Optimizing the Trickle Algorithm

The Trickle Algorithm has enjoyed much popularity and widespread use as a basic network primitive ensuring low-cost data consistency in lossy networks. Trickle is shaped by the so-called short-listen problem, hence the imposition of a listen-only period. Such a period allows Trickle to robustly address the short-listen problem at the expense of increased latency. In this letter, we introduce a simple yet powerful optimization to Trickle that can dramatically decrease Trickle´s latency with virtually no additional overhead to its scalability and robustness. Extensive simulation and testbed experiments are reported here, yielding greater than a factor of 10 decrease in propagation time.