Anuta Networks Customers’ Interview: Neustar successfully deploys ATOM
One of Anuta Networks customers, Neustar, Inc. is a leader in authoritative identity resolution, helping clients grow and guard their business by connecting people, places and things with world-class solutions in marketing, risk, security, communications, and registry services. The company is trusted by the world’s greatest brands to make critical decisions roughly 20 billion times a day.
Anuta ATOM is deployed in 26 data centers worldwide scrubbing 175 Gbps of attack traffic per day for 11,000 clients. The deployment automates best of breed network infrastructure from Arista, Citrix, Juniper, and Arbor Networks. Anuta ATOM helped deliver a client self-service portal for on-demand DDoS mitigation. The solution assures network services including BGP Flow Spec and GRE tunnel configurations in a multi-tenant infrastructure.
Listen to Mathew Wilson from Neustar, Inc. as he shares his real-world deployment of Anuta ATOM analytics and closed-loop automation software for their multi-vendor Data Centers. Network Collective hosts Jordan Martin and Eyvonne Sharp discuss the merits of a vendor-neutral automation and analytics platform that scales horizontally using microservices.
[00:00:00] Hello and welcome to this special edition of the network collective short take. Today’s short take episode is being sponsored by Anuta networks. Anuta has a platform called ATOM which stands for assurance telemetry and orchestration for multi-vendor networks. But rather than coming on to tell you about the value of the system themselves they’ve invited one of their customers to tell you about their real-world experience using ATOM and how it has helped them improve network operations.
Anuta Networks Customers’ Introduction
[00:00:40] So Matt, to get us started. Why don’t you tell us a little bit about yourself and what Neustar does?
[00:00:48] Matt: Sure. My name is Matt Wilson. I am director of network engineering at Neustar kind of give you an idea of my background. I’ve been doing DDoS protection services for about 15 years now. I’m on my fourth different company building up the network and then kind of getting it set up from an operational perspective. So Neustar brought me in. Neustar is a company that helps to create, grow, guard and guide their businesses like tying together understanding and intelligence around people places and things. So we offer a number of different things in addition to DDoS protection services, things like marketing services, rest services, registry services. And my particular area is the digital performance and defense.
[00:01:36] Interviewer: Okay. So speaking specifically about the project that led you to ATOM what problems were you trying to solve that ATOM ends up being the solution for?.
[00:01:55] Matt: Two years ago or so we started undertaking a project to completely overhaul our entire network, right. So going from four nodes to over 14 nodes going from just a couple terabytes of bandwidth over to Ten terabytes of bandwidth, multi-vendor environments things like that. We kind of very quickly identify that this was going to pose some operational scale issues. Right. So the idea that I now had to manage a lot different devices deployments concurrently around the world simultaneously, 14 different locations and more across a multi-vendor environment. This was a problem that we knew we had to look into. And so we looked at a number of different things and ended up selecting Anuta
[00:02:38] Interviewer: When you were looking at those did you consider any alternatives? I mean like who else was on the list that you were looking at?
[00:02:44] Matt: Yes we looked at a number of different ones. We primarily looked at either kind of a roll your own type of solution; whether that was using or building something from scratch or using something like ansible. We looked at that we also looked at kind of vendor specific solutions. In the end none of those really kind of seemed to fit the bill for a number of reasons. So first off we’re not a huge dev work right. We don’t have a whole lot of developers. That’s not really our forte is building those kinds of platforms like that.
[00:03:20] Matt: And most of what existed were things that simply pushed configs they had no awareness of what was actually out in the network. So they were purely just automation type of tools which definitely have their place. They can be very valuable. But what we were really looking for was more orchestration. So we were looking for something that can be aware of what’s in the network, identify whether or not changes need to be made and allow us to truly properly orchestrate and services that we deploy in our network.
Interviewer: That’s a really interesting component. Because I think when we hear about orchestration systems I think we always hear about analytics, telemetry and orchestrations as different systems or different components but you’re lumping them all together.
[00:04:06] Matt: Yeah absolutely.
Benefits of Anuta ATOM Platform
Interviewer: So what does Anuta do there? I mean what is it about, what analytics, what about that system that is driving that they combined. those stuffs?
[00:04:15] Matt: Yeah I mean one of the things I really liked about it was the fact that in the same platform I can tie together telemetry from the devices, I can tie together alerting from the devices, I can tie together SNMP from all of my devices and within the same system without having to do a lot of dev effort to create these business rules to tie all of these pieces together, which was basically when you looked at it, it was about as bad as basically just building the whole thing from scratch. So within the same system, I can take all this telemetry I can. It allows us to focus on our business rules right, the things that we want to do in the network when something happens. That’s one of the big big benefits to us is that it allows us to focus on what’s truly important to us without having to go off and build all the just what I would consider to block and tackle.
[00:05:12] Interviewer: That makes it’s very cool.
[00:05:15] Female Speaker: Yeah yeah. And so you guys specialize and indeed often attacked data from the Internet. So scale we’ve already heard several terabytes of bandwidth. Can you talk a little bit about ATOM’s ability to scale with the needs in your infrastructure?
[00:05:32] Matt: Yeah so that’s one of the nice new things that ATOM platform does for our site protect NG which is what we call our DDoS protection service. It allows us to take all of this data and it can grow with us because it meshes well with our cloud initiatives of taking a lot of these things building it into our cloud platforms and doing that in cloud services. It allows us to scale that. It scales great horizontally so we were able to start small as we are deploying we could start small we could have at polling data from a specific number of devices and as we’re growing, as we’re adding more and as like it’s; as we go from 14 and much much more. It allows for that platform to scale really well horizontally.
[00:06:23] Interviewer: So you mentioned a couple of very interesting things there. I want to expand on them for a second. You mentioned cloud platforms. So are we talking hybrid cloud here?
So is this you know on prem as well as out in public cloud offerings? So you can actually start if you have a well-established architecture that you know you kind of have what your hybrid cloud you’re running. You’re running a cloud platform within your own data centers you can start there. Really anything that’s Kubernettes based this ATOM platform can run on. So whether you want to do it there, you want to do it in a Google cloud, you want to do it in AWS. You can pretty much deploy anywhere you want.
[00:07:06] Interviewer: The other thing you’ve mentioned I think you hinted at as well the fact that as Containers scales horizontally. So I mean is this like micro-services platform is that what it’s like?
[00:07:16] Matt: Yeah that’s exactly what it does. It is it’s all based on microservices. So individual pieces can run active/active so you can scale it as much as your business requires. So obviously in the DDoS world and we’ve got customers that literally a couple of minutes of downtime can be hundreds of thousands of dollars in revenue lost. So for us having something that this platform can manage our devices, manage our network and be completely available with full redundancy. Active, active, I can independently update things I don’t have to take downtime for that because ATOM is micro-services based because it can be cloud-based or on-prem based. It really let us scale very well to meet our needs.
[00:08:05] Interviewer: That sounds really powerful. So what would you say are some of the best results? I mean we’ve talked about you know some of the technical features and some of the things that are of value but I mean to the business, what was the impact, what was the impact to your operations?
[00:08:18] Matt: Yes so one of the big things we’re able to do is to start cutting down on just human error right. So by doing this by making everything kind of programmatic, we cut down on a fair amount of human error by doing that. One of the other things we did is just the simple amount of time it took to deploy customers to make changes in the network came down quite drastically. We were able to take things that, if people were doing it by hand it might be you know a good hour worth of work and we can cut it down to about five minutes.
Interviewer: And is that because you were templatizing configurations and that type of thing? Like what was, what would allow you to reduce that?
[00:08:58] Matt: Yeah that was basically we’re able to do that because, before this, you either had to have a series of scripts and hit each individual device. And then you had them required the human to be the one that would sort of transitioning data from script to script and making sure double checking you know ensuring that everything went as exactly as you as you expected it to and hope that there weren’t any errors and you had to roll back. Anuta has basically allowed us to do that all in one single atomic operation. Right. So through their API, through our service models we’ve written and so we put in the data, we can say execute go and it goes out and does it across all the different devices across regions. And then if there’s ever a problem we can roll back atomically as well. So and I can do that even between execution. So because it’s kind of a service-based; so I deploy a service, that service might involve something a customer and a handpiece because I can do that. I don’t have to just rollback device configs. I’m rolling back a service execution.
ATOM helps with Troubleshooting
[00:10:11]Interviewer: What about troubleshooting? Has it helped you reduce the meantime to recovery or anything like that?
[00:10:19] Matt: Yes. So what we’ve been able to do. One of the big focuses and this is kind of a never-ending effort is we’re identifying things that are very common troubleshooting problems. So if I want to run say… if I want to go and run and execute something on all 14 devices at exactly the same time. I can do that and I can bring that data back and get a specific snapshot of time and bring that back. So now the operator can look at it they can identify what looks like maybe there’s a problem and then go execute. And then the other thing we’re focusing on is the information gathering, having a human look at it and make a decision about what might be a problem and then going in and executing a response right.
[00:11:06] Matt: So all the very common types of things. That’s all stuff that we’re building into the platform from a network operations perspective. And I talked before about doing it from more customers perspective, a provisioning perspective, from an operations perspective we’re able to bring down some of our meantime the resolution quite a bit by doing this.
Next Steps with Anuta ATOM
[00:11:28] Female Speaker: Very cool. And as you look at what you’ve done with the platform now, what’s next, what are you planning to do with ATOM. that maybe you’re not doing today?
[00:11:37] Matt: Yeah. So I think we’re going to start tying it more into our… In our initial phase, we didn’t tie it into our portal for like automated deployments yet. We wanted to give our operators a little more hands-on experience with it, make sure that we work out any kinks things like that. Meanwhile, you know it’s kind of our next step is to really do much more integration there and to be honest this is a never-ending thing. The one thing I love about this platform is it can literally do anything you want it to do. If you have a device or a system or network that has an API or an CLI or any way in which you can log into that thing and an issue commands. Anuta can interact with it. So you know for us I mean that can be things like if you have a network performance monitoring platform that you use. Anuta can tie into that. So we can do some really interesting things like if a mitigation starts I want to go create a service check so that I can double check what the performance of our network is through the course of that mitigation. So we can get really creative with things like that.
[00:12:50] Interviewer: That’s interesting. So I mean you’re now in, you’re using this tool. If you were to talk with someone who is you know starting today with deploying Anuta, what are some of the things you learned along the way? Maybe some of the lessons learned or tips or tricks for people who are getting started?
[00:13:06] Matt: Yeah so probably my biggest recommendation here is if you think this is going to be a plug and play type of vendor’s solution it’s not. But it’s also not intended to be right. It is…I like it because it kind of bridges the gap between being a one size fits all plug and play sort of thing that may not get you all the features you want and all the capabilities you want. But at the same time, it’s also not this late super ethereal thing. I can put it out there. I can get a lot of value out of it day one. But you can dream big with it right. You’ve got to think big with it. You can start small but really there’s no end to what you can do to kind of go back to Yvonne’s one of Yvonne’s questions. The telemetry stuff to us is one of the things that we’re really going to build out right. We’re going to really expand on that. I think there’s a huge amount of power for a lot of network operators there. The idea that I can get telemetry out of all my devices and I can automatically execute commands and multi-vendor commands in my network. You know based on things that are happening you have data that’s coming out of my network is really really powerful. I think that’s where we’re going to go.
[00:14:27] Matt: I think that’s definitely things that people ought to be looking at.
Interviewer: Sounds fantastic. Matt thanks for coming on to tell us about Anuta today. There was a lot of great information. If you’re listening to this and you want to find out more about how Anuta could help you with network assurance and orchestration, you can head on over to anutanetworks.com/collective for lots of great information, White Papers, case studies and to even demo the platform for yourself.
[00:14:53] So thanks for watching and we’ll see you next time.