data that runs through that, you know, that personality trade of your network. Once it deviates, it's going to flag that time, right? It's gonna flag that time based off a lot of information here. So we can see our problem detail is high.29:51High percentage of onboarding failures, right? So.30:11We can correlate that and add different KPIs here too as well, but I'm just want to show you, you know, it flags right at or right when it actually runs out of that deviation there, ok? So we have that, and then we can drill down into what is this impacting?30:15Alright, what's the root cause of this? What's the suggested action of this too as well? We can go, we'll go through it too as well, but I just want to point out that it it flags it right off the back for it for you. And you can add those KPIs and we'll see how that looks, right? When it doesn't deviate, we see that it's normal, there's NO flag or anything.30:32Two flags there, you get to see a lot of the issues that are syndicating into red of that deviate of that deviation, right? So the green box that.30:52Percentage and you can see the statistical models that that we have conditioned there for it too as well. Okay? Here's the who, what, when, where and how, right? So like I was mentioning just for being able to have more than one KPI is, is good enough, but.31:01Well, for this type of issue, let's say triple aaa failures, we get to see how many clients are impacted there. So how many people can't log in? What's the issue there? Is it, you know, you know, just telling you where.31:21The issue lies if there's multiple clients having this issue. There's one client that's having an issue, then that points you in another direction. Now hopefully it's not impacting different multiple sites now, right? So this one here is a single site here and then we can get to who what when where how that too is.31:36Well and we can correlate those KPIs and at different time.31:56Timeframes of where that happened. If it was busy at a certain time frame, like May was talking about certain location or you know sort of campus, then this is where we can drill down into that location there and see what was going on from that percentage standpoint there too as well.32:01And we got more, right? So here are some of the KPIs that we have, so our onboarding times not just AAA authentications or anything, we have association failures, right? So it shows you why the client or when the client was disassociated to the network and why, how you can find that out. Some roaming failures, which is a big one.32:20Too as well. So we're flagging that across the board.32:40Right, so that deviations won't show up anywhere in this once the, the AI is turned on there too as well. So this but this is we're still at baseline, right? So this is all the baseline information that it's gotten for your network and correlated for you. Here's some other impacted right so imp impacted clients, we can see some.32:43Where the device breakouts that we have there too as well. We can see that it correlates to the same time. So we're drilled down into there and then also the applications that affected and what type of traffic TXRX that is.33:03Correlating into that issue time frame too as well to do that. More of the baseline here, you can kind of like drilling out to get a bird's eye view of your buildings that you have. The the clients that are that are entering those buildings we get to see.33:19Right, so each bubble is identified as a building, and the bigger the building is the more clients have been in there. So right. So, so.33:39If you see like a bigger bubble than another one in one time, that means there were more clients in there and you were having issues with different volume inside inside that building, right? So like you see amber one will see calls to pause, but then you have.33:49A red one here around this specific time and this is where we need to drill down and see what that issue is there too as well. But you can see the value in it, right? So like being able to drawback, you know, you can go to different time frames, different time, different sites, different KPIs that you have here like DHCP time.34:05Times, association failures, onboarding each one you're able to do this and get a good baseline for it, right? But the size of the circle again is the endpoints that they have in that building issue as well, and you can see the key here for one of screenshot you can grab that now.34:25Yeah, drilling down into trends and insights, I know we talked about issues and events. We're gonna talk more about the trends and insights here and yeah, and I wanted to have a slide on this here specifically so we can see it better, right? And it's just.34:42Just pretty much showing the AP performance advisories, right? So.35:02High co channel interference that we have. We have high AP density that we can see you know based off the radios and what which one of those are impacted, the high client activity too as well so we can take a look at that, right? So we can leverage this AI, like I said for four weeks that we saw there to be able to get that information there.35:06Well that's pretty redundant, but I wanted to get a good look of how this looks here and we can get those trends trend deviations too as well. So that's the AI working force as well, ok? We can do a bar radio. So digging into low AP density e.g. where we.35:26I have this chart here, we read this chart by looking at the blue chart versus the pink chart, right? So, this is gonna be showing us the blue chart is referencing radios in your network with NO, with NO known client, right? And the experience.35:45Some problems that it may have that compares to the pink chart shows and showing information for APs that are having performance issues, right? So we can scroll over to some of the blue radios we can see what those radios look like in the network. We see the radio.36:05Is that has, you know.36:25Certain DB levels, right? So or sorry DM DBM levels, sorry. Well, for 40 % of the time here we can see a look at that, right? So we see the client experience KPIs and we can take a look at the APs that that's running with that issue as well. So we see all of that insight with.36:27In that detail to any comparisons, right? So we're comparing our devices, we can compare our location, but we also think.36:47Compare, which I'll talk a little more bit about our peers too as well. All right? Some more trends and insights for the AP performance, right? You know, this is pretty much showing the performance advisories and then the trend deviations that we can go to, and i'll spot that for it.36:55To take a look at more of the trends in it. So, so we're in the trend deviation here. This is showing more of significant data that we have there for not just the performance but just the deviation within that weekly time frame. We can see some of the occurrences if we click the link and get more information on that, you know, get into.37:13Through the three 60 degree view from there too as well. And here's our KPIs that we have here. So like client count for a specific AP radio throughput that we have for that AP too as well and what channel frequency that we're running into as well. So it's.37:32Cool lading for you for that too as well. And we had this in our, you know, our full slide there. This is the trend deviation over time. We get the beastwarm here, right? So within a four week time frame, we're able to.37:49Let this run and take a look at, you know, where and different APs that we have and different type of clients that that adds into it. Like for this one here, from a weekly client standpoint it has 36, right? So, 5.8 % of.38:04APs or or clients.38:23Where it hasn't had an issue or if it had has an issue. Like if it had, if it's in red, then we know we need to drill down into that particular time and that particular period that week. So that'll give us a good view of this over time here, right? So week one, week two, week three, week four, how are we growing, right?38:25As we grow, it's our.38:46Services degrading or if it's getting better, right? So this is how we can take a look at that systematically. Okay, and then we have heat maps, and so this is not actually AI, but one of the most powerful pressing units on our, in our planet is the.38:49And I right and that's why that's why we use heat maps, right? So what we can do is we can display the data.39:08And correlate it into patterns, right? So, we can recommend, you know, if you, well, we recommend that you have the heat maps set up so that you can take a look at each one of these KPIs here, right? So each one of these.39:15Value adds here for access points or for like we're showing access points here where we can show switches and we can show whatever type of KPIs that we have for that. So with these access points we have, you know, interference, we have failed inbound.39:33Processes and data and all that stuff, we can take a look at packet failure rates, and we can look at the interference there from it from a heat standpoint, right? So whatever's colored green that we could see the the type of radios that we have there for and we can filter all the radios here or we can filter.39:49By just 2.4 ghz. I would like to do it there too as well.40:09May you raising your hand? I did, you know that when I see 70 % utilization guys reminder for the non wireless people in the room, 70 % utilization means that your channel is pretty much shot, right? You don't have much room to keep advertising your SSID.40:13Because you don't have that air time, so if you see that real dark red.40:33Keep an eye on this thing. Yeah. Absolutely. Thank you. Appreciate that. Yeah, I mean then that's you know that's where the human eye really comes in, you know, that's why we wanted to color code it so that we can.40:39At a glance see this information here, right? And then we can change the KPI and see some of the channel utilization here and see every color code relative to that there too as well. Okay, like you said, 70 % over 70 % yeah we're, we're looking at too much.40:53Okay, and now we get into the peer comparison, which is one of my favorites, I call it spine owner competition by early you just get a good view of different people in your industry that you're able to do. Well 1st what you can do is you can compare buildings and within your network and within your environment there.41:11And, and get some KPIs running here. So our radio throughput, you know, what is our radio throughput?41:31Between building A and building B and San jose, right? So we can compare from there, we can get some of the, we can see what's improved or what's good in one site and we can kind of copy and paste that on another site to see if that helps our radio problems there too as well, right? So we.41:37We can take a look at that and so drill down into either one of these, right? So if one site is having client SNR, then we can take a look at that and kind of compare it to a good site and run through there. Another thing I didn't mention too as well, some of these slides you're seeing this little thumbs up here.41:57And really what that is, is this.42:17I will say like the manual way or let's just say the yeah the manual way on learning your network through tools. It's kind of learning your personality here. So like if you know what you whoever's logged in here, we can tell Catalyst and that hey this information is helpful for me. And so whenever you go.42:20Go to this dashboard, you come back to this dashboard, it's gonna show you.42:40More of the same, right? So if this if this two of the buildings inside San Jose that she said that this is helpful every time you come here, this is what it'll start with to give you that that processes there. So if you just, if you're just managing these buildings, you can teach catalyst center to.42:44Just show you what you want to see here, right? So, as opposed to just having to filter this every time that you as well. So it learns what's helpful to you too as well. So that's a, that's a way to do that. It's some bunch of these dashboards that we have there too as well. So keep a lookout for that if you haven't already used it, and here's.43:03The, you know.43:23Inspiring on the competition kind of deal, right? So that you can compare just like you compare builders, you can compare different people in your industry there too as well. So if you in the education higher learning will stay with amaze example if you're in the higher learning process, you can take a look at what others are doing in the higher learning.43:24Personas there too as well. And then you still have your KPIs from in that regard too as well. So packet capture rates, you know, see how you can each stack and compare to other peers in your network in your network or let's say in your industry, your radio throughputs there too as well. All KPIs are there for you from.43:45You know, every day of the week, right? So you can take a look at, you know, everything relative to that there too as well.44:04And so yeah ok alright so we're gonna talk about event analytics. Okay, so this is the last bit of me and then i'll turn it over to May to get into some AIRM. We saw that widget there, that full long widget inside our, our help.44:12Dashboard which which is our event analytics here and what's good about our events analytics is that it shows the volume of messages by categories and severity.44:33Right, and again we we get the color code so that we can take a look at, you know, what's, you know, what's important for us there and we can easily, you know, spot the the categories that we want to select here, right? So, we categorize the messages and.44:43Technically not AI too as well, but it's still helpful in using pretty much machine learning and enable to compare these issues and get a preview of your sites and syslog messages that we have here too as well because this this is wired events here that we see here and we can take a look.45:02That four long, you know, I call it whenever we have a page like this, I'll call it three 60 degree review of it too as well because we have four, we had three of them. So I just call it the three 60 degree view of syslog messages. We can.45:21Take a look at a three 60 degree view of any type of site, right? So we see globally here and we seeing everything relative to that global process, but we can take a look at our San Jose site and and just drill down into that and see all the highest.45:34Severity events that we had there. Some of the rares events that we haven't seen before, and some of the the volume events that keep popping up, right? So we can solve for that too as well, right? So and this is from wired and wireless too as well. So talk about baseline, we actually get like a full holistic color coded view of the severity.45:50The Messages that we get within our syslogs there too as well so we can resolve for that. Okay, and more of the, you know, a sense that we have more of the same that we have here with the color codes that we have here, this is just us clicking into one of these here and.46:09Drilling down, getting a closer look of some of the issues that we have there, but then drill down to the events that we have too as well. So getting closer to some of those informations that we have, what are the coverage holes that we have in there from a radio standpoint, what are some of the reachability that we can we can take a look at?46:30Either can see.46:50From the AP, from the site to the device processes here right here too as well. It's a little blurry. I do apologize for that screenshot there, but you get to see this within the drill down of these colored colors that we see here, and this will show they're red as being failed you know the.46:52The assessive type of data that we're getting here. So it's getting into.47:12Getting to red there too as well. So that's, you know, from a threshold standpoint and then some of the green ones that we have good and we can compare from that there too as well. So that's how we can see that right? And so here's the drill down dashboard that we can see here. And this is just us going into the dash.47:17For of the health and site analytics and then we can see our onboard attempts from different sites. We can see, you know, some of our failure attempts, some of our connection speeds, and we can drill down to that too as well. So the onboard attempts is gonna show all of the failed and show all the green I just use the same.47:37Key that we show for a different type of locations so we have an example of Cisco Live or the onboarding attempts probably was me and there too as well, but we can see some of the KPI values there for it and this is the key for for drilling down into any one of those, right? We'll see some of the durations.47:57Of of the Romans we can see some of the attempts that we have there in authentication failures and and the impact of different buildings that we have there too as well. So it's just like the.48:17Same dashboard that we have there for help, we can filter bands and ssids that we see there. We'll use this for being able to know you know what type of analytics that we need to focus on and from a site by site basis there for it too as well. I think this go by, go back 30.48:29Days two as well, right. So, so that's always good just for our clients there. But of course you you always have the, you know, seven days of course there for it, 24 3 h, seven days, but it should be able to go back 30 days within a map standpoint there too as well.48:48For calendar standpoint. And so more of site analytics would be not only just the duration that we have in the coverages, we take a look at some of the SLAs per site for our while it's there too as well to make sure that we're.49:05Graduated for there, and we can take a look at the proactive identification sites, investigate from there too as well. So this is more so of just showing you the coverage performance for each building that we have there too as well. So we're waited to see that.49:24Right? And I that is it talking to if I. I'm sorry another tidbit from the field. So some of my customers, they tend to use this when it comes to making things better for their.49:40For their infrastructure. They have an overhaul project, they're changing the coverage, they're changing and it can be something as simple as we're changing our ICE servers or radius servers or we have a new DHCP. So deviation and, and.50:00And, and site analytics and how everything is behaving is going to give you insights into how.50:20This is all turning out to be and is your investment in, in the right, in the right place, right? Is there something that I should suspect before my customer calls me and tells me there's something wrong? So I'm sorry. I had to, I had to get in there.50:28Jump in and talking about this bucket. Yeah NO and and just to your point, right? So this example right so we can see we're seeing here our onboarding attempts for our building 18, right? So we can take a look at building 18 and what what's the worst performance and, you know, inside one side any.50:46Floor let's just say four four was the worst performance floor within that building, we could take a look at it from like so we see the site comparison in the data, but then we can drill down into different floors which which type of floor has the the worst.51:05You know performance and then and and just drill down in and get granular from that perspective there too as well for either one of these right so that's how you can actually use that there for it, right? So you can see the floors, different floors, you can go into different sites there. So you get granular already quickly at the site there. And.51:22I am actually going to this is more so the slide here for the failures there when you have done it, but this is where you can see some of the real details of, the view details of the attempts there that we have and we can drill down into that process there too.51:40As well. All right, so, and then this is just pretty much showing you the the speed coverage that we have there for and the KPIs that we have that we can edit it based off the percentages that we need for our KPIs and setting up for our our SLA.51:58Persona there too as well so we can tailor it to these thresholds that we have here and these are configurable too as well, right? So these are the thresholds for our KPIs based off our SLA that we can use for it, right? And this is just our site analytics settings where you'll find that there is assurance and settings and it's pretty much the assurance.52:18The bottom of 2nd half menu of assurance there too as well. So you can set up any type of threshold for whatever level that you want to as well to develop the the right.52:38Noise that you want to be able to see site catalog center and to correlate this information there too as well. Okay, and you just pretty much can type in what's good, what's not, and it'll it'll look at or show you whatever you have inside of.52:51Onboarding standpoint or yeah or or deviation standpoint or disassociation standpoint there. Even one of these settings are gonna be defaulted, but then you can change it to whatever you want here. It's defaulted to 10 s here for it but you can change it to whatever you want, right? So it's configurable, completely configurable.53:08And once you have it configured, so now you are able to aggregate those KPIs into the site analytics site here, this is where you'll get to look at the percentage based off your configurable KPIs that you have there and your SLA setup for it.53:31And it'll be on the overall page and the site analytics for it.53:48Alright, and you can drill down into it, and you'll see for what's worse performing of the child buildings here. So it'll show you different sites, but then it'll show you buildings within those sites there too as well. So you can see the percentages of onboardings and connection speeds enrollments and coverages there too as well.53:54Okay, and I am going to let May take us home you know with ARRM and we'll get a demo here in this half of the presentation and thanks guys for, you know, being here and just gonna jump off mute off here for you.54:14I might. Thank you. So hello everyone. So thank you Darred. This is very insightful these these things that we've, that we've been showing you here are, as I said, very valuable when time comes.54:31To actually, to actually troubleshoot and assess where you stand in terms of, in terms of your.54:50Am I taking? Yeah, I am sharing my screen. Excellent. All right, sorry for the little snaffu there. All right, so all these metrics are super insightful when it comes to troubleshooting to understand what part is not working because they're.55:01There's nothing that looks more like a wireless problem than any problem in the world. So, and I'm saying wireless, but it could be wired.55:21It's just that it tends to be oftentimes the the wireless person, that is the culprit so as I said yes, my name is May and it's my fault. Alright, so AI enhanced RRM.55:31Let's get into it. So you're gonna find this this use case in assurance in the AI enhanced RM section, ok? And the goal of AI enhanced RM is basically to move on from, and I was, I was.55:46Doing some archaeological searches when I got prepared for this. So we want to move on for those that go way, way back to the early two 2000s, I was I was doing archaeological searches and I was sharing that with Jim Flowerwick, our.56:04Magnificent distinguished engineer, if you get a chance to see him talk, hear him talk on RF, he's awesome, so don't miss that. That being said, I remember assigning static power and channel to every AP that I had back on.56:23In the days of the university campus, and let me tell you that was long. So initially we we went from that.56:43To having APs kind of have some leeway but make their own decisions. And then we went to WDS and then RF groups and we evolved this whole thing into the traditional RM that you know now. So how does that work? Well.56:51RM is basing its decision on information that it gets from your APs, right? So it can hear information from the neighboring APs in the beacons, but there's also NDP that's a Cisco protocol that is, that is actually being.57:11Being exchanged the the the packets are, are being exchanged at the highest power, lowest trait supported on your AP. The goal of the NDP packet basically is to actually reach as far as it can, ok? To make sure that it gets a good picture.57:31Of everything that's around the AP, right? So that everybody can bring that information back to the controller and the controller is gonna.57:51Create some form of a plan that's based on the last ten or fifte ten to 15 min, most of the time it's ten, so it's based on the ten last minutes of of information that it got from the APs. So initially when the RM that you guys.58:02Now and and years every day was invented. A lot of the people thought that it was really nifty to actually have a calculation done.58:21Outside the business hours, right? So they were doing a calculation of the, of the RM and have RM recalculate and recalibrate the network at 03:00 A.M. in in the morning. The problem with that is that at 03:00 A.M. in the morning.58:32You don't have all your users? Well, unless you're a club, right? But, you know, most businesses don't have all their users, all the sources of interference, nobody's going about their business. So it's really quiet on the network. So you're adjusting for that kind of weather. And then the army of corporate america.58:50Comes to the office and your plan is not always super well adjusted to to face that. So to make that better, we've evolved our our way of going about RM and eventually we came up with the idea that maybe, you know.59:10We could evolve this, this RM into, something that was kind of more conscious of the busy hours, you know, and that could be maybe dependent on, the, the building you're in.59:30Maybe, you know, it would be cool to have more visibility because unless you're a CLI fan, it's hard to understand why, why a decision is made on the actual RRM the way it's implemented on the controller.59:47So you've got limited visibility, you've got multiple panes to go through when you want to configure RM and you don't have any simulation modes, so when you're doing a change, it's a, either you implement it or you don't. So thinking of all the.01:00:02The little perks that we wanted to have in RM, we said, ok, that's gonna need.01:00:22Compute power, we need a centralized place. What can we do? So what happened is they brought AIRRM in catalyst center. So being it in Calis Center, it gives you the torque that you need. So you've got plenty of.01:00:28Of processing power, right? Catalyst center has lots of that. You've got visibility and well, you know, if you wanted to send the information of the controller to the cloud, say.01:00:47To send that to, AI in the cloud in AI instance in the cloud, you don't want to really expose your controllers to the internet, right? But catalyst century is already exposed to the internet most of the time. So.01:01:02AIRM and caller center, you get the visibility, you get, you get all the advantages of the processing power, and then call a center can actually monitor the connection to the controller.01:01:20And it can monitor the connection to the cloud and act as the broker, right? So this is how AI RM works. So basically, what happens is caller center becomes the RM leader of your deployment.01:01:36Right? So what's the difference between AI RM and a traditional RM? Well, as I told you, right, you have this idea in traditional RM of the snapshot. You take the last 10 min and you adjust to that.01:01:57And depending on whether or not you're tolerant to change, it's changed regardless of whether you're busy, you don't necessarily know if you're, if your change is happening at the exact right time for you because I mean, even if you wanted to take the weather report every single time.01:02:16RM does its thing, well, you know, it's really a manual process to to configure.01:02:36You know, it's a pain to actually visualize everything and, there's NO like the the the metrics are very much static. You're not necessarily taking into account a feedback loop, and there's NO emulation, so.01:02:44The the AI enhanced RRM is going to base itself on the trends, so you're getting historical data, you get 14 days of that, and then the I'll show you later, it's the.01:03:04A breeze to configure, you get segmentation, your, your RF groups, it's easier to actually segment the RF profiles per building, and then you got your, your RM control center, you can run simulations, you can.01:03:18Actually get insights into your your network. So after a while, if you've defined your business hours, your busy hours as say eight to five, but really your busy hours are more like seven to four.01:03:38Insights is going to raise a flag to you and say, hey, maybe your busy hours should be moved because we see that you've got a lot of people that are coming in from seven to four or from nine to six, right? So you can adapt to that because the AI is accumulating the data and, and, and.01:03:55Bringing it to our algorithm in Catalos Center and in our cloud so that.01:04:15Everything bubbles up to those insights that provide you visibility. At least it's raising you a bigger flag than having to dig into the CLI to actually understand what's going on, right? Plus, if ever you want to bring a change, the simulator, it's.01:04:20If the click of a button honestly and you can define a change and see how many APs would be impacted by that change. I'll show you later. All right? So of course, everybody that sends data to the cloud is preoccupied by what what data is being sent.01:04:39It's important to note that everything that is being sent to our AI cloud is both encrypted and anonymized. So Jared was talking about how we can do comparisons. Of course, we're not comparing, I'm gonna take examples that, that.01:04:59That I know on top of my head, you're not comparing Princeton with Harvard, right? And Princeton is not saying I want to compare myself to Harvard. It's comparing yourself to a sample of.01:05:18The data that we have accumulated for higher education, right? It's not, it doesn't have your name on it. It's anonymized, everything is encrypted, it's stored, encrypted in the cloud. So.01:05:34Bear in mind, everything is secured. If ever your infosec team has questions on, what is being exchanged, how it's being kept, the protocols, everything, there is this link here in my presentation, but I'll show you I'll show you also when I get to the.01:05:52To the demo in your settings and all the AI options that you can actually define on your system. At the very bottom you've got some very important information that you, that is shown there and this information is also includes a link of.01:06:11Of the this this the same data sheet privacy data sheet, so you can provide that to your infosec team. And of course, you know, if ever this is not clear, they can always ask for a consult with us, we can demonstrate and.01:06:31You know, provide you with more, more, even more data, but for now, what has to be said is these are the things that we, use to actually crunch this.01:06:49To, to, to actually crunch this information and match and distill a, a right RM plan for your, for your network. So we're going to take all the L2L1L2 stats, the channels powers, channel width.01:07:05Right? I I didn't want to say bandwidth because I I thought it was the slide says bandwidth, but what we really mean is, how large of a band you're using and not the download speed. You might want also L two and L three information on the radios.01:07:23You know, information on your wlcs and everything in there is anonymized to to to the maximum possible. So, you know, we're taking decisions.01:07:43Efficiently, we're anonymizing everything and streamlining everything to make the process super efficient, right? So of course, what are the benefits? Well, it's it's not.01:07:55It's not just a question of passing the pack to, to catalyst center to gain visibility. You can run AI enhanced RRM both on, on boarded and provisioned.01:08:13Catalyst centers, 98 hundreds, sorry, slip at the top. So you can run that on any of those 98 hundreds. You, you don't have to have them, as I said, provisioned on your catalyst center. You can be using only assurance and still use that AI RM.01:08:28When you come to push.01:08:49The the required commands on your controller, the only thing that is going to be pushed is those commands required by callus center to be the master of, of the RM group, right? Because Callis Center is becoming the, the leader.01:08:51Of of the of the RM group. It's that we have a super simple workflow. I'm gonna show you both, right? You, if you are provisioned, if you're not provisioned and you're, you're actually onboarded, but the workflow works for both and it's really streamlined, it's really.01:09:10Clear it's super easy, so you don't need to jump from like point to point and go from like network to provision to this and that and 3rd.01:09:30Very simple. I'll show you. And then, you know, the power of, of, of, AIRM is that we have seen tangible enhancements to both the wireless coverage, how it's behaving, the performance. We we saw some.01:09:41Specific in gains in SNR in our own events because AIRM is something that we run in our own events. I'll show you some graphs later on, of, things that have been that have been reported using RM. But when you're talking about.01:10:01An advantage of three to six dB, for those that are not aware, this is like 1550 to a hundred percent improvement. So it's it's it's awesome, for, for, for any wireless admin there to to consider.01:10:20For that, and by the way, there are protections.01:10:40When you're enabling AI enhanced RRM so that you're not left stranded if our caller center has an outage or you lose connectivity to the internet. So I'll explain that all of that later. All right? So basically how does that go to enable? Well, what you're going to do.01:10:43Is you're going to have to, to enable the AI cloud analytics.01:11:03And once you're done with, with that, you have to also enable the AI enhanced RRM. So this is in your settings in in caller center and then you go to the workflow. So the workflow, tells you configure AI and enhanced RRM. You can enable.01:11:08With or without device provisioning. And so, you are going to be brought to a window that will ask you to define an AI RF profile that can be by building. Now there is one limitation. It is that if a controller is.01:11:28Also it is actually servicing multiple buildings. You can't go.01:11:48And enable AI RM for this controller only on this building. There is one RM process on the controller, so if you're enabling AI RM, it has to be for the entire controller.01:11:55So all the content. You can have different RF profiles, AI RF profiles for each areas, but you have to enable it for an entire controller, ok? That's, that's the only little detail.01:12:11And so defining this AI RF profile, here's an example. So, as you can see, it's everything in one pane. So you've got here, you've got which channels? Well, of course you're gonna give it a name, right?01:12:26And then you've got the, the, the frequencies that you're going to use. You're going to define those busy hours. So the busy hours are when you're more sensitive to changes, ok? So, if you're, if you're very, very sensitive to changes.01:12:45Between the hours of eight to five, then you want to define your, your, your busy hours as such. So what it's effectively telling telling the the algorithm in a way is, hey, you know, if you don't have to really change something, it's not, it's not.01:13:04Super important, but it would just make it slightly better than maybe don't do it during the the the busy hours, and if it's like something that's super major impactful, then, yeah, you, you might consider it. So depending on your busy hours sensibilities.01:13:24You know, that this this is taken accordingly, right? And then you've got all your RF settings, right? Your flexible radios if you're running those, your DCA, your dynamic bandwidth selection, so whether you want to consider doing only 20 mhz channels or.01:13:42Are you ready to do 20 or 40 or 80 e.g. in 5 ghz, so that can be also configured.01:14:02And we got TPC, so all of these things, they're usually in different tabs on the controller, right? So here they're all on the same tab, and then we've got the advanced tuning. So, e.g., in 5 ghz, maybe I want to enable preamble poundraing, I want to.01:14:11Due to zero in DFS, maybe I want to change my channel plan and use a wider or.01:14:31Or a narrower channels, so and by the way, best practice has has changed in 5 ghz. If you go to the 9800, best practices guide, you're going to realize that what we used to say is that 20 mhz was the way to go and I would still abide by that choice, but if you.01:14:38Can't afford 40, you can consider 40 safely. So that's especially true in North America with a 1200 1200 mhz that of extra that we got with 6 ghz, right?01:14:58Alright, so, the business hour, the busy hours I mentioned are the business critical, hours of your network. So, you know, do we want to adapt, adopt adopt our.01:15:12What channel plan to during those busy hours.01:15:31Yeah, that that note there is, is super important. So the busy hours are based on the local time zone set per site, right? So, bear that in mind just in case, you know, you see a behavior that you're not expecting and you're don't understand why it's doing something every.01:15:36Every day, 1 h late or 6 h late, take a look at the the local time zone, right? So that you don't have any surprises. Alright, and by the way.01:15:55Necessarily. So, and how this goes, it's very simple. It's still based on the AP through NDP is detecting information. It brings it back to the controller.01:16:16It feeds the RF data to catalyst center. Catalyst Center is anonymizing and crunching all this data, sending it up to the cloud. And then, the cloud is going to do its thing, calculate and spit back a an R an RM channel plan, right?01:16:31And so it's going to feed it back to catalyst center. Catalyst Center is going to say, ok, here's, here's how we're gonna do this and sends the decision to to your controllers. And then the controllers are populating the well, you know, they're assigning the new channels and new pan.01:16:51Hours to to the infrastructure.01:17:11Alright, so, yeah, this is an example of how you can enable this for a provisioned provisioned environment. So you can go ahead and go into system settings, you enable your AIRM, then you go into the.01:17:15Wire the deploy network profiles and you go to the wireless network settings and then you go to the wireless tab and then you're going to see that you can add.01:17:34An RF profile but you can also add an AI RF profile. So if you were provisioned, you define that profile and then you can apply it to a location, ok? So that's one way of going about it. Now.01:17:50The, the workflow is a little bit more visual for those that are more visual learners, and so you'll be able to better understand what's happening if you go through the workflow. Alright, so.01:18:07Going through the workflow, as I showed you, you know, you can, you can select your, your AI provisioning to your taste. And by the way, if ever you, for whatever reason, you're not sure.01:18:24How to, how to go about it, get lost in the interface or you're not sure about how to go about things. You can use the walk me that will, that will allow you to understand how things are actually laid out. So here are the steps.01:18:38Of enabling AI enhanced RRM, right? Alright, so allow me, I will bring my browser up.01:18:58And show you a little bit of what this looks like, right? Okay, so.01:19:20This is the cloud and just in case you don't know, we have a D cloud environment that's actually it only requires your your CCO ID so you can pre use it and take a look at what we've done here if you have, because it's an instant demo. So if it was to.01:19:35Actually enable my, my AI and his RM.01:19:55Or take a look at what is done in AIRM. I would go to assurance, AIRRM, and either I go into the workflow which I've talked to you about or once it's enabled I could come to AI enabled sites. So if I was.01:20:00Launching the workflow, I get to define a name. So as you can see, I can enable without the device provisioning or with device provisioning, ok? There's not much of a difference in in how things are done.01:20:20So there are a few requirements that are being asked of you. So of course you've got the green check mark that means that we're meeting those requirements. So, if you, if you want to take a look at those those options, you could always go to system settings, ok?01:20:37So, my devices have to be, to be capable of being.01:20:57Managed by AI enhanced RM. Most of the recent, APs especially the 91 hundreds do very well with that. All right, so I'm meeting that condition and yeah, I need device controllability to be enabled.01:21:05So, and then I could go to system settings if I had not actually enabled that. So what it's going to do and it's going to ask me to pick the locations where I want to actually enable.01:21:23ARM, so I will pick a location such as Bangalore and I will need to, I will need to sign a wireless controller. Of course we're pretending here so.01:21:39One is going to be enough, and so this will crunch it and it I will have to define my AIRF profile, ok? Because this is my lab, not gonna take any chances. I'm going to go to something where I have.01:21:57Full control.01:22:17Oh I'm going to assign primary controller to this guy.01:22:25So it's giving you a warning. Hey, you know, this controller is also managing something else. So bear in mind that there's only one RM program processed by controller, so don't forget to, right? Don't forget to actually remember to take care of the all these other guys, right? So.01:22:41So I get to define an RF profile. If I don't have any RF profiles that are actually created, I've got this link up here to create this AI RF profile. I can fill that. I have all my bets that I want to enable busy hours if I want to, I.01:23:01And define them, the bands for all these settings, alright, I get also in the advanced, in the advanced menu, I get the supported rates, the breadth of which I want to, to use for, for assigning power.01:23:22And, if I want to have RXSOP and so on and so forth. So, if I save this RF profile, I'll be able to assign it here in my list. So I'm just going to pick one for the sake of the demo.01:23:41And so, everything is going to be populated here, as such, right? So I can get a work through and then clicking next, I will get my summary. It's going to tell me also what it's going to push.01:23:59So preview and deploy, I could download the backup of the current RF config, the RF settings, alright? And so by doing that, I I have to say I agree.01:24:18You're going to see a little bit of what it's going to push to the controller. By the way, when you're enabling this on your controller, if ever there is something that goes south with the watches that are being kept, so between the.01:24:38Controllers and catalyst center in between catalyst center and the cloud, there is a watch. If ever something goes wrong, there are, it's it's it's basically going to revert to regular RM, right? So.01:24:58So you're not, you're not stranded, right? However, if you assign AIRM to a controller and you do not define AI enhanced profile for.01:25:15For for other parts of the building, then you're defaulting to static build static channels. So you you really don't want to do that. If you're enabling AI RM, enable it as a whole for the entire controller, and make sure that you've got AIRF profiles.01:25:33Defined for all the APs that are on this controller, ok? It's important. So as you can see, basically what we're saying is, hey, your new leader is going to be catalyst center here, ok? So basically that's what this config is, is saying. So if I click deploy, what you will effective.01:25:50We see if I come back to my presentation, right.01:26:11So what you're going to effectively see is that the, the RF leader is now Calla Center, ok? So basically that's, that's what it's going to change. By the way, to support this.01:26:34You need to be running iOS 1793 on your wlcs catalyst center, 2374 patch two with the advantage license. And then of course, wave one, wave two APs, the catalyst series and Wi Fi.01:26:51Six Wi-Fi 66E and seven and of course our 9800 portfolio all support this. And by the way, AI enhanced RM is also supported on Meraki, if, if you are a Meraki user, alright? And if you need to go back, you.01:27:11You can always go back and revert to your traditional RF profile.01:27:30And go back to your AI RM, right? So you're going to unassign the locations and effectively you're back to, when you've added unassigned, you're back to your usual grouped leader, that is your 90.01:27:38Hundred. So this is just an example of a Cisco event where we enabled the Cisco RM, Cisco AI enhanced RM, and you can see that we enabled it on the 26th. It took about 6 h to cook.01:27:57Right to stabilize and do its thing. And then you can see that the number of APs with high ca high co channel interference went drastically down and that's even though the client number of clients went all the way up, right?01:28:15And as I was telling you, the SNR, we enabled it on the 26th. The SNR actually raised while the number of clients actually went up. So four dB is still.01:28:33A pretty significant improvement, so you you want, you do want to consider that. I'm going to go over AI and point analytics really quickly just to remind you that it is in Catalos Center, so it's a powerful solution that provides you with deep visibility into your network.01:28:49Identifies and profiles every connected device using multiple multiple ways of classifying, so it's it's aggregating data from from sources like Mbar and ICE and it's using trust scores to proactively identify and manage risks.01:29:12And.01:29:31And it's also mitigating the challenges that are brought forward with randomized Mac addresses, e.g.. The goal is to reduce unknown unknown devices, the number of unknown devices using AI and machine learning algorithms, ok? So, this is.01:29:33Is based on deep packet inspection among other things and the ice discovery and mechanisms.01:29:52So, I'm going to I'm gonna go real quick over this because we we still got a little bit of, of, of a way to go. All right? So basically the classification is based on trust scores.01:30:00So, profiling is always like that, so you define how how trustworthy, how confident you are in the identification of this of these nodes and and then allows you to.01:30:17Decide accordingly. This is the compatibility matrix for endpoint analytics. So, it does require you to run a, a, a a higher version of code.01:30:34And I want to take a minute, guys. It's it this AI assistant is super handy in Catalos Center and I want to tell if I want to talk to you about it just show you really quickly what this is. So starting at version 23711, we have an.01:30:50Enabled an AI assistant in a code that's available at 237-11-3316. Now by the way, three, two is now GA so, it's also available in there quick quick correction May, it's it's not available in 316 but it is available in three.two.whatever.01:31:10Thank you. Thank you for that clarification. Thank you. It's true. It's true. We were all, it's kind of funny cause I remember you and me being very disappointed that it was not available in 3316I apologize for that. So 32323711, the AI is.01:31:31Assistant is available and it's really, really handy to understand what's going on in your in your network. It will help you diagnose and resolve some issues. It can automate workflows, it can crunch loads of data into.01:31:51New insights and concise you know summaries so that you can understand better what's happening and it has two mode, it has a default mode and it has an in depth mode, ok? So to enable it, if you are running the the right version.01:32:10You have to, you have to actually use your Cisco.com cco ID to register Catalyst Cloud and then you you need to go into help and DNA sense. All right? So.01:32:30The AP, the AI assistant is going to be enabled and you'll see the little icon up here show up that is the icon to call the AI assistant and you can like you can use natural language to ask it to do some kind of troubleshooting, understand something better.01:32:48You know, show you how to enable a functionality in your catalyst center and that's super handy. It's very efficient, alright? Given the time, I will skip over the demo.01:33:08But know that it's there. It is very, very helpful. The screenshots that you're seeing are from the beta, but this is NO longer a beta version. So, please check it out. It's very, very handy, and I'm gonna conclude the CIS.01:33:23By leaving you with a few thoughts. So basically, we'd like you to to to try and implement those, those AI ops. They will make your lives better. You know, you get proactive issue management with MRE, you get, you get to apply baselines and.01:33:44Check out how your network is doing, you optimize your wireless performance with AIRM. You can understand the impact of various various gestures, various maintenances, and various actions on your infrastructure.01:34:04So this is absolutely a mess. Although catalyst center is already providing you with huge visibility, you get even more with those AI ops use cases, and if you are still looking for a little bit of help with enabling all these.01:34:18Use cases if you have the virtual assistant.01:34:38They, the virtual assistant can help you and if you'd like to get a few more, a few more questions in, and discuss the, all the wonderful AI ops use cases with one of our csss. We invite you to get in.01:34:41Touch with your CXM or your sales team and they'll link us up. It will be our pleasure to assist you alright?01:35:01So, just before I let you go, I wanted to remind you that there is a post webinar survey that's going to pop open when we conclude the session. We would appreciate your feedback, so tell us what you wanna hear about. Tell us how we did, and let us know how we.01:35:08And best serve you. We're here to help. So, thank you for your time today and I'll see you next time. Thank you.01:35:28