This paper presents a comparative study on how well known synthetic network-coordinates algorithms (namely, GNP and Vivaldi) can be used to optimize a P2P algorithm. The algorithm being optimized allows a great deal of freedom in choosing each node’s partners, thus enabling network-coordinates systems to provide their best possible optimization.
The coordinates based variants of the algorithm are also compared with an almost optimal one, as well as with a version where partners are randomly chosen. The study highlights the pros and cons of the different algorithms and gives some hints of when their usage should be considered.
|Title of host publication||CRC|
|Publisher||Instituto Superior T馗nico|
|Publication status||Published - 1 Jan 2009|
|Event||Conferência de Redes de Computadores 2009 (CRC'2009) - |
Duration: 1 Jan 2009 → …
|Conference||Conferência de Redes de Computadores 2009 (CRC'2009)|
|Period||1/01/09 → …|