Skip to main content

Latency: Definition, measurement and testing

Latency
(Image credit: Getty Images)

Latency is the technical word that describes how long it takes data to get from one place to another. You can measure it with a ping. Your computer sends a small packet of data to a server, the server sends it back again, and you time how long it takes. 

Latency varies depending on three things: how fast data can physically travel through the network, which route it takes, and whether it has to queue, according to the computer hardware manufacturer Apposite Technologies (opens in new tab).

Measuring latency

Measured in milliseconds, latency is often expressed in terms of its "round trip time" (RTT), according to Frontier (opens in new tab).  RTT is the time taken for a data packet to get from one network destination to another. An alternative and less common phrase used to express latency is "time to first byte" (abbreviated to TTFB). TTFB refers to the time passed between the first part of a data packet leaving a point in the network and reaching its destination. 

Related: Quantum computing: Definition, facts & uses

Speed through the network is a significant problem for satellite internet. Most communication satellites are in geostationary orbit, 22,300 miles (35,900 km) above the Earth, according to Space.com (opens in new tab). To get from your computer to a server and back again, data has to make that long trip four times. 

SpaceX founder and CEO Elon Musk (opens in new tab) announced a concept called Starlink in January 2015, explaining that the company intended to launch only about 4,000 broadband satellites into low-Earth orbit to provide low–cost internet. For perspective, there are only about 2,000 operational satellites in orbit today, and humanity has launched only around 9,000 craft into space in all of history.

In relation to latency, Starlink is aiming to greatly reduce the RTT of data packets, cutting latency right down. This should make high-speed activities, like streaming and gaming, possible almost anywhere in the world.

Reducing data speeds

The idea of a network of geographically distant computers was originally proposed in the 1960s by MIT computer researcher JCR Licklider, in his theoretical piece on real–time interactive computing, Man–Computer Symbiosis (opens in new tab)

According to Scientific American (opens in new tab), the earliest form, ARPANET, was limited to just a few nodes in the U.S., but the development of packet switching and TCP/IP protocols (internet communication languages) in the 1970s unlocked the network's potential for expansion around the world.

Related: Firewall: Definition, technology and facts

Email was being used by the 1980s, but it wasn't until Tim Berners–Lee introduced the World Wide Web in the early 1990s that the internet began to spread beyond research and government institutions. Since then, improvements in data transmission speeds have allowed individuals and organizations to store and access ever-increasing amounts of data and send big files at faster speeds from anywhere.

Currently, fast internet access is only available in places with fiber optic cables, with fiber optic internet being 20 times faster than cable internet, according to software and computer manufacturer HP (opens in new tab). In remote locations, communications satellites provide links to the internet, but the connections are notoriously slow.  

Testing latency

Ping speed

Ping speeds of 100 milliseconds and below are typical for most broadband connections.  (Image credit: Getty Images)

Anybody can carry out a ping test on their computer to test latency. All computers that are connected to a network will have a latency testing tool built in. Web infrastructure company Cloudfare (opens in new tab) also run this test on their website. This can be used to carry out a quick data speed check, without needing to install extra software. 

The process involves sending a default of 32 bytes of data to a set destination and measuring the time taken for data to be returned to the computer, according to web hosting company Ionos (opens in new tab). The time is shown in milliseconds. 

The RTT is calculated for each test carried out and the user is provided with a summary of the results. As well as data speed, any lost data packets will be reported in this summary. At the end of the test, an average time taken between the data being sent and received again is calculated. 

Additional resources

You can find out how to improve latency for activities such as gaming by reading this article from computer software company, Norton (opens in new tab). Alternatively, to learn about how 5G impacts latency, listen to this video from telecommunications company AT&T (opens in new tab)

Bibliography

Laura Mears is a biologist who left the confines of the lab for the rigours of an office desk as a keen science writer and a full-time software engineer. Laura has previously written for the magazines How It Works and T3.  Laura's main interests include science, technology and video games.

With contributions from