What is latency?
Latency is a measurement in milliseconds of the time it takes for a data packet to be transferred from one point of a network to another. The main factor in how much latency will be present is the distance in which that data has to travel. For example, it would be expected that data requested from a local Network Attached Storage (NAS) device would have a much lower latency than accessing a website hosted on the other side of the planet. There are many other variables that can impact how much latency is experienced on a network or over an internet connection, but users can expect an average latency of between 8-40ms for a fibre internet connection, 14-90ms for an ADSL internet connection, 50-180ms for 4G and 650-850ms for satellite connectivity.
Fixed satellite services use satellites positioned in geostationary orbit (GEO) at approximately 36,000 kilometres above the equator, so the main reason for the higher latency experienced with this methodology is the comparatively large distance that data must travel. For example, any web request will leave the customers location, travel up to the satellite, back down to an earth station, out to the internet to obtain the web page and then back over the same path to complete the exchange. To put the distances the data needs to travel into perspective, the Earth has a circumference of just over 40,000 kilometres so the web request referenced will have had a journey equivalent to travelling around the world almost two times just to reach its destination.
Terrestrial internet connections such as fibre and ADSL have lower latency in general as the network and infrastructure is wired with defined distances and cable specifications, whereas communication methods such as 4G or satellite will also contend with interference from poor weather that can have a negative impact.
WHAT IS THE IMPACT OF LATENCY?
It is common for users to confuse service issues caused by latency with those of slow network performance. Latency does not impact all applications, but it can have a significant impact on what are classed as “real time” applications. Real time applications require the exchange of data as quickly as possible and so include most of the popular broadband applications ie browsing, video conferencing, Voice over IP (VoIP) and most critically, online gaming.
A familiar experience of latency can be seen with the delay in response during live conversations between an on the scene television reporter and an in-studio colleague. These exchanges often utilise satellite communication due to its non-reliance on local telecommunication infrastructure though the inherent latency in the method is evident.
CAN LATENCY BE REDUCED?
In relation to satellite-based services perhaps the most obvious solution is to bring the satellite closer to the customer; this can be seen in the range of Low Earth Orbit (LEO) services coming to market. LEO satellites are positioned below an altitude of 2000 kilometres and currently exhibit a latency as low as 43ms.
Satellite broadband technologies use a standard selection of methods to assist with reducing the impact of latency, many of which have also been adopted by terrestrial broadband service suppliers to further enhance their customers user experience. One method particular to satellite broadband, attempts to address an operational condition that exists with TCP/IP (Transmission Control Protocol /Internet Protocol), this being the suite of communication protocols utilised with internet access. TCP/IP requires communication between two computers to occur before a data stream can commence and this acknowledgement will accentuate any latency present. These acknowledgements are resolved on the satellite modem with a technique known as Web Acceleration or TCP spoofing, eliminating the need for a device to wait for a response and enabling a smooth transfer of data.
A further method employed is where the satellite gateway (the interface between the satellite network and the internet) attempts to predict the next content that will be accessed by ‘prefetching’ the web pages that have links on the content presently being viewed. Removing the step to go out to the internet again to retrieve that information speeds up the process of delivery with the obvious enhancement of the customer experience.
Data caching sees frequently used information being stored locally to speed up response times for activities such as web requests, a method also deployed by browsers on computers and smart devices in general. This is typically small pieces of information such as that associated with Domain Name Server (DNS) resolution, though this can extend to web pages and media content with the use of web caches.
DOES LATENCY LIMIT WHAT I CAN DO?
Only with applications that have critical response times. Online games that require low latency such as first-person shooters or racing games will be unplayable with GEO satellite internet as they require latency no higher than 80-90ms; in some cases, the games themselves will stop you from playing to ensure fair play. Turn based games are typically unimpacted by latency but you should always refer to any internet requirements provided by the game makers.
Real time applications that have non-critical response times (such as remote desktop software) will evidence latency with delays in actions occurring, though should remain perfectly functional.