SCALLOP: a scalable and load-balanced peer-to-peer lookup protocol

Abstract

A number of structured peer-to-peer (P2P) lookup protocols have been proposed recently. A P2P lookup protocol routes a lookup request to its target node in a P2P distributed system. Existing protocols achieve balanced routing traffic among nodes by assuming that lookup requests are evenly targeted at every node. However, when lookup requests concentrate on a few nodes simultaneously, these nodes become hot spots. Due to uneven routing patterns in existing protocols, hot spots cause unbalanced routing traffic which leads to routing bottlenecks. In this paper, we present a novel structured P2P lookup protocol called SCALLOP that delivers balanced routing and avoids routing bottlenecks at occurrences of hot spots. Among existing protocols, SCALLOP is the first one to accomplish this goal at the fundamental nature of a routing protocol. SCALLOP achieves balanced routing by uniquely constructing a balanced lookup tree for each node. The balanced tree evenly distributes routing traffic among sibling nodes and, therefore, avoids or reduces routing bottlenecks. In addition, as a load-balanced protocol, SCALLOP delivers asymptotically optimal lookup performance at the tradeoff between routing path and routing table size. We conducted a set of simulations to demonstrate the effectiveness of SCALLOP. The results show that, compared with a most-referenced and representative structured P2P lookup protocol and a graph-based extension of this protocol, SCALLOP significantly reduces routing bottlenecks while all three protocols deliver comparable lookup performance.

Workload Modeling for Computer Systems Performance Evaluation

“…This means that most of the traffic is spread uniformly across all destinations, but a certain fraction is directed at one specific destination (or a small set of destinations) [546,432,17,163,545,133]. A generalization of this model is based on mass-count disparity: for example, 80% of the traffic can be directed at 20% of the destinations, while the other 20% of the traffic is distributed across the remaining 80% of the destinations [133]. Or, 90% of the I/O operations can refer to 10% of the files, while the remaining 10% of operations are spread across the remaining 90% of the files [577].…”

Enabling Dynamic Querying over Distributed Hash Tables

“…The tree is composed of four subtrees, each one having, as root, one of the fingers of Node 2 (that is, Nodes 3, 4, 6 and 10 ). Since the spanning tree corresponds to the lookup tree, which is a binomial tree in a (fully populated) Chord network [16], also the spanning tree associated with the broadcast over a fully populated Chord ring is a binomial tree.…”

Dynamic Querying in Structured Peer-to-Peer Networks

“…The overall broadcast procedure can be viewed as the process of passing the data item through a spanning tree, rooted at the querying node, which covers all nodes in the network. Since the spanning tree corresponds to the lookup tree, which is a binomial tree in a (fully populated) Chord network [14], also the spanning tree associated to the broadcast over a fully populated Chord ring is a binomial tree.…”

“…In these cases, the generated number of messages depends on the accuracy of the popularity estimation, which is better when a higher number of nodes is queried during the probe query (that is, when V includes a finger with a high index). For instance, when r = 1%, the average number of messages is 25207 for V = {F 8 }, 14341 for V = {F 11 }, and 13169 for V = {F 14 …”