Differences
This shows you the differences between two versions of the page.
— |
education:workloadandchurngenerator [2007/08/21 14:32] (current) |
||
---|---|---|---|
Line 1: | Line 1: | ||
+ | ====== Event Workload and Churn Generators ====== | ||
+ | **Semester project** | ||
+ | |||
+ | The goal of the project is to implement two generators for an arbritrary network size //n//: | ||
+ | - a churn generator based on churn information that will be retrieved from a (very large) database, from an existing program and from online data (Gnutella/Napster [[http://www.cs.washington.edu/homes/gribble/papers/mmcn.pdf|traces]]). Churn is a measure of the numbers of peers that join and leave the system. | ||
+ | - an event workload generator that will generate //n// (network size) event workload depending on general well known distributions (e.g., we want that the overall distribution of all peers follows a gaussian law, how many events should peer //i// send?). Distributions (e.g., Uniform, Zipf,...) will be applied to peers (e.g., some peer will send more events than others) and to event //types// or content (e.g., there will be more events of type ''soccer'' than of type ''golf'') | ||
+ | |||
+ | **Language:** Java | ||
+ | |||
+ | **Responsible:** [[http://lpd.epfl.ch/mmonod|Maxime Monod]] | ||
+ | |||
+ | **Project intended for 1 student** | ||
+ | |||
+ | Le superviseur parle aussi français... |