Internet Myth #10: Putting servers close to end users solves performance problems

Team Teridion

Reality: Server proximity can help to some degree, but can be both costly and complex–reducing business agility.

You’ve worked hard to develop a cutting-edge application, but sadly, some of your users are experiencing compromised performance. The first thing that comes to mind is building or renting a closer POP to serve users from within their region–thereby avoiding a perilous journey over a long stretch of the internet. But will server proximity really solve the problem?

While the cost of new infrastructure is clear, something that may be underestimated is the complexity added when replicating data centers or POPs. This includes:

  • Code modifications required to support multiple application instances
  • Managing and synchronizing duplicated content between multiple data centers
  • More resources to cost-justify, plan, execute, and maintain all of this
  • Loss of business agility from all of these compromises

 

As anyone who has driven in rush-hour traffic or to a solar eclipse knows, congestion can be a killer. Is New Jersey logically closer to NY than London? And what is the right way to measure “closeness”? Physical distance? Ping? RTT (or latency)? So even after putting in regional POPs, the last mile can still be a problem for two reasons:

  • Congestion: Usually, the metric that determines data center location is RTT to a major user base (usually a large city). However, low latency does not translate to high throughput, and users might still see bad performance in congested routes.
  • BGP: As discussed in Myth #9, users still might be routed in loops, or even to different POPs due to BGP convergence. So even if your New Jersey data center was intended to better serve NY users, traffic could still be routed to London or some other location.

 

There is an infrastructure-free alternative that allows you to focus on your core business. Teridion offers on-demand fast lanes through the Internet, accelerating global user traffic to your data center an average of 10X to provide a like-local experience. With a performance overlay service built on 15+ top public cloud networks, your application traffic can bypass Internet Backbone problems to your main data center without dependency on any single network provider.

If you are a SaaS provider and want to learn more about how to optimize your end-to-end Internet throughput, here’s how to get started:

Book a Demo

    Interested in (please select all that apply):

    Do you require:



    Interested in (please select all that apply):
    Do you require: