MAAN: A Multi-Attribute Addressable Network for Grid Information Services Min Cai, Martin Frank, Jinbo Chen, Pedro Szekely Information Sciences Institute, University of Southern California Jan 31, 2004 Abstract. Recent structured Peer-to-Peer (P2P) systems such as Distributed Hash Tables (DHTs) offer scalable key-based lookup for distributed resources. However, they cannot be simply applied to grid information services because grid resources need to be registered and searched using multiple attributes. This paper proposes a Multi-Attribute Addressable Network (MAAN) that extends Chord to support multi-attribute and range queries. MAAN addresses range queries by mapping at- tribute values to the Chord identifier space via uniform locality preserving hashing. It uses an iterative or single attribute dominated query routing algorithm to resolve multi-attribute based queries. Each node in MAAN only has O(log N ) neighbors for N nodes. The number of routing hops to resolve a multi-attribute range query is O(log N + N × smin), where smin is the minimum range selectivity on all attributes. When smin = ε, it is logarithmic to the number of nodes, which is scalable to a large number of nodes and attributes. We also measured the performance of our MAAN implementation and the experimental results are consistent with our theoretical analysis. Keywords: Grid Computing, Peer-to-Peer, Information Services, Multi-attribute Range Queries 1. Introduction Grid computing is emerging as a novel approach of employing dis- tributed computational and storage resources to solve large-scale prob- lems in science, engineering, and commerce. Grid computing on a large scale requires scalable and efficient resource registration and lookup. Traditional approaches maintain a centralized server or a set of hierar- chically organized servers to index resource information. For example, Globus (Foster and Kesselman, 1997) uses an LDAP-based directory service named MDS (Fitzgerald et al, 1997) for resource registration and lookup. However, the centralized server(s) can become a registra- tion bottleneck in a highly dynamic environment where many resources join, leave, and change characteristics (such as CPU load) at any time. Thus, it does not scale well to a large number of grid nodes across autonomous organizations. Also, centralized approaches have the in- herent drawback of a single point of failure. Hierarchical approaches provide better scalability and failure tolerance by introducing a set of c  2004 Kluwer Academic Publishers. Printed in the Netherlands. maan.tex; 25/05/2004; 12:38; p.1