Cisco Systems, Inc. (CSCO) Earnings Call Transcript & Summary

Hey, good to see you. Welcome to the webinar. My name is Steven Heinsius, and I lead Software & Buying Programs for EMEA. And today, we're going to talk about a very important topic in the industry, which is about full stack observability. And I learned that there is a new acronym that I didn't know before, the DEM. We're definitely going to talk about that. AIOps, what is AIOps? I'm happy to have here with me a subject-matter expert on FSO, Ananda. Please introduce yourself.

Thank you, Steve, for having me here. I'm Ananda Rajagopal, VP Product for Cisco AppDynamics and Full Stack Observability.

Wow. Well, welcome to webinar.

Thank you.

Full Stack Observability, it can mean a lot of things for our customers, right? So help me, what's the definition that we can work on?

Yes, Steve. What has happened over the last few years is customers want to move away from something which is reactive and siloed to something which is more proactive and ideally more predictive as well. And that means we can't be working in a siloed fashion anymore. It is important to bring teams together. It's important to have multiple vantage points from which applications, infrastructure, security, network can all be observed and map them all, most importantly, with the business context. And when you do this together, that's when the business gets the most value by applying the technology benefits back to the business. So that's what we call as full stack observability because it addresses multiple personas and multiple vantage points where each of these domains, like network infrastructure, application, et cetera, are considered as vantage points.

So different customer challenges, customer states desired outcomes, and then making that visible and actionable, that is the problem that we're solving with the FSO?

Yes, totally. So we want to provide the most performance and the most secure experiences, both for applications and also to end users.

Okay.

So when you think about that, everybody is trying to become a digital company today. So digital experience matters a lot, right?

In a digital world, it's all about the experience is what we say.

Yes, exactly. You're using a mobile application.

100%, yes.

And perhaps you've got difficulty in checking out something that you want to buy from your favorite e-commerce company. How do you know where the problem is? More precisely, how does the application owner know where the problem is if you're having trouble? Is it because of the mobile application? Is it because the back end? Is it because of the network in between? I can already see that these are typically different themes. And we can't be looking at siloed data and expect that the superior digital experiences offered to you as an end user unless we're looking at all of it together.

Because for me as a user, you're mentioning, I have applications on my phone. And for me, the app just works, right? And I never think about...

Yes, most of it.

Yes. And I never think about it that what's behind that. But in reality, this application could be stored in a cloud environment, which is outside of my IT premise, right?

Yes.

So them being able to analyze where challenges, where a hiccup, you're making it that visible.

Absolutely. And if you unpack it even further, that application could actually be, as you said, running in the cloud environment. It may be making certain back-end calls to other Software-as-a-Service, SaaS services.

There's extra connection in there, and it could even go through another cloud environment.

Exactly. It could be going to another cloud environment. You could also be accessing a database perhaps that is kept on premises. So there's a lot of other touch points. Some of those applications could be owned by the customer. Some of it could be third-party back-end services that they're using. And so understanding this entire journey from U.S. end user, and they're doing this across millions of users. They being the provider of that specific mobile application. And that's just 1 form factor or 1 touch point. We also use our laptop sometimes, could be from a browser. And we could be doing that from our office. We could be doing that from our home. So there are just so many ways in terms of how we get these digital experiences. And it's important that for the provider of that experience, they have 1 holistic way to do that. And I go back to the retailer example because we just have so many touch points with companies that offer these offerings to us as consumers.

So the visibility and the application experience, that's clear. Is there other use cases that we are addressing with FSO?

There are. There are actually multiple use cases for full stack observability. We generally start off with application performance monitoring. So it's -- one of the things we often say is that the journey to full stack observability starts with APM, application performance monitoring. It doesn't end with application performance monitoring, but it's a journey. And the keywords there are it's a journey because we don't expect this to happen overnight. Because especially when you're dealing with large enterprises, they've got pretty complex processes and people. We all know it's not just about technology, it's also about the people and processes that we should be mindful about. So when you look at that, it's a journey. So there are many use cases we have identified. One of the use cases is application monitoring. Application performance monitoring could be done for something which is more of a hybrid workload, which perhaps could be a more traditional workload. Perhaps it's got code that was written maybe 15 years back, monolithic application. And maybe there are some components of that, that is beginning to leverage newer paradigms such as micro services. Or you could have something which is completely cloud-native, and I'll call this cloud-native application monitoring. So that's 1 major use case. Now we live in a world where security is no longer optional, right?

Absolutely.

Nobody wants the data to be compromised. There are regulations in fact, in many countries, in many continents where that's an important factor. So there's also business risk observability. So how do you look at application security? Or how do you look at data security for these cloud-based data stores? So that is under the use case of business risk observability.

So those are like assets auditing and security policy checks or...

Think of it as vulnerability assessment, but mapping those vulnerabilities in software to a business context by assigning a business risk score to it. So with vulnerability scores, you get a score such as this is a CVE that has got a score of 8. Okay, that looks -- that's already scary on a scale of 1 to 10.

On a scale of 1 to 10, it's not good. Yes. I would say, "Oh, no, risk of 9 to 10, that is high."

Risk is high. So it's high. It's a very high risk, right? But you don't really know what is the impact on the business unless you've kind of mapped that to a business context. So one of the things which we try to do in FSO is mapping that to the business context by analyzing the business transactions and saying that this CVE has got a really high business impact because it is being used for a customer-facing business transaction, right? Whereas maybe something else that they have the same CVE score perhaps that is used...

Lower priority.

Yes, something which is more of an internal private infrastructure and therefore, the exposure for that or the ability to actually capitalize on that for a trajector is not as high. So that would have like a lower business risk.

All right. So we have the application experience. We have the business risk observability. What else? .

You spoke about digital experience monitoring. That's the third one.

Okay. That's the DEM.

That is DEM.

The digital experience monitoring. Okay.

Right. The fourth one I would add is cost insights and resource optimization. So Steve, I often like to say that everybody loves to go to the cloud, right, and make use of -- the agility is awesome. The flexibility is incredible. But there's also 1 unsaid thing, that the cloud bill only goes in 1 direction, right? And the finance people don't like that.

They don't like that, no.

So they're looking for help. Many of the customers that we speak with are looking for help in terms of give us some insights in terms of what is the actual utilization. Is my Kubernetes workload really requiring all of the resources? And when we look at customers' workloads, we often find that only about 20% or 25% of the resources are used many times, which means it's up to 75% to 80% of wastage in the resources. That is how much of over provisioning that's happening.

So by making that visible, the customer can make a more conscious decision on the scoping and the size of the virtual machine, for instance, and then save a lot of money by...

Save a lot of money. Also have an accomplishment or attain an objective from the point of view of sustainability because you can almost think of cost as like a proxy for savings as well, energy savings as well, right? So the lesser that utilize, it means that the energy consumption also of a virtual machine or a container, for example, goes down. And that leads into the sustainability initiatives that [ equity ] might have. And furthermore, there's a second part to it where we also provide recommendations by running AI or ML algorithms experiments, as we call it, in the background to help provide recommendations in terms of how those workloads could be optimized. So that's called an application resource optimizer. So cost insights and application resource optimization, very important for FinOps, financial ops people, who are responsible for budgets and also for Personas who are responsible for making sure that they are making the best -- getting the best out of the infrastructure. So that's another big initiative.

So what I'm hearing is that, of course, in fintech and financial industry definitely, but also for the IT manager and the IT operator, being able to have full visibility and run a report to show how with the FSO solution, we're actually saving money can help them justify getting new budget for new projects, right?

Exactly, exactly. Tool consolidation, as we say, is a pretty big prerogative for many organizations. Built a study in 2023 and found that organizations use anywhere between 10 and 100 tools for observability. I'm not even talking about security tools.

Yes, only observability.

Just in observability space. There's some where actually closer to 100 tools.

Oh, my god, because they all have their own dashboard, their own language, their own settings. And to get the most out of it, you need to properly configure it. And do that for 100 tools would probably have a team of 4 or 5 people max to run it.

Yes. And oftentimes, this starts off because there was perhaps a particular problem statement that they started with a narrow use case. And they made the right decision at that point in time in terms of deploying that particular tool, but it's like going down a cul de sac, the tool doesn't necessarily solve or provide enough value beyond that specific use case. So we see a lot of initiatives from customers who are now questioning, do I really need so many tools because there's so much of overlapping capabilities. And why don't they have a true open platform, a true extensible platform, something which is flexible that kind of helps bring this together and fund that particular initiative. And sometimes, the answer to, "I've got 15 tools. I need to consolidate," could be, "I first need to introduce a 16th tool," right? Because it first integrate on my journey towards consolidation.

I get it. I get it. Ananda, I do want to dive in a little bit into the new 3-letter acronym that I didn't know yet, the DEM, the digital experience monitor. What is it? What's the outcome? What is the customer problems that we're fixing there?

Yes. It's such an important topic, Steve, in terms of digital experiences for the end users that a company serves, right? We spoke about an e-commerce company before. But actually, it addresses everybody. Even at cisco.com, we've got end users coming in who are accessing our website. So they are also digital users. So if you look at the various touch points, there's users who are accessing a front-end application from a browser, could be from a mobile endpoint. It could be also using a back-end application that is running in cloud or on-premises, as I mentioned. They could also be using back-end services that are other SaaS services that are being used. They could also be the intervening network. It could be the public network or it could also be a private network, what we call as a critical network. And so there is a variety of technologies that are being used in order to help solve these problems. So there are technologies like real user monitoring, where you're actually instrumenting using JavaScript to understand how the browser is responding, or understand what the mobile experience is. You may also need to understand how those SaaS services are performing. For that, you need something called a synthetics. You may also need to understand how the SaaS API endpoints are performing. Again, you use synthetics for that. You may also need to understand how these long-tail sessions are performing using a capability called session replay. So think of this like a network DVR, right? And just if you can go back in time.

A digital video recorder.

A digital video recorder. Thank you. So just as you sometimes have to go back in time to understand what actually happened.

You want to see it again.

Yes, exactly. Especially in sports, right? You always want to go back. I want to see that goal being scored again. So it's a similar capability that's being brought into sessions, where sometimes, what happens is when applications are built, we always deploy them in a lab environment. We test it out. Or even if it is in a production and even if it's in a real environment, it's not at the same scale as a production environment. And so there's always this long tail of problems that comes up, which are very hard to diagnose in a lab environment or during your trial phase. So this capability of session replay, it's a technology that we acquired from a company that we acquired based in Europe called Smartlook.

Okay, yes.

And it's less than 8 months since we acquired them, and we've integrated the technology into the Cisco Observability Platform. So that is integrated with these release monitoring technologies for browser, and we'll also be doing it for mobile release monitoring and provides the session replay capability. And last but not the least, the synthetic capability comes from [ tows ice ] and enables measurement of the Internet performance in terms of packet loss, jitter and latency. And for private networks or for critical networks, as we call it, Accedian, another company that Cisco purchased. The network assurance for those critical networks is brought in. So all of these are components that go into this digital experience monitoring solution. You can see that we are covering practically all points that could potentially be failure points that a customer could face, and that's why we are super excited about the potential of what digital experience monitoring. We are -- in my mind, we are redefining the definition of what DEM is because only Cisco, Steve, has got the assets. Only Cisco has got the assets across the network, across the application performance, across the synthetics and mappings are to the business context.

That's built on 40 years of experience in networking, right?

Exactly.

Amazing. So I'm excited about what you're sharing here. How are our customers responding?

They totally love it. They love the fact that there's an integration of technologies. They really appreciate the fact that Cisco is bringing together these technologies from different businesses that we have. And we've done a very quick job in terms of integrating these technologies into 1 solution. So one of the nice things is they're seeing the cisco Observability Platform truly is something which is providing this cross-domain insights. So many times, we speak with our customers, who are perhaps in an ITOps or application operations responsibility. And they say, "Well, the first thing I need to do is mean time to innocence." It's not the domain that I'm responsible for, which is, is it the network? Is it the application? Is it the front end, right? So triaging, we should respect the fact that that's a very important aspect in terms of how organizations detect and get to the root cause effect. So the fact that all of this data is in 1 unified data model, which is built on 1 extensible, flexible platform. It really doesn't matter if it is 1,000 eyes or a third party, even non-Cisco, or if it is developed something from Cisco AppDynamics or developed completely organically within Cisco. It doesn't matter because that's the beauty of an open standard, and it's built on open telemetry.

It's data. It's open and extensible, can be northbound, southbound, shared in other applications, fantastic. So I heard you mention AIOps as well. And I know SecOps. I know DevOps. AIOps is new to me. What is AIOps and how are you contributing to that?

Right. So AIOps is short form for artificial intelligence operations. And the term AIOps is kind of stuck with a lot of customers. And the capabilities in terms of AIOps have begun to mature. One of the reasons our customers began to gravitate towards AIOps is because they look at these various tooth that were siloed. And one of the...

You're Crossing the bridge.

You're crossing the bridge. And in terms of the life cycle of how incidence gets managed, there's also the big domain of IT service management or ITSM. There is oftentimes a configuration management database, which is kind of an inventory of all the assets that a company may have. There's also IT service intelligence. There is then the network aspect. It is business units that could be using something else. So there's like a variety of things, which a company could be using. In fact, a fantastic story is what Cisco ourselves is dead as customer 0 for the Cisco Observability Platform. We look at bringing in all the data from multiple assets that we had, multiple sources, whether it's Cisco or non-Cisco. And we had ServiceNow, we had Cisco Network Infrastructure. We don't use non-Cisco infrastructure, by the way, in Cisco IT.

What a surprise.

What a surprise, exactly. But there were other non-Cisco elements that are being used in terms of software. They use the Cisco Observability Platform as the foundation in order to build an AIOps application, wherein all these data sources were brought together, unified view, cross-domain correlation and providing, most importantly, the business health view across these domains.

In a visual way.

In a visual way, right. And also doing this correlation across them. They've developed it so rapidly, we began to demonstrate this in 2023 to many customers. They loved it so much and said, "Why are you just demoing this? We want to actually see the same thing and use the something like that." So in 2024, we've announced an AIOps application on the Cisco Observability Platform that now customers can get exactly the same benefits that Cisco IT has delivered for all of us as employees of Cisco. So super excited about that. And we also believe that strategically, there is a convergence that's happening between the world of observability and the world of AIOps. And I've spoken with so many customers who say, "You know what, why should I be having a separate AIOps of platform?" Logically, this sits within the domain of absorbability. And so that's just such a logical use case for full stack observability.

So to tap on to a little bit more of use cases for customers. I've recently seen a demo on modules that you can build on the FSO platform. And I was super impressed because it's open and extensible. Actually, that's the reason why you can integrate so quickly. But we also have partners now building on the FSO as a platform and can really build their own [indiscernible], so use cases. And not just that, but they can publish it in -- I think it was the FSO Exchange, is that...

That is correct.

FSO Exchange. And there's even specific use cases for specific segments and industries. So even if our customer and the audience is thinking, "I have this business challenge. Could we solve that in FSO?" The answer is yes. I don't even know what use case, but the answer is yes. Because it is open, programmable, and you can tailor the data to the needs. Do you have a good example of...

Yes. First of all, the answer is a resounding yes. And here's the reason why. When you build it on the right architecture, a unified data model, which means it doesn't matter if its metrics or events or traces or locks. It's all represented as 1 common unified view. If it is built on top of the right open standard, open telemetry, second most popular open source project in the CNCF after Kubernetes, or that's got mass momentum behind that. Two years back, people were wondering, probably it's still early stages. Now every single customer I speak to is either doing open telemetry or is on a path towards having a project on open telemetry in their organization. What does it mean for partners? Partners can actually build and bring their innovations on top of the Cisco Observability Platform. What does it mean for customers? It means customers are not just beholden to innovations coming from Cisco alone, they can actually get innovations from partners. So between June of 2023 and January of 2024, we have released modules at s rate of 1 per week. That's how fast the flywheel of innovation is spinning. Let me give you 2 examples in terms of...

That's '23, '24 modules.

Yes, so '24. So that is kind of the impact of this innovation that's happening. Let me give you 2 examples in terms of how we are kind of bringing this to bear. Today, everybody is deploying machine learning models, our AI models, to help get the benefit of data, the massive volumes they have. As more and more machine learning models get deployed, now the question in terms of bias of the ML model comes into play. What is the drift? Because things don't stay static. Things change. So is that model beginning to drift over time? How accurate is it? Are the predictions actually matching up with the reality? So there's this entire field of MLOps, machine learning ops, that is coming up, just like we spoke about AIOps, which is front and center of many customers' minds. Now we don't claim to be experts in MLOps, but there's plenty of companies out there. So we partnered with a company by name Evolutio that has got a very strong services business and also has a very strong development practice. And they built a module called MLOps on the Cisco Observability Platform, which is powered by technology from DataRobot, who is one of the leaders in machine learning ops. That's just 1 example. There's another company by the name Aporia, which has an ML observability offering that has been introduced to the market. Another important one, a second use case, which I would say, is sustainability.

Yes, a big one.

Exactly. Climatic has actually come up with a sustainability module that enables you to measure the sustainable impact of the resources being used.

So you can actually show in a report what is your carbon footprint, what's your energy consumption and how much you would save if you...

Exactly. The carbon footprint is a good example of it. So we've got lots of use cases that we are opening up, could be for SAP, for example, because that's massive in certain verticals. There could be business insights that we could offer. There could be service level observability that we offer along with machine learning operations network and infrastructure. CloudFabrix is one of the partner of ours that's developing some great technologies on top of us. So these are all examples in terms of how extensibility of the platform could be achieved and customer outcomes can be achieved by combining Cisco offerings with partner offerings.

Great. So Ananda, I'm going to show you here my crystal ball, all right? So I'm going to ask you, how does the future look like in FSO? And how does the future look like with the DEM and the AIOps?

Yes. So as you look forward, we certainly see a big consolidation towards 1 unified model, right, where teams come together in order to build the right architecture for the next 5, 10 years. And customers today are asking questions in terms of what is that right architectural model. So whatever solution that they choose needs to kind of subscribe to some of these standards in terms of openness, in terms of flexibility, in terms of scalability because it's a big data problem, and making sure that you have that extensibility so that it's making use of innovations from multiple vendors in the ecosystem as opposed to just be beholden to 1 vendor, right? So I think that ability to make the best out of what is available in the market there is something that puts a customer on a very strong footing for having a scalable architecture for the next 5, 10 years.

Amazing. Wow. So Ananda, I have learned a lot today, including new acronyms. I think that our customers would like to hear and read more. We are sharing all the links here on the page. Please use them. Please visit and reach out to us if you want to hear more. I'm pretty sure we're going to share the link to the FSO Exchange as well so they can see the use cases developed. We're going to share all that. Ananda, do you have any closing thoughts?

Well, here's 1 suggestion. When you get back to your next staff meeting, your next Board meeting, ask your team how full stacked is our observability. Do we have redundancy in our stack? Are there ways in terms of how we could achieve these business outcomes by looking at digital experience monitoring or business risk observability or application monitoring, or many other use cases that are perhaps unique to your organization? And do we have the right scalable architecture for the next 5 to 10 years?

I think that's going to be a very interesting conversation.

Indeed.

Great. Thank you very much.

Thank you, Steve, for having me here.

Have a great day. Thank you. Thanks so much.