Development and Implementation of B-Tracker on a BitTorrent Client

State: completed by Andri Lareida

Trackers are used in peer-to-peer (P2P) networks for provider discovery, that is, mapping resources to potential providers. B-Trackeroffers a fully-distributed, pull-based tracker, by extending DHT functionality by distributing the tracker load among all providers in a swarm. Bloom filters are used to avoid redundant mappings to be transmitted. This results in the important properties of load balancing and scalability, while adding the ability for peers to fetch new mappings instantly.

While the B-Tracker approach has been described and validated through simulations, the opportunity of further new results and insights shall be taken in the following ways:

  1. Motivation: use real BitTorrent data (or run a series of KAD nodes) and show that load balancing and efficiency are indeed a problem.
  2. Approach: establish theoretical bounds and complexity of B-Tracker operations.
  3. Simulation: investigate the impact of changing some constants used on the paper (replication factors etc) to better justify chosen values.
  4. Experiments: run experiments using a BitTorrent simulator.
  5. Optional alternative approach: look for a solution based on unstructured P2P, since those probably show even better load balance.
  6. Optional improvement: investigate locality-awareness of secondary trackers to reduce inter-domain traffic.
70% Design, 30% Implementation
Java programming knowledge, knowledge of DHTs, previous participation on P2P lecture a plus

Supervisors: Prof. Dr. Burkhard Stiller, Dr. Thomas Bocek

back to the main page