Multicomputer object stores: the Distributed Texas experiment
نویسنده
چکیده
The paper is about design of object stores for distributed memory computers. Problems of distribution within such stores are described through the design of Distributed Texas, a system derived from the Texas store [6] and implemented on a scalable multicomputer (a Fujitsu AP1000). Problems discussed include creating a single image of an object space across multicomputer architectures, tradeoffs between page-based and object-based granularity, the distribution of store level services and the management of concurrency and persistance within those distributed services. Problems of building layered software on multicomputers pervade these issues. Preliminary scalability performance figures for Distributed Texas are also presented.
منابع مشابه
Multicomputer Object Stores: The Multicomputer Texas Experiment
The paper is about design of object stores for distributed memory computers. Problems of distribution within such stores areThis paper appeared in the proceedings of the Seventh International Workshop on Persistent Object Systems described through the design of Multicomputer Texas, a system derived from the Texas store [8] and implemented on a scalable multicomputer (a Fujitsu AP1000). Problems...
متن کاملRecovery and Page Coherency for a Scalable Multicomputer Object Store
This paper presents scalable algorithms for recovery and page coherency in multicomputer object stores. Recovery and coherency are central to object store engineering and distributed memory multicomputers are fundamental to scalable computation. Efficient recovery is implemented through a combination of local logging and a localisation of the transactional workspace model. A vector of update co...
متن کاملHeROD Flavoured Oct-Trees: Scientific Computation with a Multicomputer Persistent Object Store
Using a persistent multicomputer object store should greatly simplify the writing of distributed memory parallel programs operating on irregular, object-structured data, by removing from the programmer the burden of managing data referencing, distribution and coherency. In this report we explore the development of such a program using the HeROD persistent multicomputer object store, in a comput...
متن کاملGeneric Containers for a Distributed Object Store
In this paper we report of an experiment about how generic container classes can be used to build up a distributed main memory object store. This approach is inspired by the C++ standard template library (STL) but uses set-like associative structures instead of the arraylike sequential structures mainly used by the STL. The basic structure is a container with only fundamental functionality. The...
متن کاملThe M-Cache: A Message-Handling Mechanism for Multicomputer Systems
This paper presents the design and evaluation of the M-cache, a small, fast and intelligent memory for handling messages at the processing nodes of multicomputer systems. The M-cache is neither a conventional cache nor a message bu er, though it has some characteristics of both. It provides hardware support for the message search operation often performed in message-directed programming, a prog...
متن کامل