The impact of DHT routing geometry on resilience and proximity

Abstract

The various proposed DHT routing algorithms embody several different underlying routing geometries. These geometries include hypercubes, rings, tree-like structures, and butterfly networks. In this paper we focus on how these basic geometric approaches affect the resilience and proximity properties of DHTs. One factor that distinguishes these geometries is the degree of flexibility they provide in the selection of neighbors and routes. Flexibility is an important factor in achieving good static resilience and effective proximity neighbor and route selection. Our basic finding is that, despite our initial preference for more complex geometries, the ring geometry allows the greatest flexibility, and hence achieves the best resilience and proximity performance.

On the Cost of Participating in a Peer-to-Peer Network

“…This paper takes a first step towards quantifying the level of collaboration that can be expected from each participant, by proposing a model to evaluate the cost each peer incurs for being a part of the overlay. Such a cost model has several useful applications, among which, (1) providing a benchmark that can be used to compare between different proposals, complementary to recent works comparing topological properties of various overlays [7,12], (2) allowing for predicting disincentives, and designing mechanisms that ensure a protocol is strategyproof [16], and (3) facilitating the design of load balancing primitives.…”

Efficient Distributed Web Crawling Utilizing Internet Resources

“…The above method can be classified as a proximity identifier selection [17] method used to solve the DHT's topological inconsistency problem. However, different from the existing examples such as Topologically-aware CAN [18] and Proximity-aware CAN [19], our system not only maps the peers (crawlers) from NC to CAN, but also maps the resources (Web hosts or pieces).…”

Churn Tolerance Improvement Techniques in an Algorithm-Neutral DHT

“…The number of messages sent to make replicas is identical in both cases, but efficiency of those messaging differ if node IDs reflect network proximity. The responsible node is better if the distance between node IDs in the routing algorithm reflects network proximity with proximity identifier selection (PIS) [5]. The efficiency is identical in both cases because our implementation does not take account of network proximity currently.…”

“…They also measured and evaluated techniques to reduce the required time to perform routing. Those techniques include TCP-style timeout calculation and proximity neighbor selection (PNS) [5].…”