Best Practices in AIOps Management with Payal Kindiger of Riverbed

This is Textron tv. Hi everyone. Welcome back here to techron tv.

I have a first time guest on our show to introduce you to here. Her name is Py Kinder and hopefully, I, I pronounce that right, PA We welcome, welcome to Text Junk tv. It's a pleasure to have you here.

Thank you, Alan. It's nice to be here. Absolutely.

So Pyle is senior director product marketing at Riverbed, and we're going to get into Riverbed and we're gonna have a little discussion around ai, what else is new, um, you know, here. But before we do that, I wanted Kyle to maybe share a little bit of her own journey with you and our audience so you understand who she is, where she's coming from, and what brings her here today. Pile, I, no, you know, no pressure, but Sure.

If you would in a little bit of your background with the folks out here. Yeah, absolutely. So, as Alan mentioned, I'm senior director of product marketing for Riverbed, and I cover our AI driven observability suite of products and platform called iio.

Um, and so I have over 20 years background in, uh, software SaaS, B2B space. Uh, I started my career at Deloitte doing leading large scale implementations, um, in the IT network systems areas. Um, my focus was on process design and change leadership and basically getting organizations ready to, um, accept, accept those deployments and worked with a lot of global, um, uh, fortune 500 financial services companies and, and other enterprises.

Um, and then I joined a company in, um, you know, as a, we were a solution and services provider, um, an IBM partner, uh, for many years. And, um, focused on the Tivoli suite of products. Also worked with BMC software, uh, products like, uh, remedy at the time and Sure.

And then, um, and then ultimately we, uh, bought a pre-revenue product called Resolve, um, and which was focused on, you know, runbook automation. So this was, you know, what, you know, well over a decade, uh, ago, this was at the time, uh, an early market, um, product around runbook automation. And so we grew that and I was there and, and I let the go to market for that until, uh, I did a strategic exit.

Very cool. Very cool. Um, pile Riverbed, I'm sure some folks in our audience have heard of it, and I'm sure some folks in our audience haven't heard of, of Riverbed.

Um, and the ones who have heard of it may not be a hundred percent sure exactly who they are and what they do, but they've heard of it. Nevertheless, why don't we, if you don't mind, we'll start real high and say, okay, here's who Riverbed is and here's what we do. Yeah, so Riverbed is a, you know, our focus is around unified observability, so AI driven, unified observability.

So, you know, whereas, um, today, you know, a lot of observability is tied to a PM vendors, Riverbed has a slightly different view, and that is, we believe in unified observability across the network, the infrastructure, the applications, and the end user experience. Um, so Riverbed's been around for many years. Um, I actually just joined about a year ago or so.

Um, historically, you know, they were focused in networking side and, you know, had, um, you know, in monitoring and, um, you know, solutions are on SD WAN optimization. For example, in the last few years, it's taken, um, a strategic focus on specifically unified observability, um, uh, with this brand of, uh, products called uv. Very cool.

How do you spell that by the way? Olivia UV A, yes. Uh, A-L-L-U-V-I-O.

Okay. com or io or is it all at underneath the Riverbed site? com.

Excellent, thank you. So Kyle, I just got back last week from Chicago where I was there for the Cube Con Cloud Native Con event. And, and of course, observability is such a huge piece of this cloud native landscape now, you know, open telemetry, Prometheus, I mean, all of these huge open source projects that are really fueling so much of what we see in observability.

Um, you know, and then, then so many commercial entities kind of are, you know, using that foundationally, but then building their product and their value props over that. And it, it's, I mean it's, I I, I've watched this develop over the last maybe two, three years and it, you know, shows no signs of slowing down. Certainly part and parcel of that observability landscape though is also the use of what we were calling AI ops.

But a funny thing happened to the AI ops on the way here originally, I think when we were talking about AI ops, it was very much machine learning, ML ops, more than AI ops perhaps, but now we're starting to see a AI ops or ML combined with generative gen ai, right? As like a two, a one-two punch, you know, first we, we we, we observe using, you know, machine learning pattern matching and all of this sort of stuff, and then some gen AI on top of that to synthesize, right? What we've learned or what we've, what we've recognized.

And wow, I mean, this, this is game changing kind of stuff, but I'm wondering, you know, is that the kind of things Riverbed is seeing and the kind of stuff you guys are working on? Yeah, a hundred percent. I think, you know, especially with our clients, we're seeing more integration of AI chatbot and large language, um, models.

Um, and I think that's gonna continue to, to happen to deliver those fast, rapid insights and ask the, and answer the question, you know, tell me what I need to know. Um, and in general, you know, you talked about sort of the, the, the juxtaposition of observability and AIOps. You know, observability is about understanding the unknown unknowns.

Um, in other words, you know, these solutions provide enhanced visibility, um, that includes comprehensive contextual insights, um, into the state of the entire infrastructure. So what AI capabilities do is they bring that intelligence layer to observability platforms, uh, to achieve, you know, uh, a few benefits. You know, one you just talked about, you know, predictive analytics and insights where new data means new behavior of machine learning models without human intervention.

Um, so the goal there is continuous learning and infrastructure and system behavior to prevent those incident outages. Um, you know, second is early warning of emerging incidents, and then the third I would say is automated remediation. Um, you know, we have our AI driven automation, uh, capabilities called, or, or, or we call it intelligent automation.

And I think with ai, um, and the combination of observability data with high fidelity telemetry, there's a lot more broader applications of, uh, remediation or incident response than we could have done, um, in the past. So, um, so yeah, those are, that's definitely a, a big focus, uh, for beded. Absolutely.

Absolutely. Now, look, depending who you talk to, pa, pa, right? Probably some people say this is coming, but we have time.

Other people say, you know, the time is now we, if we don't, you know, if we've gotta start making some decisions, we've gotta start preparing for this. Other people are saying, Hey, you gotta jump it right now and start using it. Forget preparing.

If you're not using it now, you're behind the curve. You know, and, and there's always going to, you know, that's that crossing the chasm model of early adopters, early mainstream, late mainstream laggards. You're, you know, it comes in all shapes and sizes along the continuum, but, you know, from where you sit and recognizing, you know, you're sitting at a place where, where the current is moving fast, what, what's your advice for IT teams now regarding AI ops observability, if they can't use the latest and greatest right now, what can they do to get ready to use the next latest and greatest and for what's coming down here?

Yeah, I mean, I think I would start with gaining alignment first on how and where they think AI ops is gonna provide the best value. So maybe start with a targeted use case where there's an overarching problem where perhaps AI could uniquely solve that, that, um, those pain points. So is it about freeing up, um, time from overwhelmed, you know, IT staff, uh, that maybe choose a recurring problem that consumes significant operational cycles, right?

Um, in that case, it ops teams can deploy AIOps, uh, with anomaly detection to obtain those early warning signals of a developing issue and then kick off, you know, automated incident response or, or re uh, remediation, sorry, to, to eliminate the, the operational, the wasted operational cycles, or is it about increasing availability for mission critical applications, um, then perhaps considering net using network observability, uh, platforms, um, that have embedded AIOps capabilities, um, so that you, you know, for example, could put all of that underlying root cause analysis and insights into a single ticket. So I think understanding what the use case is, um, and maybe starting with a targeted approach is probably the best way to go. I, I don't, don't disagree with you at all, you know, so I, I, I myself have been in technology 30 plus years, right?

And I think one of the things I've learned through all the various cycles, you know, boom, bust, boom, bust hype over hype that I've seen is, you know, how, how is our audience out here? How do we separate myth from reality? What's real, what's not real?

What, what is worth our time? What's not worth our time? What advice pile would you give to our, our audience in terms of, you know, what, these are the things you need to focus on, but here are some things that maybe they aren't so real right now and I wouldn't prioritize them.

Yeah. Um, you know, I think there's, you know, the, you know, the AI AIOps, um, market is, is generally broad. Um, you know, and then there was some skepticism early on, you know, with some of these, um, AIOps models where you couldn't, it was sort of a sort of a black box approach where there might've been some more advanced, you know, didn't quite know exactly how those models worked, and, um, maybe the outputs, you know, didn't quite match, uh, the expectation.

But I think our AI ops capabilities over time, um, the last decade, if so, uh, have come a long way. Um, so I think, you know, there's a few key dimensions of where AI ops can really help IT teams today. Um, and I would say there's probably three, uh, three main areas.

One is proactive operations. You know, a lot of AIOps vendors focus on predictive models as, as you mentioned, and leverage advanced analytics and machine learning to identify potential issues, um, before they become critical issues such as a network outage. Um, and this predictive capabilities empowers IT teams to, to take preemptive measures.

Um, second, you know, where AIOps, um, I think a key capability for AIOps is intelligent automation. Um, so it can enable network operation centers or noxs to shift left. Um, and it does this by reducing the number of repetitive tasks so that they can focus on the more complex and the higher priority issues, uh, and more importantly, not have to frequently escalate tickets to IT specialists could be better served doing more high value work such as, you know, innovative projects.

And third, um, where AIOps absolutely does help today is user first insights to drive prioritization, especially in the context of being unified observability. Uh, you know, our AIOps capabilities, for example, at Riverbed can provide user first insights and then put all of that relevant performance data and analytics into a single view with all the necessary context. So for example, you know, understanding how many, you know, if there's some downtime, how many users are being impacted, what are the group of users that are being impacted, what's the root cause of that issue?

Which locations are being impacted by that issue, et cetera. So this not only helps to reduce the meantime to repair, but also lets it focus on the most important issues, which is so important today when it is being overwhelmed with tickets and alerts. Agreed.

I, I know that. I think that's excellent advice. Good stuff right there.

Um, Paal, Paal, people maybe want to get more information about AIOps and is it alluvial, right? Is that how you pronounce it? Yeah, it's, it's El Luo, that's right.

Mm-Hmm. com and click on through to alluvial there, or is there somewhere else we could send it? com and, um, you go to our website under there you'll see our brand of products for alluvial, and, and you'll see all of our products as well as our solutions, uh, that, that we address with our combined capabilities of unified observability in the AIOps.

That's excellent. Um, I'm just trying to think of anything else you wanna share with the audience. Um, yeah, I think, you know, I think what, you know, one thing I just wanted to share, you know, around the intelligent automation piece of it, um, you know, it's, it's come a long way.

I think our capabilities and, you know, we do intelligent automation for service desk, you know, it ops, um, and, and network teams. But like putting into context, kind of what I was mentioning about, you know, those actionable insights and driving AI automation, et cetera. Um, what I love about what we do today is that, you know, let's take, um, an example where we have end user experience monitoring and, um, an observability.

com, um, and all of a sudden, you know, the response time is slow. If they have alluvial in their, in their environment, Eluvia will be able to, um, proactively detect an anomaly, basically collect and correlate all of that relevant performance data from the laptops, the networks, the apps, and then determine that root cause. And then once it's figured out that root cause, it will then replicate advanced investigative processes, you know, like an expert would do.

Um, and then using, you know, I've been saying intelligent automation, another way to say it's called adaptive automation. Sure. It basically could employ the right set of actions, um, to fix that problem.

And then if that problem is too complex to nuanced, all of that information then would be placed into a ServiceNow ticket, um, so that it would be routed with the right priority to the right level, and to the right team for resolution. So something that would normally take hours to get resolved, uh, would now be, you know, reduced two minutes. So I think that's what I, um, just love in terms of like how we're really combining, uh, the capabilities of observability and, and using AI to bring intelligence, um, and, and make people's lives easier.

Agreed. Fantastic. Thank you so much, PA Alpha being with us today here, uh, do come back and keep us posted on what's happening with Riverbed and Alluvial and, and the world of AI ops.

Okay? Okay, will do. Thank you, Ellen.

Appreciate It. All righty. Appreciate your time.

We're gonna take a break here on Text Drunk tv. We're gonna be back in just a moment with our next guest.