To search, Click below search items.


All Published Papers Search Service


Considering The Nodes Join/Leave Behavior in The Analysis of Chord Stabilization in The Heterogeneous Network


Sri Wahjuni, Kalamullah Ramli


Vol. 11  No. 12  pp. 57-61


Since each node in the heterogeneous network has a various device and connection capabilities, the hierarchical architecture is a suitable structured P2P architecture for this kind of network, where the more stable nodes (super nodes) acts as a gateway for other less stable nodes (normal nodes). While structured peer-to-peer (P2P) has benefits of its scalability and capability in performing a successful query lookup, its stability is suffered from the dynamics of networks that form the P2P overlay. The dynamics of network, called churn, is affected by the nodes joining and leaving continuously in the network. Previous works on the analysis of impact of churn in structured P2P have not yet distinguished the impacts of nodes join events and of nodes leave events. In this paper we investigate the effect of churn to the performance of Chord protocol by observing the different impact caused by nodes join events and nodes leave events. Distinguising the impacts of nodes join events and nodes leave events is an important point in maintaining the stability of Chord overlay that implement hierarchical architecture. In the previous approach, when a super node (which is a member of Chord ring) experiences a failure, the normal nodes should perform individual rejoin. From the experiments we found that the increasing of node join events lead to a significant increasing of the successful lookup latency and decreasing of successful lookup rate, even more significant than the impacts caused by nodes leave events. Regarding to these results, the individual rejoin, as commonly implemented in current Chord-based hierarchical P2P, will suffer the overlay stability. In addition, we also note that in order to hold the relative-stable performance, nodes leave events produce higher traffic load than nodes join events.


Churn, hierarchical P2P, stability, heterogeneous network