Humble Avalanche Replay of HTTP Archives

(Harhar, for short)

tl;dr: Check the Installation, Usage and Sample pages.

Running performance tests is important and if you're reading this document it means you know that. What is also important is to run realistic performance tests. RFC 3511 is good, and pre-defined protocol mixes are good too. However there are cases where you want to simply, precisely reproduce a real website, web service, or anything HTTP based (like Adaptive Bitrate or SOAP services). There are many reasons why you might want to model one or more of those. What is important is having the ability to do it.

This is what Harhar enables you to do. Harhar stands for “Humble Avalanche Replay of HTTP Archives.” A HTTP archive is, simply put, a JavaScript Object Notation (JSON) file formatted in a pre-defined fashion. It's actually a W3C standard. The specification is available here: https://dvcs.w3.org/hg/webperf/raw-file/tip/specs/HAR/Overview.html

A pre-requisite to using this tool is to have good knowledge of Spirent Avalanche.

Harhar is a program that will take one or more HTTP Archive (.har) files, read them, rebuild the requests and responses that were capture and convert that to an Avalanche Action List. In short: you record yourself accessing a website, automatically save all the pages, images and other resources you accessed, and convert that in a way that make it easy to replay in Avalanche – at a thousand fold load.