直接评价节点诚信度的P2P动态信任模型

Dynamic trust model used to directly evaluate nodes' reputation under P2P environment

为解决现有P2P信任模型计算开销大、动态适应能力差、推荐可信度的动态性未得到足够重视等问题,提出了一种新的动态信任模型。通过直接交互节点的局部评价加权其推荐可信度计算节点的全局信誉值,避免了迭代过程,降低了网络开销。采用基于时间帧的方法更新节点的全局信誉值和推荐可信度,以抑制节点提供服务和推荐两方面的动态性。仿真实验结果表明,新模型较现有模型在网络开销、节点动态性推荐的抑制等方面有较大改进。

To solve such problems as heavy message overhead, poor adaptability and ignoring the dynamic characteristic of recommendation credibility in current P2P trust models, a new dynamic model is proposed, which computes global reputation value through the direct interaction peers＇ local evaluation weighted its recommendation credibility. It avoids the iterative computing process and reduces the message overhead of network. Moreover, this new model uses a time-frame-based method to update the global reputation value and recommendation credibility in order to control the dynamic characteristic of peers. Simulations and experimental results indicate this model has advantages in reducing message overhead and controlling dynamic recommendation over the existing trust models.