Revisiting the swarm evolution: A long term perspective

Although peer-to-peer system is scalable for content distribution, its lifespan is shorter than traditional systems. More and more researchers pay close attention to swarm lifespan and propose specialized methods to extend it. Bundling technique and SRE (Share Ratio Enforcement) mechanism in Private Tracker system are two successful enhancements which are proved by practical systems. However, they have side effects to system users. One of the reason is that the enhancements do not take effect on the key factors. So there is a question, what factors influence lifespan? A model which can answer this question is needed to help system designers. Facing the challenges, we measure a popular BitTorrent system with a macro view which is different from previous works. Based on our measurement, we model the long term evolution to find out factors that influence swarm lifespan. In order to unveil the trend of a swarm, we propose a half-life based metric. According to the metric, our model fits the real swarms more closely. By the help of our model, we analyze the strengths and shortcomings of bundling technique and SRE incentives.