Leader Election is one of the fundamental problems in distributed systems. A leader is a correct process that can be used to coordinate the work of a set of processes. An algorithm has to implement two properties to solve the Leader Election problem: 1. Safety: At any point in time there is at most one leader. 2. Liveness: Eventually there will be one leader forever. We showed in our previous work how to solve a weaker form of Leader Election in an asynchronous system with local clocks [1]. But for electing a leader the system needs to stabilize: a leader is only guaranteed to be elected during periods in which the system behaves synchronously. In this work we show that the stabilization property is not necessary for the Leader Election problem. We do this by examine the ability to solve Leader Election in the FAR model [2]. The FAR model does neither assume the existence of an upper bound on the communication or computation delays nor that the system stabilizes. Instead it assumes that the system is in a certain balance: computation is not infinitely fast, the communication subsystem has rudimentary congestion control and the average response time between two correct processes is finite. Our contribution is twofold: (1) we show that Leader Election is not solvable in the pure FAR model and (2) that it becomes solvable with local clocks with a bounded drift rate. Especially interesting is the implementation of the liveness property: we reuse an internal timeout originally introduced to provide liveness to a failure detector implementation.
展开▼
