LIRNEasia is a regional ICT policy and regulation think tank active across the Asia Pacific

REQUEST FOR PROPOSAL: We are inviting Proposals from potential Bidders to conduct a nationally representative study of ICT access and use in Nepal with special focus on the disabled. Click here for more information. Deadline for submissions: 29 January 2018.

BB QOSE: High download speed does not mean better latency

Contrary to popular belief, higher bandwidth and high download speeds does not necessarily mean better performance in terms of latency or round trip time (RTT). RTT is the measure of how long a packet takes to reach the destination server and return to the client (i.e. the PC or laptop or any device used to make the request). So the less time it takes the faster the web page will load (in other words, the lower the RTT the better). Higher bandwidth and high download speed is important for downloading large chunks of content such as video, music etc. However, when it comes to web browsing (where there are many short connections between the client and the server), RTT is where the bottleneck lies.

LIRNEasia’s research on broadband quality of service experience (BB QoSE) proves that higher download speeds does not translate to lower RTT (which is what is needed for web pages to load fast). The most recent diagnostics report average download speeds of over 5 Mbps for fixed broadband in Colombo (SLT and Dialog LTE) and an average of approximately 1.1 Mbps – 3 Mbps for mobile broadband (Etisalat, Dialog and Mobitel), in the International domain (the download of a 10 MB file from a server located in the US). The corresponding RTT measures over 500 ms for both fixed and mobile broadband.

The location of the server also matters. The value of testing in multiple domains as the LIRNEasia methodology suggests, is that the variation when accessing a locally hosted server vs. an internationally hosted server can be analysed. In most cases there is understandably better QoSE when accessing content from a locally hosted server (this article explains the physics element further). However, there are exceptions; for instance when service providers use HTTP proxies to cache certain web sites, the results can be reversed.

Diagnostics were carried out in February and March 2014  for both fixed and mobile broadband (via USB dongle) connections in the following cities: Bangalore, Chennai, Delhi, Dhaka, Colombo, Male, Karachchi, Kathmandu, Thimphu, Bangkok, Manilla and Jakarta. The report can be accessed here.



Research Mailing List

Enter your email for research updates:


Flickr Photos