Latency = Control

NL-ix values low latency, and one of the big drivers behind that is Jan Hoogeboom, CEO of NL-ix. In this article, Jan explains his big drive for latency and why latency = control.

Jan: People sometimes ask me why I go on about latency. Does it really matter? Aren’t there more critical factors, like security or stability? I normally reply that in a world of distributed apps, if you don’t know your latency you are no longer in control of any of these factors.

The App Explosion
Look at the number of apps most organisations now use. In the hundreds and growing fast. The rapid proliferation of apps creates major headaches. As everyone knows, it takes just a few weeks for everyone to move to a new app, then, in no time at all, the business is completely dependent on it. These apps are increasingly run in the cloud, with 50-100 interconnections to the outside world. Even if you operate in just one market you will be running apps from Paris, or Dublin. And the more apps that run over existing apps – think partnerships and security – the more and more data-hungry, and slower, they get.

The Usual Suspects
So how can you control performance and delivery across a fast-moving multi-point business? Let’s look at the usual interconnection suspects.

You should have a Cloud Connect link with your IaaS workload. This works well, but is limited as most apps are not “Cloud-Connectable,” living either in the public cloud or at a small-scale local level. Your leased lines can’t scale up to 50/100 people. Your business is spread over many markets and in many locations which even the largest data center provider cannot reach. IP Transit? Unfortunately, IP Transit is often a black box which is like running your business blindfold – you don’t know the efficiency or otherwise of the routes you are paying for. And on the peering side nearly all Internet Exchanges provide an island of dense interconnection rather than a geo-dispersed spread.

Taking Control
So how do you establish that control of performance, redundancy, security or cost? You need control of the geographical application paths, the separation of traffic flows, and the ability to monitor for issues. And the only tangible indicator that your network has capacity and that you are in control is latency. It makes your network visible which allows you to manage, plan and troubleshoot.

Latency also delivers stability. Let’s say your published latency between Amsterdam and Frankfurt is 6 milliseconds. Maybe one or two milliseconds don’t really matter, but you know that you can trust that route, and that you aren’t paying for something that could take 200 milliseconds.

Peace of Mind
This is why at NL-ix we place such a premium on latency – and why I go on about it! It’s the reason we invest in new lower latency primary and backup routes, and share point-to-point latency with our customers in real time.

Building a faster network and sharing data and access to it means we can offer our customers a unique level of performance, control and peace of mind as they operate in a world dominated by distributed apps.