Using PVM to Implement a Parallel Database System 1 )

نویسندگان

  • G. Bozas
  • R. Lehn
  • A. Listl
  • M. Pawlowski
چکیده

MIDAS (MunIch Parallel DAtabase System) is a parallel relational database system. It is well suited to serve as a platform for the exploration of various strategies in order to parallelize a rela-tional database system. This short paper depicts how the abstract architecture of MIDAS looks like. A detailed description shows the implementation of a MIDAS prototype on a network of UNIX workstations. The core of the MIDAS protoype is implemented using PVM. But pure PVM does not fulfil all demands in order to implement a parallel database system. We demonstrate how our implementation combines the PVM programming model with standard UNIX mechanisms. We use UNIX shared memory to enhance the efficiency of our implementation and we use access privileges to UNIX files to secure the database against unauthorized access. MIDAS has a client/server architecture (see figure 1). The MIDAS server provides a high level interface (SQL) to retrieve and manipulate efficiently the contents of a relational database. The database itself is stored in the filesystem provided by the implementation platform. MIDAS clients are database applications. They are sequential programs issuing transactions to the MIDAS server. Each transaction on its part consists of retrieval/update queries. A MIDAS client can run on any computer having access to the MIDAS server. 1) This research is directed by Prof. R. Bayer Ph.D. and is part of the more extensive project " Tools and Methods for Using Parallel Computer Architectures " executed at the dept. of computer science of the Technische Universität München since 1990 (sponsored by DFG contract SFB 342).

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Network recourses (NOW) to remove the work load for Deriving rule for Semantic Query Optimization and Speed up answering queries

The rapid growth in the size of databases and the advances made in Query Languages has resulted in increased SQL query complexity submitted by users, which in turn slows down the speed of information retrieval from the database. The future of high performance database systems lies in parallelism. Commercial vendors’ database systems have introduced solutions but these have proved to be extremel...

متن کامل

Offering Parallelism to a Sequential Database Management System on a Network of Workstations Using PVM

The considerable growth of on-line document searching and consulting brings much of the data providers to reconsider their database management systems (DBMS) capacities. Parallel DBMS then appear as a good solution, but the involved changes in administration and cost limit their breakthrough. To overcome these drawbacks, we propose an hybrid structure, which adapts a parallel extension to an ex...

متن کامل

PVM Experiences in Developing the MIDASParallel Database

We are currently developing a parallel database system that exploits PVM as communication platform. In this paper we describe the communication characteristics of this application and specify to which extent PVM innuences the overall performance of our system. Our experiences with PVM are shown, advantages and some shortcomings are summarized, and hints for eeciently implementing applications w...

متن کامل

Experiences with Modelling an Open Pvm Environment Using Computer Simulation

Parallel Virtual Machine (PVM) is a widely{used software system that permits a heterogeneous collection of computers networked together to be viewed by a user's program as a single distributed{memory parallel machine. In this paper, the application of simulation to develop a functional and performance model of the PVM system and PVM applications running in an open workstations environment was i...

متن کامل

A Multiplatform Distributed FEM Analysis System using PVM and MPI

A multiplatform computational system for parallel finite element structural analysis using a distributed memory environment is described in this paper. The complete system is comprised by integrated programs, each of one responsible for a different task: pre-processing, mesh partitioning (necessary to perform the parallel analysis), structural analysis and post-processing. The main focus here i...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 1994