Unified Observability in Modern IT – Dave Donatelli, Riverbed Technologies

This is Techron tv. Hey guys, thanks for the throw. We're here with newly appointed Riverbed, CEO, Dave Donatelli, and we're talking about the rise of unified observability and where AI is gonna fit in as we go forward into the new year.

Dave, welcome to the show. Great to be here. You just joined Riverbed.

So what attracted you to the company? What do you see as the opportunity here? 'cause we've been talking about observability for a little while now, but I always feel like it's in the context of say, DevOps, but maybe it's a bigger and broader thing.

Yeah, I think it's, uh, a much bigger and broader thing. You know, the, the truth of it is people's IT environments keep getting more complex all the time, as we know, through the pandemic work from home really became a norm work from Starbucks. So with all that complexity comes a need for automation to figure out how to diagnose and solve and prevent problems, whether they're application problems or problems with people's work, you know, environment, either workstations, PCs, or phones.

As we approach that, are we gonna see more convergence of job functions and roles? Because historically we've had all these silos and sometimes I wonder if they work against this because there's the networking team, the app dev team, the IT operations team, data ops, there's more ops teams than I can shake a stick at. So do you think that we might start seeing some convergence in these functions and roles as we go forward?

I, I think I wouldn't see that in the, in the near term. I, I think the reason why there are that many roles that you see, particularly in more sophisticated and technologically advanced companies is the comple is the complexity of all the technology. I mean, simply put, the, these large sophisticated companies are extremely complex.

What you're really seeing in the marketplace is a combination of technologies, meaning moving from more point products to products that can understand more of the topology and better understand root cause where problems are and fix them through automation versus the way we do it today, which is a bunch of people on a conference call trying to figure out what's going on. Where will AI fit in? I feel like if we unify observability, we're starting to collect more data in a central location.

And is that kind of gonna be a foundation upon which we build our AI models? Uh, you're exactly right. I, the whole idea is, you know, with either AI or machine learning, the better your data pulled learn from, the better your solutions will be.

The more automated you can make things. So a key part of that is collecting the data and collecting accurate data that you can then make good automated decisions from. And you see that happening in our product, uh, product lineup as well.

We now have products out that collect data across your environment. Using that data and applying machine learning to it, you can start to automate both problem identification, but even all the way up through problem resolution without human intervention. One of the more subtle things about observability is that historically we have tracked a bunch of predefined metrics and call it monitoring and theoretically what observability we can launch queries to go find out the root cause of issues.

But do you think people know why questions to ask in the first place? Well, again, I think the challenge happens when you start to cross boundaries. So when you, when you're wondering is this an application problem?

Is this an operating system problem? Is this a network problem? It, the, the talk, when I speak with customers, you hear 'em talk about meantime to innocence.

So in essence, they want to be the person on the call that says, I can prove it's not me still don't know what the problem is, but I can say the problem's not with me. And so again, that, that challenge is when, when these problems do occur, is resolving quickly. And in order to do that again, you need tools that look out at more broader, you know, part of the problem so that you can get to that resolution faster.

The meantime, the innocence process usually occurs in a war room and everybody stays in that war room until they are, uh, allowed to exit because they have proven their innocence. Is that an obsolete process or can we get to something that's a little more sophisticated? Uh, well, I can tell you as a person who's spent a lot of my life on those calls, you know, trying to figure out what's going on, I hope it becomes an obsolete process very soon.

'cause it's no fun for anybody. And I, I can tell you that a lot of people fear that call from the CEO, the best problems I like to say is a problem that never occurs at all. And this is really the promise of technology in, in this space.

And, and one of the things that really attracted me to Riverbed is understanding that, you know, this complexity is getting worse. You know, so your challenge is getting higher and, you know, the old way of everybody on a conference call just doesn't scale. So by using automation, um, to help really augment, I'd say your smart people we're not looking to replace 'em, we're looking to help make their job easier, um, helps make those, uh, you know, terribly long.

Conference calls, hopefully a thing of the past very soon. There are a lot of people throwing around the term observability these days. What ultimately differentiates Riverbed when it comes to that particular function and task?

Yeah, I'd say, you know, first of all around observability, there's obviously a, a broad definition of what it can be. And being in a space for a long time, we've gathered a ton of knowledge about how the, the inner workings of networks, the inner workings of applications, inner workings of endpoints as we call them, which could be, again, your laptop, your pc, your desktop. So having that knowledge of them being able to unify it into a common product set, um, I think is very much what the market's looking for.

You know, I was speaking to a customer the other day, they told me they had 58 different observability point tools, if you will. So various little pieces of software they buy and then they have to string it all together to try and get a view of what's really happening. So, you know, what they'd really prefer is to buy from fewer suppliers who cover more of the problem.

Uh, for many reasons it simplifies. Um, you know, who they have to work with. It simplifies the amount of tools they have to deal with.

And ultimate, you know, more things you can get outta one tool. It simplifies both problem determination and problem resolution. And that's in essence what our strategy is, is to help do that, you know, cover more of a customer's problem set and to also do that in a way that's open.

So meaning not just restricting to our own tools and our own products, but to work with their other tools out there so that they can, you know, use those, for instance, in our ability to automate problem resolution versus just working with Riverbed tools. So I think being open is always customer friendly and I think covering more of the market is always a customer friendly position. And that's what I like about what we're doing When we see some rationalization of those tools.

'cause I think some folks are trying to figure out how to pay for the observability platform. And to your point, we have a lot of tools that people are using and each of those ones has its own licensing agreement and associated costs. Yeah, I mean certainly within, within our technology, I'll say you definitely get rationalization.

For instance, uh, we have a technology we're in beta with now, which is a common agent across our products. So, you know, as you're probably aware, if, if you're to do a poll, and I do this all the time, uh, customers hate agents, they want fewer, not more agents for a lot of specific reasons. So the fact that we can have a common agent across our tools, again, simplifies their environment, simplifies the amount of products they have to have to buy.

I'd say also across the marketplace in general, yes, you know, there is a trend, again for fewer tools that cover more versus a proliferation of more tools that cover less. Mm-Hmm. You mentioned agents, uh, we have seen the rise of agentless over the years, but it seems to me there's just a lot of things you can't do without an agent.

So do we need to kinda strike a balance here? I mean, I may hate something, but I gotta recognize the fact that it adds some value, right? Yes, it does add value.

And, and as much as everybody like to go to an agent agentless world, we're gonna have agents for a while. So again, our approach is okay if you have to have them, as you said, they need to add specific value, you know, the reason why they're there. And then second of all, you have to make it a lot easier on people, um, in order to deploy them.

So again, by having more products covered by a single agent, we certainly reduce the amount of agents people have and we manage them through the way we manage. And it's so interesting, you know, customers really are a guide and I, I was talking to a customer the other day and I said, how many agents do you have per PC in your environment? Their answer was 18.

Okay. That's a heavy load, uh, to have to manage and, and care for, and also performance burden on those. So again, by reducing the down to, you know, fewer, more essential agents, we, we help make that a lot simpler for people.

Historically, a lot of this software has been deployed and managed by the internal IT team. Do you think that the burden for that is shifting more towards vendors such as yourselves? Is it gonna be more of an ad service kind of platform or maybe managed services?

How do you think this might play out? I think in the bigger accounts, it's a little bit of all the above. Okay.

So, so if you, if you look at us, um, a lot of our great customers are the largest companies in the world. And again, they're, they're big sophisticated IT environments and they're also companies that are always acquiring. So just when you think they've standardized on one way of doing things, they just acquire another company who's got another whole set of it that they need to go ahead and integrate.

And therefore, in many of these accounts, you know, it's, it's a com combination today, right? They have SaaS products, they have things out in the cloud, they have mission critical applications that they run in house. Um, and much as everybody's trying, I think that's the world we're gonna live in for a while, which is a, a mix of all these different environments.

And again, that is one of the challenges that they face in, in supplying a great digital experience to their end users, is it's a complex environment and I think it's gonna remain. So for a while You just joined Riverbed. What is the definition of success look like for you a year from now?

Well, we're very excited. Uh, you know, we've previewed with our customers a lot of product releases that we have upcoming over the next, let's call it nine to 12 months, you know, starting Q one this year. So a, a big definition of our success is getting those products out to our customers.

I'm very excited to see our customer's reaction to those products. You know, we previewed 'em with them. They're very excited about our direction and they believe it helps solve some of this complexity problem that we've been speaking about.

And that's where we'll start to ga products like our common agent that we spoke about. Um, our IQ product continues to evolve. That's our AI machine learning product that looks across the industry as well as new management, uh, technology, again, that, that looks across all of our products versus just individual products.

So a big measurement for our success is to get those products in production at our customer sites and having them help solve some of the most critical problems. That's what I'm really looking forward to. As I mentioned, we've been, you know, in active discussions with all our customers about it, and they're, they're very excited and they believe it's really spot on for what they're looking to do.

Um, for those organizations that are not your customers, what would you say to those folks about, you know, taking another look at Riverbend? I would say this is the great news is the things I I speak about with Riverbed are, are working in the real world today. So our customers today, who, who who have chosen be River Riverbed customers are some of the largest, most sophisticated technology companies, you know, users in the world, the world's largest banks, the world's largest airlines, uh, largest government agencies out there.

So our products have run in the real world for many years. They're a fail field proven, and they scale to great, you know, to whatever size an organization you have. So they're proven products that work at scale, again, with a very active roadmap that, that really is bringing out the technologies that people are looking for to solve some of these problems of complexity.

All right. Well, Dave, best of luck to you. And as you all heard it here, it's a brand new day, Riverbed, check 'em out because, well, there's just all kinds of new ways of thinking about observability, so don't limit yourself to the platforms you have in hand today because it's a different day tomorrow.

Hey Dave, thanks for being here. Enjoyed it. Great to see you.

All right, back to you guys in the studio.