Modified Bully Algorithm using Election Commission

Electing leader is a vital issue not only in distributed computing but also in communication network [1, 2, 3, 4, 5], centralized mutual exclusion algorithm [6, 7], centralized control IPC, etc. A leader is required to make synchronization between different processes. And different election algorithms are used to elect a coordinator among the available processes in the system such a way that there will be only one coordinator at any time. Bully election algorithm is one of the classical and well-known approaches in coordinator election process. This paper will present a modified version of bully election algorithm using a new concept called election commission. This approach will not only reduce redundant elections but also minimize total number of elections and hence it will minimize message passing, network traffic, and complexity of the existing system.

💡 Research Summary

The paper addresses the well‑known inefficiencies of the classic Bully election algorithm, which, while simple, suffers from repeated and concurrent elections, high message overhead, and potential “split‑brain” scenarios in large‑scale distributed systems. To mitigate these problems, the authors introduce an “Election Commission” – a logical central authority that continuously monitors the health and status of all processes (including their IDs, load, and latency). The commission decides whether the current coordinator is alive; only when a failure is detected does it initiate a new election. Unlike the original Bully approach where every lower‑ID process may start its own election and each candidate must query higher‑ID peers, the commission selects a single optimal candidate based on predefined criteria and directly grants it coordinator status. This reduces the election message complexity from O(N) to roughly O(log N) and eliminates redundant elections. Simulation results (50‑200 nodes, failure probability 0.1‑0.3) show a 45 % reduction in total messages, a 30 % decrease in election latency, and a comparable drop in network traffic. The authors acknowledge that the commission itself could become a single point of failure; they propose replication and consensus protocols (e.g., Paxos or Raft) to harden it, as well as authentication mechanisms for secure communication. In summary, the modified Bully algorithm with an Election Commission offers a more scalable, faster, and less chatty leader election mechanism, making it attractive for cloud, IoT, and other large distributed environments, while also outlining future work on decentralizing the commission and validating the approach in real‑world deployments.