Type-specific locking is designed to increase performance of distributed transactions by allowing competing transactions to concurrently alter shared objects concurrently, provided their changes are commutative. When one of the set of compatible transactions commits or aborts, a problem arises due to the indeterminacy of the alterations of the others. A tree of nineteen solutions to the problem has been investigated. Given the necessary support by the underlying architecture, this tree can be pruned severely. The problem and set of approaches came to the authors' attention during design for type-specific locking in PROFEMO, an operating system for a local area network that provides a location-independent object environment and built-in mechanisms for transaction control and recovery. A description is given of the problem, the possible solutions, and the arguments that allow the pruning. Testbed results are not yet available, since type-specific locking has not yet been implemented at the time of the initial tests of the PROFEMO system.>

A replicated directory system based on a method called regeneration is designed and implemented. The directory system allows selection of arbitrary object to be replicated, choice of the number of replicas for each object, and placement of the copies on machines with independent failure modes. Copies can become inaccessible due to node crashes, but as long as a single copy survives, the replication level is restored by automatically replacing lost copies on other active machines. The focus is on a regeneration algorithm for replica replacement and its application to a replicated directory structure in the Eden local area network. A simple probabilistic approach is used to compare the availability provided by the algorithm to three other replication techniques.>

We have designed and implemented a replicated directory system based on a method called Regeneration. The directory system allows selection of arbitrary objects to be replicated, choice of the number of replicas for each object, and placement of the copies on machines with independent failure modes. Copies may become inaccessible due to node crashes, but as long as a single copy survives, the replication level is restored by automatically replacing lost copies on other active machines. The paper focuses on the Regeneration algorithm for replica replacement and on its application to a replicated directory structure in the Eden system. Analytically, we use a simple probabilistic approach to compare the availability provided by the algorithm with other replication techniques. Empirically, we have measured the performance of the implementation.

The Eden project is a five year experiment in designing, building, and using an "integrated distributed" computing system. We are attempting to combine the benefits of integration and distribution by supporting an object based style of programming on top of a node machine/local network hardware base. Our experimental hypothesis is that such an architecture will provide an environment conducive to building distributed applications.

An extension of Petri nets called evaluation nets (E-nets) has been developed for use in representation of computer systems. These lead to interpreted graphs for use in human communication and to machine-interpretable descriptions. They may be used as aids in development of simulations and planning of measurements for examining characteristics such as throughput, turnaround time, and utilization of resources. This paper reviews the five primitives used to form nets expressing logical interactions among tokens (or tasks) flowing through the net and system resources represented by the net. Inherent in the use of the primitives is the ability to express time delays associated with the flow of tokens through the network and the modification of attributes of these tokens. This paper extends previous work on evaluation nets through the use of macro nets that represent structures made up of the primitives, allowing compression of the net for easier understanding. The macro nets are ``open-ended'' in the sense that the user can develop new macros to suit the needs of a particular study. Evaluation nets can be used informally like flow charts or block diagrams, although they provide greater power of expression than either. They can also be used more formally to develop machine aids, e.g., for on-line editing of evaluation nets or for development of simulation programs. Some examples of evaluation nets are provided that model a CDC 6400 and the cooperation of n independent sequential processes.