innovative system offering application developers an extensively flexible group communication model is described. The emergence of process-group environments for distributed computing represents a promising step toward robustness for mission-critical distributed applications. Process groups have a “natural’ ’ correspondence with data or services that have been replicated for availability or as part of a coherent cache. They can be used to support highly available security domains, and group mechanisms fit well with an emerging generation of intelligent network and collaborative work applications.

Traditional object-oriented design methods deal with the functional aspects of systems, but the do not address quality of service (QoS) aspects such as...

This paper presents the Maestro Tools, a distributed object layer built on top of the Ensemble group communication system developed at Cornell. The Maestro tools include a visual application development environment; an interface to and implementation of fundamental distributed object abstractions, such as CSCW (cooperative-work), client/server, and publish/subscribe objects; and a set of adaptive control/policy modules which are used to adjust Ensemble failure detection and group multicast protocols in accordance with the application's quality of service requirements. The Maestro tools have been used in a reliable CORBA system and in a number of other projects. Results reported in this paper are work in progress: about 2/3 of our goals have been achieved, but there is still work to be done before the full-scale object layer is completed. 1 Introduction The rapid convergence of reliable distributed computing and object-oriented technologies is a reality of modern network computing. Thi...

This paper presents the design and implementation of a Java package called Filterfresh for building replicated fault-tolerant servers. Maintaining the correctness and integrity of replicated servers is supported by a GroupManager object instantiated with each replica to form a logical group. The Group Managers use a Group Membership algorithm to maintain a consistent group view and a Reliable Multicast mechanism to communicate with other Group Managers. We then demonstrate how Filterfresh can be integrated into the Java RMI facilities. First we use the GroupManager class to construct a faulttolerant RMI registry called FT Registry|a group of replicated RMI registry servers. Second, we describe our implementation of the FT Unicast|a client-side mechanism that tolerates and masks server failures below the stub layer, transparent to the client. We also present initial performance results, and discuss how general purpose RMI servers can be made highly available using the Filterfresh package. 1

Modern, distributed computing systems need to cope continuously with changes. We identify two kinds of changes: low frequency infrastructural changes, such as software upgrade; and frequent changes in the execution environment, such asnetwork bandwidth, memory availability. This paper proposes 2K, a component-based operating system architecture for rapidly changing environments. In 2K, adaptation is driven by architectural awareness: the system software includes models of its own structure, state, and behavior. Change in the dynamic structure of the system triggers adaptation. To implement adaptation, 2K incorporates a re ective middleware layer that admits dynamic customization through the dynamic loading of components and modi cation of object mechanisms such as method dispatch. The architecture also addresses adaptation in response to software and hardware evolution through code distribution, implemented using the active networking model. Prototype implementations of 2K ideas had provided encouraging results. We also present an application study, identify problems that arose while implementing the application on a traditional system, and analyze how 2K features would alleviate those problems.

OMG CORBA applications require a distributed naming serviceinorder to install and to retrieve object references. High availability of the naming service is important since most CORBA applications needto access it at least once during their lifetime. Unfortunately, the OMG standards do not deal with availability issues# the naming services of many of the commercially available CORBA object request brokers introduce single points of failure. In this paper we describe the design and implementation of a replicated, highly-available CORBA name server that adheres to the OMG Common Object Services Specification. Our naming servicecan bereplicated at run-time, while many applications are installing and retrieving object references. We compare our approach with the approaches taken by the ILU, NEO, Orbix, and DOME object request brokers. The performance of our name server is measured for various degrees of replication.