Past the Hype: Intelligence in Assurance

So Jonas and I are here to present like the assurance strategy of Cisco, right? And this is not the first time you're sharing it. We have shared it last year as well.

In this right forum, uh, what we are going to showcase is first, um, talk about, do like a quick recap of what we shared last time. Show a snapshot of the progress we have made so far, and then jump onto your demo to show all the capabilities that we have. Now, last time in this forum, we talked a lot about, um, the direction the assurance is taking strategy wise, right?

Um, and that was mainly focusing on leveraging various Cisco technology, how we can use those and build a sophisticated assurance solution, a solution that works for full stack visibility, a solution that can provide our customers, uh, visibility into their user experience, right? Um, we talked about four key elements that power assurance or power, the user experience, the health of client network, devices, infrastructure, connectivity, and application. And our commitment remained the same.

Still, we are still enhancing the visibility into these four key elements. Every time we build a new assurance feature or trying to improve an existing one, we always go back to four, three major pillar of our assurance strategy. Firstly, detection, right?

We want to make it really easy for our customers to detect and visualize issues in real time. Secondly is remediation, right? We want to provide our customers a clear, actionable remediation steps so that they can quickly resolve issues and reduce mean type resolution, right?

Then the final thing, the key thing is confidence. Confidence of, of our customers. How do we gain the trust of our customers, right?

We want to empower our customers with all the tools that they need and the real time data that they need to gain trust into the insights, the why behind the insights that we are providing, right? And that's, those are the three main pillar that we are equally putting a lot of effort in, uh, to provide a, uh, provide a visibility, and also help them provide great user experience, right? So that they can detect and remediate issues with confidence.

Now, this is the exact slide we shared last year with you guys. Um, like this is the roadmap we shared and we told you like, okay, these are the, all of the stuff that we are working on and plus some new, new stuff that I, uh, you see on the slide, right? We worked, um, over the last year, we made a lot of progress, uh, from making assurance easy to use, more intuitive and intelligent.

Now there's a, as you can see, there's a lot of green check marks here, right? So instead of going into each one of them individually, it, it'll take us the whole day. And I think right now we have only 20 minutes for our session.

Um, we are gonna go into the demo and showcase some of this feature, bring them to life and show you live, like what the, what these capabilities are. So this is the assurance of our view page that provides our customers the visibility into the health of their network, full stack network. We designed it in a way that it works and scale at in full stack networks.

Yeah. I just have a clarification. You had said earlier they a couple times and customer and client, who are you referring to?

The, So when I'm IT guys who are managing Meraki network mm-hmm. Or the actual end user who's experiencing the problem? So the customers are the IT managers who are managing the network and end users are the one that they're supporting that using their networks.

Do end users have access to see this assurance information? They don't. It's the, it's about providing those IT managers visibility into how the, um, experience of their end users when they're accessing network or trying to access resources over the network, right?

So this is kind of an, uh, what it provides is a continuous and proactive monitoring, right? As you can see with one score, we are trying to showcase, uh, the health of the network. So it, it gives you continuous monitoring.

It also lets you know if there is an, uh, degradation happening in any part of your network. Now, how do we come up with this score? It comprised of the four key elements that I mentioned before, right?

The health of clients, devices, infrastructure, connectivity, and application. Now, if you look at the client section, the way we are doing this as based on connection type, how are clients connecting to your network via wireless wired or remote? Are they having any problem in any of the steps that they, they're taking in connecting to the network?

Uh, like association authentication, IP address, are they seeing any kind of a loss latency? Um, all this sort of stuff, right? And then you can quickly go in and then look at which clients are impacted by which issue.

So you can see there is one client impacted and it's impacted by not yet p response. So you can get more details and more visibility into, um, what sort of issue that your clients are having. Then similarly on the network device section, this is more about the device health, not the infrastructure connectivity, like your STP connection or connection with your, between your AP and switches, but more about the health of individual devices.

How's your PCPU memory doing? How's your, um, security appliance utilization is, right? So you can take a look at them from here.

If there's an issue, it'll show up here and you can see like, hey, uh, okay, so the device utilization is picking around this time and it's an issue, right? It might be impacting your, uh, the, the clients in the network, uh, when the device utilization is picking, What kind of utilization would that be? Are you talking like CPU utilization, your ISP bandwidth utilization?

So this one that I just showed is based off of, uh, security appliance CP utilization. Good. But what part, you said the, the WAN security.

WAN security. So when you're looking at WAN appliance, that's looking at the security appliance, CPU, uh, usage and then also like same way if you look at switches, the CPU usage of the switch individual switches, um, and how they're doing it, just the actual Hardware Component. Yeah.

Then the infrastructure connectivity that looks at across 20 metrics monitor, 20 metrics across lan, van, VPN and RF environment. Then application thanks to thousand ice. Now we have visibility into application as well, right?

The whole thing is to provide visibility, provide visibility end to end. Mm-hmm. Like how is your, how's the connection between a client and an end application?

Right now infrastructure connectivity is one of our, uh, newest edition. We just recently added it, and you can quickly see here, um, hey, there is a, like a high latency that was happening on when one, and you can quickly surface that here, right? But, um, all this good, like you're detecting the issue.

Uh, one question that network engineers always find themselves asking when, whenever they're seeing issue and everything was working fine and there was an issue is what changed, right? So if you, if we go here, we see the score timeline, what the last day, and we see that there is some problems over here with infrastructure connectivity, and we just saw there was an issue with the when appliance, and if I want to know what changed, I can just simply look at the configuration timeline and then look at what exactly changed in the network that is causing this issue, right? And also at the same time as you can see, when the score is reverted, the score goes back up.

Now that's the power of this, um, at a glance view of like, hey, configuration changes and all of the network data is in one place. Yeah. Could this be added with AI to have told you that rather than having to go into the timeline to pull it up, you have the data mm-hmm.

And it's just a query saying, if I made a change, did I have what went worse? Mm-hmm. Then report it.

Yeah. So we do, right now there is some intelligent into like how we are detecting this configuration changes and if they're impacting network, right? To rely on our experts and then see like, okay, these are the changes most likely to impact your network performance.

Let's try to monitor them whenever there is a ion, right? But with ai, we can build solution and we plan to kind of go heading that way where we, we can use AI to improve the accuracy and improve the of accuracy of that correlation. How is this change impacting your performance?

And right To, to build on that, are you baselining some of these where you're over time you're establishing specific baselines or are they hard set, you know, so latency for example, right? Mm-hmm. Latency on a wan up link can be because the ISB change or something like that.

Are you baselining that latency or are you just setting a hey, if it's over 200 milliseconds of latency, that's bad. Yeah. So the way we are baselining this is we are looking across the board like all of our customers right across Meraki and how this latency laws, these metrics that we are using here and the thresholds that we need to use, we define those threshold based on that, right?

So that is kind of the first step towards having more, um, having thresholds that are more, uh, relevant. Like what have meaningful impact on your network. Sure.

That's how we are determining it, but soon what we are planning on doing is how can we personalize these thresholds, right? Sure. Just looking at the specific network that you're looking at, looking at that data and showing, coming up with those dynamic thresholds that We can find.

Okay. So they're fixed threshold values for Yeah, for most of those. Okay, cool.

Thanks. I noticed when you shared the web, um, latency being high, it was still showing the, when connection is green mm-hmm. Yes.

And it's still good. Yes. And that is a bug When, And that is a bug.

Uh, We're supposed to see that, but, uh, yeah. Come on, Peter. Uh, yeah, there is a, there's a lot of, uh, improvement that we are doing to this page continuously, and one thing I would just wanna call out the amount of effort that the, the entire, our entire team is putting into this is the moment our customers use one of these feedback button and tell us like, something is broken, Hey, this is not how it should be.

Those becomes a bug and we fix them as soon as possible, right? Um, and this also helps us because one thing, um, to do, like one thing to keep in mind with this kind of score, they're really subjective, right? How can we make them more objective?

This is a very challenging task. So this kind of feedback loop from our customers is, is helping us get there, right? Yeah.

To, To, uh, just a quick note for this, uh, I've, I've did some troubleshooting with Meraki, and I have to say it has improved quite a bit since last time I did. Um, one thing that would be useful then for the feedback is be able to give a score, but maybe for specific clients, because I was doing like, um, troubleshooting for hospital and then they had issues with like iPad and ultrasound machines, but then the VIP headset was working well. So if I can say, you know, the score doesn't really, and the score was good, the core didn't really reflect what we were seeing in the field.

So if I could say, you know, the iPads are having a little bit more trouble then my VOIP phones, then maybe it would be, um, you know, it would be nicer for the, the admin. Got it. Yeah.

Yeah. So Jonah says actually something he's gonna share today. Okay.

And that will, I think, answer your question, like how you can provide feedback when there's like a specific client that you're troubleshooting and it's not lining up with Yeah. The score is not lining up with like how the problem that this client is seeing. Yes.

So this is nice to have like an overview. Yeah. But usually when we do troubleshooting, we are looking at one client and we're, we're trying to kind of pinpoint what's wrong with that client.

Got it. Alright. So now this is a, this is a great tool to, when you're troubleshooting just one network, but most network engineers, they're in charge of multiple sites, multiple networks, right?

So how do we, how can we help them scale that? Soon we are going to launch this organization summary page. Some of you might have seen this already on dashboard, um, but it looks different because we are revamping that page to be more assurance centric.

Um, the way this page helps our customers, the first thing is off the bat you can see any of your critical devices are down. Are they offline? Are they disconnected from the cloud?

Right? And you can just get that information from the first section here. Then you can quickly look at your organization insights, how many networks are impacted in the last day, and which networks are performing the worst, right?

Or which networks are trending down. Then you can also look at any network from this place and then see what is the top issue and what kind of impact it's having on a network in terms of clients and network devices. That's the question.

Just what, what it might be a simple question. What do you define as a network, Uh, network here? Um, so the very different network is based off of like, uh, the network you create on the dashboard.

Those are the network that shows up in this. Yeah, you can call it site id. Yeah, yeah, yeah.

You can call it a site. Maybe let's say in this building, if you have one network, you can either have it on one network or you can divide it out into two different networks on dashboard, and that becomes the definition of network for us as well. Right.

I know we're running out of time here. The last thing I would say is these are a good tool, like as I mentioned, the three pillars, detection, right? It's good in terms of detection, but how do you remediate issues, right?

Recently we have added RCA for around 115 alerts. Whenever see an alert, the fundamental questions you ask is like when you talk it, how frequently it's happening, what does it mean, what kind of impact it's having on your network, right? Once those are answered, you can just simply go to the suggested actions and then take or do a, like a root cause analysis and take steps to fix the issue, right?

From here, instead of trying to navigate across different pages on dashboard, trying to figure out where a device is located, where the switch is located, and then, uh, do the troubleshooting, you can do all of this from here and then solve it from here, right? This is available already? Yes.

So, uh, there are like 115 less. So the one that you're seeing here, this is more like a detailed RCA, but there is, uh, another type of RCA that we did is, is more, uh, basic. It's basically shows you those answers.

You uh, it provides the answer to those fundamental question. Mm-hmm. And then also give you such distraction that are more text-based, not like as intelligent as this one that I'm showcasing here.

Okay. But those 150 NRCS all are available right now. I feel like this page is, uh, hidden too far in some menu.

I feel like this page is hidden too far in some menu. It needs to be brought out more. Mm-hmm.

Just, just in my opinion. No, that that's a, that's a valid bit feedback. Uh, um, alright, so I know we are running out of time, so I'm gonna hand it off to Jonas, uh, to talk more about like, hey, you don't always have the time to sit in front of a laptop and monitor a dashboard.

How do you get notified? Right? So thank you.

Right now I'm gonna grab a chair because yesterday was leg day, so Need a little bit support. Uh, first and first. Um, personally it's been on my bucket list to present to you because I've been a fan of the event for years.

I'm really happy to show you something cool today and as Iba was talking about, um, we definitely wanna provide you with an option to get the right level of attention for the things that are happening within your organization. And the way that we did it, it's, we took a page that you've probably seen previously and actually he was showing it, uh, a couple seconds ago, the alerts page and you being requesting a lot, uh, for a feature that basically lets you get the alerts that you really need. Those alerts are in really, really important for your business needs.

And what does it mean by that? Well, it's simply called alerts profile. So now you can configure alerts that are gonna get triggered only for the things that you care about from the places that you care about your networks.

And also we're gonna send a notification in a mechanism that makes sense to you. So to give you a quick demo here, we have, uh, a profile that is called East. It includes certain type of alerts.

You can see here it's called Reachability and I've chosen some bad internet connection alerts, ISP issues from reachable devices. You can choose anything you want. I also choose which networks are gonna get triggered from, right?

I'm gonna assume these are all my East coast networks. And then I choose what email and what web hook servers I wanna send these alerts to. You might have a web hook server that is just for the east side of your deployment.

You can choose it here if you want. Have one that is for the West coast, then you can choose it here. You say that and one, once one of those alerts gets no, gets triggered, then you cannot receive that notification.

You can configure several, uh, profiles. It's all a matter of what makes sense to your business. Any questions on this one?

Okay, Now can you, Can you zoom in so we can read what you're doing? Zoomin on the page? Yeah.

Oh, thank you. Mm-hmm. Alright, so I'm gonna move next to a feature that is gonna try to address the concern that you brought up.

How can you troubleshoot a single client at a time? We have identified that most of you probably spend more than 50% of your time probably shooting a single client, a single, a single device in your network. And to be honest, this is probably the page that you are pretty, pretty, pretty familiar with.

Uh, when those situations come up, right, you open the page, you find a client and you need or you want to guess what's happening with it. Now for those of you that are not familiar with dashboard, this is what we call the detailed page view for a client. It has all the information for a particular client, but there's a couple of things that this page does not a good job on.

First, it only shows you the current information for the client. It doesn't, it doesn't let you figure it out. When an issue began to happen, usually your tickets, like tickets that you get will ask you a question, when did the issue begin?

It might have began three days ago or, or way, uh, like a week ago. Go. Also, you don't know what the issue was from here.

There's no way for you to tell you. Now if you wanna guess what the issue was like on Monday, what you will do is you're gonna start jumping between page logs or event logs on dashboard and you can start opening tabs and you wanna waste like 15, 20 minutes doing end. That to be honest, you don't need to do it.

We already have all the data that that tells you what was wrong with the client. And we have done that work and we have brought up a new revamp page that's gonna just tackle that use case. Word of caution.

This is live code, this is a live network. Uh, and this is the first time we're showing this to the public. And I wanna put your attention, call your attention in this timeline view here.

This timeline will show you every onboarding issue that the client has seen on the timeframe that you have selected. Now you don't need to be jumping between pages. All those events will be shown here.

And what I mean by onboarding issues, well association authentication, the SP and DNS problems. And that means that if you are asked what happened with this client on this time, on this timeframe or this particular timestamp, you can just click on it and go back in time. And now we're gonna tell you exactly what the issue was.

Here we can see if we have a DS DS P problem and I can expand and even tell you if that issue was caused by another Meraki device within the network. Quick example here we can see that the client couldn't connect to the network because he have a GCP problem and we see that the next device is complaining about not having any more IP addresses to give. Now you don't need to be guessing and opening logs across pages to put two and two together.

We're gonna show all of that in within one page. More importantly, if there's an RCA for one of these events, like the one that the back has shown, we're gonna bring that RCA window here and you can fix a problem right away. You don't need to be moving to a different page just to fix a problem.

There are two more points that I wanna mention really, really quickly. One of those tickets that you get, uh, usually don't have a good information about the location of where the issue was seen. For example, they might say, Hey, I was on the fifth floor and I don't know what access points I was connected to when I saw the problem.

So we are also solving for that because at topology view that you see here is now going to include historical information. What that means is that if you click on an event, we're gonna do something very similar to the virtualization feature called Snapshot. We're gonna take a snapshot of the, of the topology and once you click on it, we're gonna bring that in here.

So if this event happened on the fifth floor, the topology here will show the topology on the fifth floor, but if this one here happened on the second floor, we're gonna show the information for the second floor. So you don't need to be asking the user, Hey, which window were you close to on the fifth floor to figure out what was wrong And something, yeah, go ahead. Real, Just real quick.

Um, so you mentioned authentication is one of the things that you, you're classifying on. Does that include roaming issues? Uh, no roaming issues, just pure authentication like radio by password and so on.

We're gonna improve the data we bring in that include other, uh, expanded features like authentic, uh, uh, roaming and on, um, pretty soon. Okay, thank you. Sure.

Last point really quickly. It's applications. A lot of you usually get tickets from users saying they cannot open an application and it's usually not the network is the application having an outage.

So thanks to thousands we're gonna bring that data in and if there's an outage for office, we're gonna show that here. And that will be an event on the timeline. So you'll click on it and you can see, oh, there's an outage for office.

All my line is good. I don't need to do anything else. It's just a problem.

Uh, on the uh, application provider, You said that was a separate application bringing in that? Yes. Uh, this is thanks to the integration we launched with Thousand Ice last year.

So if the integration is active, you're gonna see that icon here. Alright, so that's the new feature we're gonna be launching soon. Um, if you wanna participate in our private beta, please feel free to sign up.

I'm gonna put the QR code real quickly here. Um, there's two more pieces that I want to call out. If you also wanna be part of the beta for AI assistant, please scan the code as well.

And if you're gonna be at Cisco Live, I'm gonna be talking in depth about all these features on the technical seminar Sunday. So feel free to, uh, sign up if you wanna know more. It's four hours of just assurance and two hours of me talking about the feature that you just saw.