Evaluating Traffic at IXes and Data Centers with PeeringDB Filters in Kentik

Peering managers, network planners, and edge strategy experts are always looking for ways to improve their network's performance, increase reliability, and reduce connectivity costs. But getting the data to support strategic decisions like peering partnerships can be a difficult task. In this video, I'm gonna show you how you can use PeeringDB filters in Kentik to help evaluate traffic between your network and networks that are present at internet exchanges and data centers worldwide. It's gonna help you answer questions like what internet exchanges and data centers should I be at based on my network's traffic profile. Am I optimizing the use of my existing exchange ports? Are there members at that exchange I could potentially peer with but am not currently. And how does this all impact our bottom line? So let's hop into Kentik and get started. So in this first example, we're going to be an access network that is already a member at AMS-IX in Amsterdam. And I wanna make sure that we're optimizing our port at this Internet exchange. So what I'm looking to uncover really is if there are any other possible candidates that I could peer with at AMS-IX. So I wanna see how much traffic I have on my transit connection to other members at AMS-IX. So within Kentik, I am in the data explorer here, and you can see we have our sidebar here for querying And I'm gonna go down to the filters to tell Kentik, what traffic I wanna look at. So since most of my traffic is coming inbound. I wanna filter to see external traffic coming from source networks. So I'm gonna set the network boundary to external. And then I really only wanna see transit traffic. So I'm gonna make sure that we filter to only see transit connectivity type. And then this is where I'm going to be using those PeeringDB source filters to only show traffic from ASNs that are members at AMS-IX. So I'm gonna go in here and search for our new PeeringDB set of filters I'm gonna select ASN is member of IX. And then I'm going to select AMS-IX. And then I also wanna set up some filters, to only look at members who are gonna be potentially open to pairing with me, So I wanna filter by the peering policy, making sure that I'm only looking at open. And selective peering policies. So let's run that. These results are pretty interesting. It looks like we have a healthy amount of, transit traffic here that could be potentially offloaded to peering at AMS-IX. So now I wanna look at, you know, who are the members that are sending us this traffic. Right? Now I'm gonna go into the dimensions and select which dimensions that I wanna see. So I wanna pull in connectivity type just to make sure that it's pulling as transit. I wanna look at the provider, the transit provider. And I wanna look at the AS number. Yeah. Rearrange those a bit. Click save. And then I'm also gonna adjust the visualization for this to a sankey. Great. So now you can see in in this chart here, we can see all of the source networks that are sending transit traffic through looks like a couple of main providers here, Verizon AT&T, Cogent. And so this this really gives us a nice visualization to see where this transit traffic is coming from. There may be reasons, right, for, traffic to be transit. But in this scenario, we're gonna assume that we can move all of this over to peering at AMS-IX. So in order to, figure out who we would contact at those source networks to set up peering, You could easily go here, select PeeringDB record, and this is gonna pull up the integrated PeeringDB data within Kentik for that, for that source network. Then we can also, if we wanna look at how much this could potentially save us and cost savings, we could hop over to the Connectivity Cost portion of the Kentik portal and look up the current price that we're paying right now for transit traffic on AT&T, Verizon, etcetera. And we've got our average, our ninety fifth, and our max metrics here. So everything that we need to do is here, for it to to conduct those easy calculations. So to make sure that I can quickly come back to the same query without having to rebuild it all over again, I am gonna create this as a saved view. Give it a simple name. AMS-IX IX. And then I'm gonna make it private and save. And so now anytime if I wanna revisit this this query, change it up a little bit, it is now stored in my library under saved views. Alright. Let's try another example. Imagine where network that wants to build a pop in the Ashburn area, And we wanna figure out which internet exchange it would make most sense for us to join. So for this example, we're gonna look at three exchanges in the Ashburn area. Linx NoVA, Equinix Ashburn, and Digital Realty. So first, I'm gonna select line chart. For my visual, as that's gonna give us the best comparison for this example. And then I'm gonna hop down to those filters again. Tell Kentik what traffic I'm interested in seeing. So I'm gonna set the network boundary to external. Again, I wanna make sure I'm only looking at transit traffic. So I'm going to set the connectivity type to transit. And then I'm gonna have those same filters for the peering policy, making sure that we're only looking at networks that have open. Or selective peering policies. Click save. And then now we're gonna go into the dimensions area, and we're gonna use the filter based dimensions. And this is really gonna enable us to have a comparison and see those three comparisons in one line chart for each IX. So I'm gonna give this a simple name IX comparison Ashburn. And then here, we're gonna create essentially a series of filter dimensions based off of PeeringDB filters for each IX. So I'm gonna start out with Linx NoVA, and I'm gonna use those same PeeringDB filters to select Linx NoVA. I'm gonna create one for each IX that we are interested in evaluating. Then last one, Digital Realty. Perfect. So now we're gonna click save. And so what these results are showing here is the total amount of transit traffic that could be offloaded to each of those IXs if we were to join them. So based on the results, we can see that there's, a clear winner here with Equinix Ashburn, based on the traffic standpoint. Right? We can see the most traffic. Highest amount of traffic could be offloaded to peering there. Of course, there are other costs and other factors to consider as well, but this gives us a pretty good jumping off point to further investigate. So from here, if I wanted to just drill down into Equinix Ashburn and, you know, see what, those potential providers, and transit networks would be, I can go to the previous, view that I set up for AMS-IX. So I'm going to my saved views, open up that previously saved AMS-IX query, and all I would have to do is adjust this query. In the filtering here. Instead of looking at AMS-IX, now we wanna look at Equinix. Click save. And now we're off to the races. We've got our our source networks. We've got our transit providers. We've got all the data we need to investigate further. And potentially see, how much we could peer off it at Equinix Ashburn. So in this demo, we took two scenarios that have been traditionally a complex and challenging task for networking and peering teams to solve, and we've reduced that data gathering and correlation time down to minutes with Kentik. You can try Kentik's peering features for yourself by signing up for a free trial on our website or requesting a personalized demo for your team.