Grid Information Services for Distributed Resource Sharing
Karl Czajkowski Steven Fitzgeraldy Ian Fosterzx Carl Kesselman
 Information Sciences Institute y Department of Computer Science
University of Southern California California State University, Northridge
Marina del Rey, CA 90292 Northridge, CA 91330
z Department of Computer Science x Mathematics and Computer Science Division
The University of Chicago Argonne National Laboratory
Chicago, IL 60657 Argonne, IL 60439
Abstract
Grid technologies enable large-scale sharing of re-
sources within formal or informal consortia of individuals
and/or institutions: what are sometimes called virtual or-
ganizations. In these settings, the discovery, characteriza-
tion, and monitoring of resources, services, and computa-
tions are challenging problems due to the considerable di-
versity, large numbers, dynamic behavior, and geographical
distribution of the entities in which a user might be inter-
ested. Consequently, information services are a vital part
of any Grid software infrastructure, providing fundamental
mechanisms for discovery and monitoring, and hence for
planning and adapting application behavior.
We present here an information services architecture that
addresses performance, security, scalability, and robustness
requirements. Our architecture defines simple low-level en-
quiry and registration protocols that make it easy to incor-
porate individual entities into various information struc-
tures, such as aggregate directories that support a variety of
different query languages and discovery strategies. These
protocols can also be combined with other Grid protocols
to construct additional higher-level services and capabil-
ities such as brokering, monitoring, fault detection, and
troubleshooting. Our architecture has been implemented as
MDS-2, which forms part of the Globus Grid toolkit and has
been widely deployed and applied.
1. Introduction
Grid computing technologies enable wide-spread shar-
ing and coordinated use of networked resources [15]. Shar-
ing relationships may be static and long-lived—e.g., among
the major resource centers of a company or university—
or highly dynamic: e.g., among the evolving membership
of a scientific collaboration [17]. In either case, the fact
that users typically have little or no knowledge of the re-
sources contributed by participants in the “virtual organiza-
tion” (VO) poses a significant obstacle to their use. For this
reason, information services designed to support the ini-
tial discovery and ongoing monitoring of the existence and
characteristics of resources, services, computations, and
other entities are a vital part of a Grid system [13].
Such information services find uses in a variety of Grid
scenarios. The following examples illustrate but do not
exhaust the range of applications that rely on information
services, and the variety of information sources and infor-
mation access and management methods that are associated
with these applications [13, 30, 36, 11, 34].
A service discovery service operated by a community
records the identity and essential characteristics of the “ser-
vices” that are available to community members. Such a
discovery service might enable a physicist to determine that
a new university that has just joined his consortium has 100
new CPUs available for approved use. Here, information
sources are relatively static and the information itself relates
primarily to availability.
A superscheduler routes computational requests to the
“best” available computer in a Grid containing multiple
high-end computers, where “best” can encompass issues of
architecture, installed software, performance, availability,
and policy. Here, information sources are computers, and
information can include both relatively static information
such as system configuration (architecture, OS version, ac-
cess policy) and more dynamic information such as instan-
taneous load and predictions of future availability [40, 10].
A replica selection service within a data grid responds
to requests for the “best” copy of files that are replicated
on multiple storage systems. Here, information sources can
once again include system configuration, instantaneous per-
formance, and predictions, but for storage systems and net-
works rather than computers.
An application adaptation agent monitors both a run-
Proceedings of the 10th IEEE International Symposium on High Performance Distributed Computing (HPDC-10’01)
1082-8907/01 $10.00 © 2001 IEEE