#packet #loss #classification #networking

packet_loss_classification

Packet loss classifiation algorithms

1 unstable release

0.1.0 Dec 31, 2019

#8 in #loss

Apache-2.0/MIT

23KB
288 lines

Packet loss classification

Crates.io docs.rs docs ci

This crate provides some popular packet loss classifiers.

Packet loss in networks happens due to congestion and wireless errors. Depending on the reason behind such errors an application might have to take different measures to enhance performance.

This crate provides five classifiers (MBiaz, Spike, ZigZag, ZBS and Trend) for packet loss classification. Each performs well under certain circumstances and it is up to the user to decide on the best fit. ZBS being a hybrid version of the first four can lead to good results across a number of network topologies based on topology estimation.

For the theory behind all algorithms the following two papers (where theory and algorithms are taken from) are highly recommended:

  • Cen, Song, Pamela C. Cosman, and Geoffrey M. Voelker. "End-to-end differentiation of congestion and wireless losses." IEEE/ACM Transactions on Networking (TON) 11.5 (2003): 703-717
  • Hsiao, Hsu-Feng, et al. "A new multimedia packet loss classification algorithm for congestion control over wired/wireless channels." Proceedings.(ICASSP'05). IEEE International Conference on Acoustics, Speech, and Signal Processing, 2005.. Vol. 2. IEEE, 2005.

No runtime deps