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

Great. Thank you. And in turn, J. P., I'm kind of going back to your -- like your recent participation on the Cisco Live. What have you observed in terms from that event and the industry in general?

So I would say Mark summarized it very well, even though I'm not sure if he was there in the event. But I want to add to -- what we have seen is, as Mark mentioned earlier, IT organizations are challenged to constantly reduce cost, right? And what happens as an example, one example, is they pushed towards maximum utilization, right? And when you push towards maximum utilization, you are kind of like working on the edge. And without having a complete visibility, analytics and some of these tools, you are flying blind to provide that positive digital experience. And in today's world, it becomes even more challenging to make sure that your distributed workforce, your distributed applications are catered very well and have a positive digital experience. So with that said, right, in today's environment, not only visibility and analytics, it's also can I predict something, can I prevent a problem happening, is even more important to mitigate, right?

Great. Thank you. And Brandon, what's your observation on this particular?

Yes, certainly. So as I think about some of the trends that are driving edge networking and specifically SD-WAN deployments for organizations, we talked about how the edge is becoming much more distributed today. We have more distributed users and devices accessing distributed applications. And so we're also looking, as J. P. mentioned, to -- organizations are looking to do more with less, to be more efficient, to increase their level of automation. And then, of course, security is so important for any technology buying decision. And so all of these requirements that folks are looking for, for technology deployments, from incorporating the network edge to increasing their automation to having more security integrated into the network, all of these things are incumbent upon having visibility and analytics and observability platform that can support these goals of the organization. And so in a highly distributed world, the absorbability must stretch across all points of the enterprise network edge, whether that be users and devices or campus and branch sites or multi-cloud endpoints or IoT endpoints. It's important for the visibility platform to be able to enhance automation platforms to be able to, as Mark pointed out, this idea of not just identifying when there's a problem in the network, but actually either recommending ways to fix it or proactively making that fix before it impacts users. And then, of course, with security as well, there's this idea that you can't secure something in the network unless you know what's actually happening and in the network. So observability platforms provide a foundational layer for being able to enhance the security footprint of the enterprise network edge as well. So that's sort of how I see some of these trends and the importance of visibility across them.

I'll just add to that, Robert. As -- Brandon said it well, SD-WAN is -- when I look at it from a standpoint of analytics, it represents a lot of what's happening within IT overall in terms of people trying to bring together analytics and observability to, again, execute that concert of the digital infrastructure. You have security. You have cloud measurements. You have performance measurements that are related to the networking. You have digital experience. All those things are places that are particular emphasis in many different parts of IT. And within the SD-WAN environment, you have them all wrapped up into one and are all concerns, all drive the automation that goes along with it. And again, all serve that predictability that we want to move towards, that proaction in terms of management approach.

Great. In terms of talking about SD-WAN. So what's your thought about it -- how SD-WAN can leverage network analytic -- or what can we expect from that perspective? Brandon?

Yes, certainly. So as we've said, some of the trends that we see driving the SD-WAN market today in terms of adoption and use cases, organizations are looking to optimize the user experiences for users, who are accessing these distributed applications, multi-cloud applications. And so SD-WAN plays an important role in terms of enabling and optimizing connectivity to these cloud-based applications. And so it's important for an SD-WAN platform to not just provide that connectivity to these multi-cloud platforms, things like having direct connections into IaaS and SaaS-based clouds, having on ramps to be able to use the public cloud backbone for interconnect for site-to-site or site to multi-site use cases. But it's important to be able to have the visibility and analytics that underlay some of those connections into the multi-cloud as well. And so it's important as I think about applying analytics to SD-WAN, I mentioned this earlier, to have insights across multiple parts of the network. So that includes both the underlay and the overlay. So you want to be able to have a platform that will be able to not just tell you insights in terms of what's happening at the SD-WAN management layer, but into the underlay as well. A platform that can help you understand if there is some sort of issue that's impacting users. Where is the source of that? Is it in the SD-WAN management layer? Is it in the underlying connectivity layer? Is it in the application layer? So having an observability platform that can stretch across all these different points of the network becomes increasingly important. I would also say another trend that we're seeing is organizations looking to have more integrated management across the campus and the branch across the LAN and the WAN. And so again, as you're looking to have more integrated management across these disparate parts of the network, having a reservability platform that can provide a centralized view across both the campus and the branch becomes increasingly important as well. And then I'll just touch on the last important trend that we're seeing in SD-WAN and that's around security and SASE-based architectures. So as organizations are looking to have security more integrated into their SD-WAN deployments and their networking overall, again, visibility and analytics play an important role in terms of enabling that those use cases in terms of being able to provide insights into the users and devices that are on the network, which applications they're accessing. This is foundational information for being able to set policies in terms of which users and devices are allowed on the network, which applications they're allowed to access. These absorbability platforms really provide foundational information for really being able to provide a security posture on top of the enterprise network edge.

Thank you, Brandon. And Mark, do you have any perspective from an analytics standpoint with regard to SD-WAN? Mark, you're on mute again.

Sorry. The -- yes, relative to analytics, I mean, analytics that are more tightly weaved into the infrastructure is vital, certainly, with analytics focused on SD-WAN environments. I mean, SD-WAN is going to be smarter, more -- it's going to be more predictable. It's going to be easier to fix. I mean, the tie-in between specific domains within your infrastructure and analytics is really key. And I can tell you, from surveys and research we've done, we see wide variances in terms of people using observability and analytics tools within their different parts of the digital infrastructure in their applications, in their computing side of the fence, the different domains of the network, wireless or SD-WAN and such. So from that standpoint, we really need very strong effort to associate analytics with the different domains within the digital infrastructure, not just the network but outside the network as well. And then a roll up, if you would, of that analytic information, the data, the analysis so that we're getting, again, a better feedback mechanism. What is causing that problem in the app. Isn't it nice when systems automatically diagnose themselves and indicate problems instead of having the staff running off and validating, the network is fine or the WAN is fine or the computing infrastructure is fine, the cloud service is fine -- everyone is running around using different tools. Wouldn't it be nice for systems to work together so that, again, you're getting that cross-domain intelligence? And that has an exponential return in terms of how you're using analytics to affect that concert, if you would, of the digital infrastructure.

Great. Thank you. Now let's take a segue to the Cisco SD-WAN. So Cisco SD-WAN connect any user to any application, right? It's build up on the 3 pillars, the multi-cloud, security and application optimization, all under a SASE-enabled architecture. So J. P., can you share your insight on Cisco SD-WAN?

Absolutely, Robert. So typically, any SD-WAN, right, these days, the table stake item is have [indiscernible] pulled off from brands into your respective multi-cloud communication, right? In the Cisco scenario, [indiscernible] from taken the solution to yet another level. And we are bringing a whole lot of new capabilities as part of our multi-cloud offering. And here, the idea is to connect all the users and devices irrespective of where they are located. Two applications irrespective of the type of cloud and location of the cloud, they are connecting to, right? And this is what we combine into our multi-cloud access. Without going to too much detail, idea here is to provide the connectivity to applications, whether they are Infrastructure as a Service, Software as a Service, or any variation of the services. Second piece of the solution here is secure connectivity. And again, the objective here is to provide security at the right place. And here, what we are doing is we are defining or applying the security centrally on the SD-WAN controller. But then it is applied, distributed. What I mean by that is operating system that we are running on our router is fairly mature and rich in terms of feature set. It supports a lot of security capabilities. And this embedded security is what gets applied to the SD-WAN solution when we enable it centrally from the vManage, right? Additionally, we have integration with our own umbrella or Zscaler cloud player that scope type of SSE solution. So if customers or users are interested in taking advantage of centralized cloud-based security -- or in quite a few scenarios, customers, they want to try both. They want to have security defined at the periphery of the network as well as they want to leverage certain components centrally as well. So we support all of those solutions. And that's what I mean by right security at right place. And then on the right-hand side, and that is the key topic that we -- of the discussion today is provide complete visibility. And here, what we are doing is we are gathering a lot of information, flow-based information, telemetry data, whatnot, from these routers. And these routers then feed centrally into our controllers. And then controller put the information into data lake. This information is for the process and is shown on our analytics dashboard. And that's where you see a lot of visibility, a lot of information that comes into a user basically for usage in terms of day 2, in terms of troubleshooting and whatnot. Next slide, please. So this slide basically talks a little bit more about how we are simplifying, right? So I'll double-click on this very quickly. At a very high level, as a screen shows single dashboard and our controller becomes a single dashboard for any type of operation you want to do, whether it's a configuration, troubleshooting, monitoring, whatnot. Over the last 2 years, we have simplified our solution drastically. But as we all know, simplification is not a onetime activity, right? It's a journey that you have to constantly look back, improvise based on technology, based on requirement. It's a constant change that is happening, I would say. In recent few releases, what we have done is we have provided complete automation for the -- or rather enhanced automation for the configuration that end customers are doing in terms of guided workflow. So like no one needs to open a manual to look into how to configure Cisco SD-WAN, right? And this configuration is built with smart defaults. So you don't have to put in a whole lot of parameter, just a couple of parameters, 1 and 2 click, and it's done. Why? Because a lot of information we have gathered around industry and we know what type of defaults are required for what type of industry deployment, right? So for example, there are retail deployments, retail SD-WAN deployments, we know, okay, these are the parameters, these are the type of application majority of the time our end user and customer use. So we take advantage of those things. Of course, if you want to override, you can definitely do that. Monitoring is another active topic here. And as I was saying earlier, is we provide a lot of information that is collected by our devices in a central dashboard, including health and enhanced topology. Also, we provide information in terms of network performance, like key parameters of loss, latency, jitter. And that we provide in collaboration with application. So now you know if application is performing poorly, and network is not performing poorly, then you know it's an application-specific issue, a problem in application stack. If application is performing poorly and network is also performing poorly at the same time, right, you know it's a network related problem. And you can correlate all of these things with a single dashboard, single insight. So complete 360-degree view to very quickly double down on what is happening in your network. Not only that, we provide information related to SaaS applications like Webex or Microsoft 365. And those dashboards show in a similar fashion, right? This is extended to customer application as well. So not only customers can take advantage of the predefined SaaS application, but also they can extend to their customer application because every one is pretty much running some form of variation of their own customer application to run their business. Enhanced troubleshooting that we bring in is -- to give you an example, right, we show you topology, and right there all the necessary tools, like ping trace, packet flow or network trace, all of those things are available with the right plan of single button, right? So making it very simple and very intuitive. On the prediction, we will talk in a subsequent slide. But again, we are looking into AI/ML type of use cases to bring in a predictive type of information. Let's go to next slide. So what -- as I was explaining, the bottom left-hand graph shows you have an SD-WAN fabric. And this SD-WAN fabric basically is connecting information and feeding into vManage. And from vManage, we put into our data lake that goes into the vAnalytics, right? And from vAnalytics, we take the information and ingest into what we call it as data analytics engine. And this data analytics engine processes 1 month of data. So it's built based on the similar concept of AI/ML. And here, taking advantage of the intelligence that we have built into our engine and 1 month of actual data that is run -- that is collected on customer network, we can predict problems related to customer network. And that is what I'm showing in the center of the page here, center of the slide here. So going from left to right, you have Google application, GoTo Meeting, Office 365, so on and so forth, right? And if you look at it -- it's a little bit small, but if you look at it into right below the application, okay, how many users are there, what is the path quality. And we provide recommendation. And if you use our recommendation, we can tell you what would be the recommended path quality or the change in the path quality. Of course, these numbers are -- you can see it displayed in the lab, wherein it is 99% or 95% going to -- little bit improvement. But in real life scenario -- we cannot show customer data here. In real life scenario, you will see customer, let's say, is experiencing a path quality of 85%, and based on recommendation, if we can move to 95%. So that's a level of influence we can make proactively for our customers. And as you can imagine, this -- today, the fallback loop number three step is manual because we know it will take time for customers to trust any type of automated engine, right? Down the line, the hope is that customer does trust this engine and they say, "Okay, I want to enable it automatically," right? So we will provide that option in subsequent release wherein you say, "Okay, I will take all the automation that is suggested by this engine and apply it automatically." In that case, you are into auto steering mode, right? It's like a Tesla or autopilot in a flight scenario, right, wherein is automatically self-correcting and adjusting and giving your application the highest level of digital experience, right? And why is it important? Because as we see networks are not static. They are like living, breathing animal, I would say, wherein it's constantly changing. So your sales team, all of a sudden, decides to use newer sales application, IT team may not even know. Or your engineering team is starting to use the new tool, IT may not even know. Or they come to know once there is a complaint from sales side or engineering side saying, I'm having poor experience, then IT team will look into it to say, you know what, I never planned for it or my policy is not defined for this new application or changing behavior in the application, right? So you are kind of flying blind. And we try to bring in that visibility. Not only the visibility, kind of prediction and awareness towards these changes that are happening in your network. So with that -- it's a long answer, Robert, to your question. But to -- short summary is we are simplifying our solution, bringing in a lot of visibility. And then not only visibility, we are bringing prediction into our solutions, so that we can predict, and you can avoid a problem and are in -- not in reactive mode, but in proactive mode.

Thank you, J. P. And this is also really aligned to -- the predictive path recommendation really aligned to our vision of the predicted network. Now I'll kind of pivot back to Mark. So I think -- I'd like you to kind of share your view how AI or ML are at flying -- no, analytic from what you experienced in the marketplace, and what your point of views are?

Yes. I mentioned this before, but collecting all the data, having all the intelligence is one thing, but utilizing that intelligence is really what contributes to more precise automation, problem avoidance, all the things that you want, certainly within your network environment and, again, even across your larger digital infrastructure. So from that standpoint, systems that learn, systems that are able to evaluate more deeply, systems that are able to basically relieve the burden of analysis from operators, you're no longer staring at the screen and depending on the expertise of the operator figuring something out, representing -- that was the old days of monitoring. You had the screen with all these bar charts and all these nice things moving around and reds and oranges appearing, and it was left up to the operator to figure out what's the problem. Root cause analysis, anomaly detection, threat -- used especially in threat detection, all of these things require, A, processing a lot of data and also applying a lot of smarts over time. So again, getting smarter as you analyze problems as you look at situations that are occurring, as you're analyzing the trends, certainly looking out is becoming a much more vital requirement these days as opposed to just simply helping you identify a problem. A lot of people with analytics get very focused on shortening MTTR times. And that's great. Certainly, every second saved is great, but avoiding them is a lot more effective in terms of, quite frankly, hiding problems from the user. Again, fixing things before they even happen. We're just avoiding them completely so that they don't happen at all. So from that perspective, AI/ML, tremendous value in processing a lot of data, applying intelligence to that data, looking at changes, looking at solutions over time and starting to understand about what is best for a particular situation. And then subsequently, taking action. The system taking action itself. Now we have to build up a trust in our automation systems and our AI systems and our analytics systems to feel comfortable with that, but the reality is that -- to lead to a more dynamic network and one that's better protected and better serves the user on a more consistent basis, we have to bring this intelligence to bear and provide for a more dynamic network environment because digital is all about speed and dynamics and changing. And we've seen the benefits of how this can be applied in the computing space with cloud services for the last 15 years. So the more we can bring dynamics into the networking environment, again, with analytics, with automation driven by analytics, with all of the data that's feeding the analytics machine and, again, using AI and ML to make that analytics, that analysis, as insightful as possible. And again, linking it to actions that need to be taken, either by the less trusting operator that wants to do it themselves or by automation programs or run books or other kinds of things, or directly built into the devices themselves.

Great. And then Brandon, do you have any unit perspective on AI/ML as well with regard to the network?

Yes. Thanks, Robert. Yes, as I think about applying ML and AI specifically to SD-WAN and the WAN in general, I think that the WAN is really an ideal use case for applying some of these ML and AI principles because there's such a tremendous amount of data that flows through the WAN. If you think about the streaming telemetry that you can get from the underlying WAN connections that organizations are using within hybrid WAN scenarios that are managed by an SD-WAN. If you think about all of the user experience metrics that can be gleaned from, again, those distributed users accessing these distributed multi-cloud devices. Given the importance that we talked about in terms of optimizing that digital experience as well as the IT operations management perspective, ML and AI and the components of enhancing an automation perspective, like we've been talking about, really make it an ideal fit for applying these ML and AI algorithms to the WAN and SD-WAN. But I do agree that it's going to be a journey for folks to be able to get along here. It doesn't have to be an all-or-nothing thing where you're either using it or you're not. It's not binary. There's going to be different ways that you can start to use some of these ML and AI techniques. And at the end of the day, I think this is really about ML and AI platforms understanding what normal behavior is and then recognizing when there's abnormal behavior, and then being able to alert the managers to when there is a performance degradation, a security incident. And as Mark said, hopefully, be able to fix those either manually or via self-mediation before they impact any users or their organization. And so I think that's the real value of what an ML and AI platform can enable. But there are steps along the way where we can take advantage of some of these abilities of ML and AI without turning your entire network over to an ML or AI platform.

So I want to jump in here and quickly mention that it totally resonates with me what Mark and Brandon had shared. And that is a use case -- a similar use case I was sharing earlier as how Cisco in a Cisco SD-WAN is using AI/ML in terms of bringing in more predictability in the solution. And that was a one use case that I mentioned. There are quite a few use cases that we are working on and will be sharing with the audience and our users soon in terms of similar cases where how we can bring in more and more predictive behavior, show the trend analysis, and things like that. In terms of helping our end user with day 2 operations as well as doing some type of projection in terms of bandwidth projection, in terms of their capacity planning and so on and so forth. So a lot of exciting things are happening on the Cisco SD-WAN side, which we'll be sharing sometime soon.

Thank you, J. P., Mark and Brandon. Now as we get close to the top hours, in terms of the [indiscernible] Mark, do you have any closing remarks for today's session?

Yes, if we could put up that keys to success slide, it would be great. And we've mentioned this a couple of times, but again, the digital experience is a simple ultimate measurement of the success of IT services. So -- again, it's a good thing to focus on. It's where it all starts and ends, quite frankly, for the user, and certainly for the business. So you're focused on that. Streamline your tools, your techniques, even the teams, get teams working together better, get your tools working together better, sharing of data. Tie in your analytics and management systems to your individual devices in different domains within your digital infrastructure that allows for, again, deeper insights into how those individual components are operating, but also as they're able to then graduate it to see how the components are interacting together. You want to strike a balance between resolving things and being resilient and being responsive and being ready for the next pressure point. So make sure that when you're focused on analytics, again, you're looking outward. You're having that, as it says in the final bullet there, you're taking a proactive approach, you're looking out, you're predicting, you're preventing, you're prescribing. So stay ahead of the game there. Focus as much on responding and readiness as resolution and just day-to-day, minute-to-minute resiliency. And I can't emphasize enough the staff focus. We see this time and again at IDC in surveys, in interviews with people, in consultations, is -- ease up on your staff, redirect your staff, get them focused on things that provide more value to the company and provide more value to their own careers. Allow them -- allow networking people to understand better security, to understand cloud. And again, SD-WAN is that touch point for all of those key technologies and certainly serves as a way of allowing staff working together better. And again, streamlining your teams and your tools. To a great extent, network and digital infrastructure success is really driven by having your staff do less, not more. So again, let the systems operate themselves. Let the systems get smarter. Let them have more data available to them. Again, less reliant on the staff for doing minute-to-minute monitoring or day-to-day operations or month-to-month deployments or configuration changes or maintenance, all of that time saved translates to them, again, adding greater value, again, to the business, to the company, to the organization, but also to their own careers. And that drives retention, satisfaction in their jobs, which is key to keeping good talent, which is, again, in short supply these days.

And just to add to what Mark said there and to emphasize a key point in terms of how to prioritize a lot of the trends that we've been talking about today, to me it comes down to the IT team and networking teams specifically, should be working to further enable the goals of the overall business. And so that is the North Star that organizations should be focusing on in terms of what are your organization's overall priorities, and then what are the technology solutions that are needed to help enable and accelerate those goals from a digital transformation perspective, from a revenue growth perspective, from a customer satisfaction, employee satisfaction, sustainability. These are some of the goals that we see they -- overall organizations have. And it should be the role of IT and technology in general to help support those. And a lot of the things that we talked about today are ways that you can help accelerate to ensure that IT and networking is actually helping your organization meet and exceed its overall goals.

Thank you. So with that, thank you for joining our Insider Series today. And we have additional resources, including IDC report, authored both by Mark and Brandon on network analytic as well a video and podcast. And the link will be posted shortly. And with that, I will conclude our Insider Series today. Thank you for your time.