Proven techniques to measure and reduce network latency

Network latency is the term used to refer to the length of time required for a visitor to a website to be able to be connected to its web server, for the processing of their request and for data to begin to be sent by the server.

There are a number of things that can have an impact on latency time, including the likes of server performance and round-trips.

There is a definite link between the performance metrics of a server such as available RAM, used hardware including HDD/SDD drives and server speed, and the latency of a website.

A round-trip is the term given to the journey that an object request such as stylesheets, script files and HTML files takes to get to a web server and back again, with distance being the main cause of the time taken in addition to various intermediate points that the connection has to travel through.

High network latency can have a dramatic increase on the load times of web pages, resulting in interruptions to streaming audio and video and sometimes making applications completely unusable. Even a small latency increase can negatively impact some applications enough to result in the web page being abandoned by impatient users.

But what can be done to reduce network latency?

How to Measure Network Latency?

The first step to improving network latency is to make an accurate measurement of the length of time it takes for network communication to get from its point of origin to being fulfilled. There are a variety of tools that can be made use of by network managers to accomplish this, including the likes of Traceroute, My traceroute (MTR) and Ping.

Ping is able to record the round-trip time a packet takes to get to its final destination from the original source. Traceroute, on the other hand, is able to visualise the path taken by packets over IP networks and record the latency time between each of the route’s hosts. Elements of both of these are blended by MTR in order to find the delay times between devices that are on that path in addition to the final transit times.

Reducing Network Latency with CDNs

CDNs such as Imperva can be made use of as a way to reduce the latency of a website as well as to improve UX and the overall performance of the site. There are a number of methods through which this can be accomplished.

One such method is connection optimization. CDNs are able to optimise the connections between origin servers and users via network peering, TCP pre-pooling and the reuse of sessions. Premium CDNs are able to further increase the speed of communication with the use of traffic routing via a tier 1 network backbone with the lowest possible number of hops.

Another method is content caching. CDNs are able to cache and compress the mirror versions of web pages and then store them in strategically located data centres. Users can then be delivered the content depending on their location, cutting down on latency and round-trip times.

By reducing latency and improving network speed, business processes will also be able to make giant leaps in terms of high performance and efficiency.