Advanced Monitoring and Management of SD-WAN
How does SD-WAN shift the status quo for the Operations teams? SD-WAN technology possesses sophisticated, in-built capability for operations and management, and advanced automation to address network problems in real-time. However, the Operations teams needs to seamlessly integrate this network and health status intelligence from their new SD-WAN infrastructure into their existing monitoring and management infrastructure.
Additionally, monitoring solutions must provide visibility across all network architectures including highly dynamic and complex hybrid clouds and other SDN elements.
This session covers the integration of CA Virtual Network Assurance and CA Performance Management with Viptela SD-WAN. The key topics that will be discussed include:
- Eliminating visibility gaps between traditional networks and SD-WAN providing administrators with insights to optimize and transform their large hybrid networks
- Providing relationship mapping of major IT events
- Visualizing physical and logical dependencies in applications and service chains
- Extending visibility for operators with network performance monitoring across cloud, SDN and SD-WAN networks (including service chains)
- Enabling operational insight into how all elements of traditional and overlay networks affect service integrity
Tim Diep is director of product management at CA Technologies, Enterprise Management Business Unit. A veteran of the cable and networking industry, he joined CA in 2014 to lead the concept development of SDN assurance. Before CA, he was with Juniper Networks, where he developed early-stage products for cable broadband, router services, and streaming video delivery. Tim holds patents in “applying differentiated services” and “granular access control management.”
Jeremy Rossbach is a Sr. Product Marketing Manager for a variety of Enterprise Management solutions at CA Technologies. Prior to joining CA, he spent over 15 years working in IT, across both public and private sectors, managing data centers for startups, healthcare, financial and federal system integrators. His previous roles as a data center administrator, engineer, architect and manager provided him invaluable insights into the challenges and goals of IT consumers.
CA Technologies: Advanced Monitoring and Management of SD-WAN
Jeremy: All right, thank you very much for joining us today. Good morning, good afternoon, good evening, wherever you are in the world. My name is Jeremy Rossbach with CA Technologies. I’m very happy to be here today, very proud that we were able to take part in the SD-WAN Summit. And thank you to our partners for inviting us. We will be discussing in this session, Advanced Monitoring and Management of SD-WAN Technologies. I am the Product Marketing Manager for our product CA Performance Management, and CA Virtual Network Assurance. And joining with me today is the Director of Product Management for those two products as well, Timothy Diep.
So I know time is of the essence here, so I’d like to get started right away. So, just as a brief introduction as to who CA Technologies is, at CA we definitely understand that software is at the center of any business success, when it comes to succeeding in the application economy. We understand that your applications are the primary means of interacting with your customers. It can be the face of the company, and the primary influence of how a customer perceives your brand.
So at CA, we definitely understand about software. We know software. We’re not in the hardware business. We have always been in the software business. It’s been our focus for our customers for 30 plus years now. We know how to build quality solutions for our customers. Again, over 30 years’ experience and especially in the infrastructure management arena, where we have thousands of customers using our solutions today. As you can see here, we are a modern software factory, from end to end management, from project and portfolio management, out to the cloud through mobile and into mainframe.
We are definitely going to focus our area of attention today in the DevOps space, specifically under infrastructure management. Within the three core elements of our portfolio strategy is DevOps. It’s where we provide a flawless application experience. We aim to optimize for performance and comprehensive monitoring and advanced analytics. So monitoring those applications and the underlying infrastructure from end to end, from cloud through mobile, even back into the mainframe with scalable unified solutions.
So again, we’re going to focus our session today on infrastructure management and monitoring, specifically network performance monitoring of SD-WAN Technologies. So I’m going to turn it over to our Director of Product Management, Timothy Diep. Tim, go ahead.
Tim: Thanks, Jeremy. So, as Jeremy was saying we’re very, CA Technologies, delighted to be participating in this event. We have a good partnership with Viptela. And I’m very excited to share what we’ve been doing together, in terms of monitoring SD-WAN, monitoring Viptela. But before we talk about SD-WAN, I want to take a moment, and take a step back and look at the network as a whole, how it’s changing and how it’s affecting network monitoring. Networks are changing, and you know maybe even better than us, how it’s changing, because you’re living it today.
And the changes have tremendous benefits to organizations, in terms of automations and speed and agility. But at the same time, it has introduced a tremendous amount of challenges in the form of monitoring. And that’s what we want to talk about today and specifically also SD-WAN. There are really four major components that is forcing this change onto a network, and therefore challenging the monitoring. First is App economy. And for those of you who are familiar with CA, you probably heard us talk about the App economy. But very simply, it’s about businesses today, whatever they make, whatever they sell, they have to start adopting software applications to become a digital business.
And by doing so, there are a lot more applications in your network. And these applications are, you can think of it like more users in your network. However, they are like executive users. They are very demanding in terms of bandwidth. They’re very sensitive to latency. They want availability. They want something fixed right away when there’s a problem. So as a network, as a network owner or a network monitor admin, you have to respond to the apps, because these apps are your new business. These apps is the new business model for your company.
The second force is cloud, and we all know that cloud has been changing the way that we run workloads and host workloads. You know, the appearance of AWS and Azure is really changing the way that we set up our data centers. What’s happened with that type of transformation is the traffic pattern has changed. Whereas in the past, the traffic pattern is within your premise, within your enterprise. These traffic patterns now are moving outside, into the cloud, via service provider network, via the Internet, via your wide area network. And these networks, these links, these connectivity, these locations are not equipped, weren’t designed to run, to support this type of traffic.
Hence, we have SD-WAN. Hence, it’s the emergence of SD-WAN. So, that’s why we’re talking about SD-WAN today. But this whole cloud transformation is not just about a WAN. It’s also introduced problems in terms of visibility. As the workloads move to the cloud, you still need to understand the network performance of the workloads. And today, it’s very difficult and challenging to do that. Next is SDN. SDN is everywhere. SDN is in the WAN. SDN is in the data center. SDN is in the service provider networks. SDN is even showing up in optical networks.
SDN introduces a lot of complexity. It introduces overlays. It’s not just one overlay, but many different overlays. That’s complexity that a monitoring system has to deal with. SDN introduces new objects, new devices to monitor. These are not physical devices. These are virtual, and in some cases because of automation, some of them are logical devices. SDN also introduces variances in your network. And these variances are due to vendor-specific technologies, or open source specific technologies.
And then finally, IoT. This is a big topic and IoT is many things, and different for different companies. Obviously, different verticals have different type of applications. But I think one thing we can agree on is IoT will introduce more things to monitor in a network. And if you’re changing your network monitoring system, if you’re revamping your monitoring system, you should ask yourself this question. Is my new monitoring system ready for IoT? So faced with this network transformation, and a need to revamp network monitoring, that traditional monitoring systems and the administrators who deal with these systems have a lot of challenges.
One is that using a traditional method, you’re not going to be able to understand the different layers in SDN technologies, the new object, the new component that shows up in your new network. At the same time, the variances coming from different vendors, as well as open source, there are just too many variety of architectures and technologies to deal with using the traditional method. Traditional network monitoring system would have a hard time correlating the network impact to apps. Remember what we talked about, in terms of App economy.
It’s not enough to just root cause a problem in the network to a particular network component. It’s just as important, and maybe more important for some companies, to identify the impacted applications, so that remediation can be done fast enough. With the appearance of virtualization and automation, logical objects, it’s also critical to relate these new objects, these new devices to physical devices. And using the traditional method is not going to work. With the cloud transformation, now there’s a gap, in terms of understanding network performance of the workloads in the cloud. That’s the gap that using a traditional network monitoring system, you will not be able to solve.
And then with all these challenges, you still have to deal with the legacy network. Because for most of you, you’re not going to change your network completely overnight. You’re going to do it spot by spot, focus on the ones with the best return. Therefore, you’re always going to deal with a big part of your network that’s legacy. So, that’s the challenge is dealing with all these new challenges, but you still have to deal with the legacy network.
Here’s a study that was done by EMA about a year ago, and it supports a lot of the points I was making. It was a study done, about 250 respondents. Most of the respondents are enterprise, about three quarter. About a quarter are service provider. But you can see that a large majority, as they look at specifically at SD-WAN, that they’re really concerned about this appearance of virtual components. They’re thinking ahead about oh, when I stop throwing in virtual CPE and virtual network functions, how do I deal with that? How do I deal with the appearance and disappearance, and the movement of these virtual components?
And they haven’t even talked about logical components yet. There’s the concern about variety, variances, multiple service providers. That’s going to be a recurring theme, as we move forward with SDN and SD-WAN is that there will be a lot more variances to deal with for enterprise, for service providers. And then, the third point I want to make is the word legacy. Legacy is an important part of addressing these new networks and monitoring these new networks. Is you can’t just focus on the new entities. You have to focus also on the legacy entities, and be able to monitor both side by side.
Whenever there are network changes, architectural changes, over the past few decades, there are a few things that from our experience, at CA, you know, working with customers, enterprise and service providers for years, there are a few things that we have identified as common pitfalls that we see a lot of companies do. And with SDN, with SD-WAN, it’s even more important to avoid these pitfalls.
The first is to jump to building another monitoring silos. This is something that happens a lot, in the past. And some companies have been able to get away with it, and overcome it with spending more money, in the end to unify all the tools. But in this era, with this network transformation, it’s even more critical to make sure you don’t build another silos. Because you’re going to be dealing with issues, in terms of integration, and integration will be much harder later. As well as vendor lock-in. And then, building a tool that’s not applicable to the operational team. The tool that, particularly the EMS tools that will be more fitting for the engineering teams.
The second part is something, we in CA, in the monitoring discipline called Operational UI. Operational UI are the features and functionalities in the UI that you have adapted into your company’s workflow. And this is important, because you train your personnel for years to use the tool a certain way. You ask a vendor like CA, and other similar monitoring vendors, to build the necessary UI features in the software for you, so that you can build processes in your company to adapt to it. You can’t abandon that just by entertaining a new technology. The key is to maintain the same operational UI and apply it to the new technology areas.
And then the last point is something that often is overlooked in monitoring the software scale. In today’s network architecture, you have to think about scale. If you only think about one part of the architecture, and not think about the bigger picture, you’re going to have systems that cannot scale. Scale is critical, because the way new technologies like SD-WAN and SDN and NFV are introduced, it’s using overlay technologies. And as I mentioned before, overlay technologies and automation technologies introduce virtual components, as well as logical components. They are equivalent to another interface, another device that you have to monitor.
If you’re looking for SD-WAN monitoring, or you’ve started to monitor SD-WAN already, there’s some considerations that we want to call out. You should not just focus on SD-WAN only. You should look at virtual CPE also. Because more than likely, whether you’re doing SD-WAN by yourself, or you go through a service provider, the concept of virtual CPE will appear. Because it makes a lot of sense. It’s going to help you scale. It’s going to reduce cost for you. It’s going to make your WAN that much better. So, it is almost certain that virtual CPE will appear in your SD-WAN network one way or the other. So therefore, when you monitor SD-WAN, you have to consider virtual CPE also.
And as I said many times already, you have to keep an eye on monitoring legacy network. You can’t just forget about the legacy network and build yet another system. You have to think about, how do I monitor SD-WAN, virtual CPE, and legacy together? You have to think about cloud visibility, because a big part of SD-WAN, a big reason that you move to SD-WAN is about workloads to the cloud. And with that workload moving to the cloud, you have this visibility gap. So, your SD-WAN monitoring solution has to be able to extend it to cover this visibility gap in the cloud.
And then, network analytics. This is the important investment for any monitoring system. If you’re buying a monitoring system, it should have a roadmap towards analytics. Because this is where you can harness and harvest all your data to build intelligence down the road, to do things like self-healing, to look at things like correlating between applications and systems and networks. So, CA has a concept called Converged Network Monitoring Platform, and what this starts with is a foundation platform that has a lot of shared services. It’s scalable.
It has a flexible data model. It’s highly available, with high availability and disaster recovery. You have a lot of core monitoring services, like thresholding and grouping and multi-tenancy. You also have a collection of different single pane. We don’t offer you one single pane. We’ll give you in fact three different single pane. We will give you a native single pane that you can use to see the variety of data. We give you the ability to build your own single pane, using something we call Open API, CA Open API. And then we also give you a more flexible single pane that you can use to build reporting engine on. And this single pane can also get you information about the systems, about the apps, about the end user experience, provided through CA.
We’ll give you the ability to integrate fault into the Performance Management Platform. In addition to that, we’ll give you the ability to get deep visibility into the network through flow, through packet analysis. And the important part about what we provide, in terms of fault and deep visibility, is it’s a concept called in-context. The in-context means if you’re looking at a particular area in the network, whether it’s an endpoint, a virtual endpoint or logical component or a physical component, we allow you to travel deep into the visibility area, focusing on that one context. And that’s critical for troubleshooting, for triaging.
We’ll also give you the ability, through the platform, to add coverage. And this is important with SDN, SD-WAN is to take this foundational platform tool with all the capabilities and applied to SD-WAN, data centers, NFV, or even any type of non-SNMP networks like IoT, for instance. And we’ll give you the ability also to do something called Relationship. Because we understand that the way you maintain visibility and control over virtual and logical networks is to apply Relationship. And so, this is what you get when you get the platform with these additional capabilities.
Then finally, Network Analytics. The products that we provide have the investments, in terms of being able to feed the information, feed the data into an analytics engine. So that you can use the engine to do, well whatever type of intelligence that you may decide to do. In some cases, will be intelligence that we provide you. In some cases, intelligence that you do in-house in your own company. Why is converged important? In our definition, converged means visibility plus coverage.
Visibility is to allow you to start with fault management, go into the performance, the historical performance of a particular network, a particular device, a particular endpoint in-context, and travel to understand flow, understand the traffic analysis based on the packet. Also, be able to relate to systems and apps, relate to the application experience, and eventually end users. That’s the goal of convergence. And be able to take that visibility, the deep level of visibility in-context, and apply it to different type of network architectures: starting with legacy network, SDN, Network Function Virtualization, SD-WAN, data centers, cloud, and IoT.
Visibility coverage based on a common platform that’s scalable, that has a lot of shared services that you can use, and more importantly, you have a common operational UI, common operational workflow. This saves you Opex, because you don’t have to train a new group of people, just because it’s an SDN or it’s a cloud part of the network. The flexible data model, so that you can customize in the future for new things that appear in your network, example IoT, and then relationship mapping. For you to get a handle on virtual and logical and elasticity of these components, you need relationship mapping. You need relationship topology.
In terms of monitoring SDN, we don’t just monitor SD-WAN. We monitor SD-WAN, SDN/NFV and Software-Defined Data Center. Our perspective is to monitor the entire pathway to the cloud, starting with the performance management base platform with fault management from the spectrum, and then adding the CA Virtual Network Assurance. Now, you’re able to monitor the legacy network, as well as new network, again across the entire pathway to the cloud, with the WAN, with the service provider network, and the data center. With a number working with an ecosystem of leading vendors, and Viptela being one key part of our ecosystem. But there are other vendors, as well as open source technologies, that we support.
So now, we’re onto the middle of our presentation. Jeremy, do you want to guide the audience on these polling questions?
Jeremy: Sure. Tim, great job so far. And we’re about 26 minutes past the hour, so a quick time check. Guys, I’m going to push this poll question out. I believe I’m doing it now, so hopefully you see it. We’re very interested in understanding where you may be in a SD-WAN adoption or implementation. So, are you currently researching? Are you currently testing any? Are you already in production, which would be awesome? Or you just don’t have any plans yet. So, just take the next 10 seconds to put in your answers there.
You know, some EMA research on the top business drivers for SD-WAN or improved app performance, optimized access to the cloud, improved network security, reducing operational costs, like Tim talked about, and rapid deployment of network devices. So I think that all makes sense to everyone in the audience. All right, so I’m going to shut down the poll. It looks like we have about 50 percent currently researching, and 50 percent in production. So, congratulations. Some currently testing, too, so really great stuff.
All right, perfect guys. And as I stop the vote, it looks like currently researching about 58 percent. Testing about 33 percent. And around 10 percent or so in production, which makes sense. It’s a new technology, and there’s going to be some time for testing and still researching, and understanding what’s out there. So, thank you guys for that quick polling and voting. Tim, I’m going to hand it back to you for the second half of the presentation. And again, we got about 15 minutes, because we want to leave some time for Q&A.
Tim: Okay, thank you very much. So Jeremy, there’s also a question that popped out, if you can help me answer that, while I continue the presentation.
Tim: He says the monitoring SD-WAN specifically – first of all, our solution can monitor your current WAN, or call it legacy WAN, and SD-WAN. So I want to make that very clear that it’s not just tailormade to monitor SD-WAN. You can monitor existing WAN. It provides you a single pane or converged monitoring, as I mentioned before, for visibility. That means you get flow. You get packet analysis. You get fault, if you choose, as well as you get coverage of course, SD-WAN, software-defined data center, legacy networks, as well as virtual CPE.
You get the ability to monitor, not just the performance of the WAN, but also the intelligence. So, we’re going to give you a new type of KPIs, to understand how well is your SD-WAN intelligence, the control-plane? How well is it making decisions? How well are the policies that you created for the different application classes performing? And also something that we do, we focus a lot on from the beginning, is with the introduction of overlay technologies, don’t forget about the underlay. In fact, whenever something breaks, more than likely it’s due to the underlay.
So, we put a lot of effort to make sure that we provide the overlay and the underlay correlation. That exists in all the SDN monitoring technologies that we provide, including SD-WAN. And then, let’s not forget about what’s the next step after SD-WAN – virtual CPE. That’s something that we can support along with SD-WAN. So, if you endeavor to move your SD-WAN monitoring to monitor a virtual CPE, we can do that for you.
I want to give you a few screenshots to show you the opening dashboards, the opening UI that you’re going to see when you look at our SD-WAN monitoring solution. Here’s the opening home page of SD-WAN monitoring. You can see some summary, statistics on top. You see a geo map. You see locations of your branch offices, and the underlay routers within, and you also see paths. You can also see the actual path that is designed for the different application classes.
If you go to the dashboard, and these are the out-of-the-box dashboards. These are the things that you’ll get immediately when you purchase the software, is that you have immediate starting points for all your monitoring. You have specific sets of monitoring dashboards for SD-WAN. But you also get monitoring dashboards for legacy networks, for software-defined data center or Cisco ACI data center, and then you also get virtual CPE monitoring, as well as any form of SDN or NFV service chaining type of deployments. So again, the concept of convergence is legacy and new. And not just one type of new, but a variety of news.
Again, this is a view of the startup dashboards. And again, you can see the summary of the different sites. For this particular tenant or particular company, the application classes, whether it’s video, voice, or mission-critical, or best effort type of classes, the actual path themselves and then the underlay routers, very important to be able to identify the overlay to the particular underlay. Because again, a lot of cause of problems could be in the underlay. You can see that from this home page view, you can see overlay performance in summary, and you can see underlay statistics already.
This is another view of overlay and underlay. You can see that in the table below, as you look at the different path, you can click on a particular site, a particular edge router that’s there, and then bring up an underlay specific page for that specific router. And then within that page, you get to understand the traditional type of data plane performance of that router on a per interface basis. But at the same time, you can also see application class data. Another example of how we combine overlay and underlay data in the same view, so that you can do quick assessments between new overlay technologies, and traditional legacy underlay technologies.
As I mentioned before, another aspect that we introduced is monitoring the intelligence. In the past, we monitor data plane. We monitor the performance of the network and how it’s doing. But now with SDN and SD-WAN, it’s important to monitor the decision-making process of these new networks. So monitoring the intelligence as a new KPI, and here’s an example of what we call a scorecard view that gives you the intelligence performance of each type of intelligence, whether it stays on a preferred path or not, how often it stays on a preferred path, and various statistics about the control-plane intelligence.
But what’s also unique is it allows you to break it down by different levels of hierarchy. So you can collapse this view, and look at different hierarchy. And this is a traditional operation UI context grouping that allows you to generate different type of dashboards, different type of reports. So, this is a good example of how we’ve taken something that we learned over the last 20 years, in terms of how enterprise and service providers like to monitor, and how they like to create different hierarchies and groups, and will apply it to SD-WAN.
We’ll give you also insight into different application class. In fact, we created what we call Context Pages. And context pages again, is an operational UI term, in a monitoring space, for detailed data on a particular object, a device or a component. And we’ve gone ahead and used that concept, and applied it to application class. So in the same way, in the past, that you’re able to go into a particular router, context page, and do really detailed analysis on why or how it’s performing, and what anticipated issues that might occur, now we’ve taken that troubleshooting, that analytical technique and applied it to application class. In this case, it’s video.
I talked a little bit about relationship mapping. This is a big part of what we provide in the SDN monitoring space, and it’s a big part of how we help you solve the virtual – how you handle the virtual and logical components that appear in your network, is through relating it to the rest of the network, the underlay network. Here’s an example of virtual CPE. This is a virtual CPE service chain with different network functions. Through relationship mapping, we are able to identify the associated underlay building blocks, the vSwitch, the host, the VMs themselves. And you’re able to see the individual context pages again of each of the building blocks.
That ends my part of the presentation. I’m going to pass it back to my colleague, Jeremy, to wrap up our presentation, and do another poll question here.
Jeremy: Thanks. Thank you, Tim. And I think we timed this perfectly. We have a few more minutes left. Guys, I’m going to put out the last of the poll questions, a pretty easy one. Just we’d be interested – it looks like there’s around 82 folks on. We’d just be interested in how many folks are current customers of ours, versus not. It should be pretty easy to answer that yes or no. And as I move forward here, and you take time to answer that poll question. If you want more information, please very easy, just visit ca.com/performancemanagement. Our contact information is as follows, on the bottom there.
We are happy to have more conversations regarding the technology we talked about today, and very easy to follow us on Twitter and LinkedIn. So please, continue the conversation with us. And we do have a couple questions from the audience. I’m not aware that I can answer them inside of the GUI. So real quick, Tim, if you want to jump in here, when I try to answer these. CA has a number of network management products, as you mentioned, which is the best for monitoring SD-WAN and Viptela technology.
Again, starting with the base products that Tim talked about, CA Performance Management, which is going to be your network performance trending analysis. With the plug-in CA Virtual Network Assurance that extends that functionality out to the new SDN world, whether it’s SD-WAN, SD data centers, NFV, etc. Anything to add to that, Tim?
Tim: No. I think that’s the minimum starting point is CA Performance Management and CA Virtual Network Assurance. Beyond that, if you want additional visibility, you can buy Flow Monitoring and Packet Analysis. These are value-added use cases that will give you additional visibility, but is not mandatory as part of the base of the product.
Jeremy: All right, we do have one more question. And Tim, you may want to answer this. The question is, CA has a lot of customers, what are they telling you about SD-WAN and Viptela? What are the trends that you’re seeing in the marketplace?
Tim: We have, as the question mentioned, we do have a lot of customers in the enterprise space. Most of the Fortune 500 enterprises all over the world are our customers, including service providers. And they are – everyone is very bullish, very excited about SD-WAN technologies. They think that the WAN has been neglected for years. That now with cloud transformation, with the need to reduce cost, everyone is excited about the availability of the solution. I think that, you know informally, I would say eight out of 10 of our customers are seriously looking at WAN transformation or SD-WAN.
So, this is definitely the right area, particularly around the whole cloud transformation. They feel that it’s strategic to the whole cloud transformation, and it needs to be done, because the WAN has been neglected for years.
Jeremy: Excellent. I do not see any more questions. Guys, thank you so much for staying on. We are going to end a few minutes early here, some recommended FutureWAN sessions still to attend. But again, thank you so much. I’m sure I speak for Tim, when I say it’s been an honor to be invited to this Summit. We’re very happy to be able to have talked to you guys today. You do have our contact information that we showed earlier. So we’re very happy to continue the conversations around monitoring and management of SD-WAN technologies. So again, thank you so much for attending, and have a wonderful afternoon and evening.