search query: @keyword bundling / total: 4
reference: 1 / 4
« previous | next »
Author:Bayat Pour, Hossein
Title:Web Content Delivery Optimization
Publication type:Master's thesis
Publication year:2016
Pages:(6) + 55      Language:   eng
Department/School:Sähkötekniikan korkeakoulu
Main subject:Networking Technology   (S3029)
Supervisor:Manner, Jukka
Instructor:Raj Moktan, Gautam
Electronic version URL: http://urn.fi/URN:NBN:fi:aalto-201608263096
Location:P1 Ark Aalto  4440   | Archive
Keywords:HTTP request minimization
HTTP compression
base64
bundling
IUC
Abstract (eng):Milliseconds matters, when they're counted.
If we consider the life of the universe into one single year, then on 31 December at 11:59:59.5 PM, "speed" was transportation's concern, and now after 500 milliseconds it is web's, and no one knows whose concern it would be in coming milliseconds, but at this very moment; this thesis proposes an optimization method, mainly for content delivery on slow connections.
The method utilizes a proxy as a middle box to fetch the content; requested by a client, from a single or multiple web servers, and bundles all of the fetched image content types that fits into the bundling policy; inside a JavaScript file in Base64 format.
This optimization method reduces the number of HTTP requests between the client and multiple web servers as a result of its proposed bundling solution, and at the same time optimizes the HTTP compression efficiency as a result of its proposed method of aggregative textual content compression.
Page loading time results of the test web pages; which were specially designed and developed to capture the optimum benefits of the proposed method; proved up to 81% faster page loading time for all connection types.
However, other tests in non-optimal situations such as webpages which use "Lazy Loading" techniques, showed just 35% to 50% benefits, that is only achievable on 2G and 3G connections (0.2 Mbps - 15 Mbps downlink) and not faster connections.
ED:2016-09-04
INSSI record number: 54308
+ add basket
« previous | next »
INSSI