| EPSRC Reference: |
GR/J53348/01 |
| Title: |
A PARALLEL FUNCTIONAL TRANSACTION PROCESSOR |
| Principal Investigator: |
Peyton Jones, Professor L |
| Other Investigators: |
|
| Researcher Co-Investigators: |
|
| Project Partners: |
|
| Department: |
School of Computing Science |
| Organisation: |
University of Glasgow |
| Scheme: |
Standard Research (Pre-FEC) |
| Starts: |
01 April 1994 |
Ends: |
30 September 1997 |
Value (£): |
156,786
|
| EPSRC Research Topic Classifications: |
| Information & Knowledge Mgmt |
|
|
| EPSRC Industrial Sector Classifications: |
|
| Related Grants: |
|
| Panel History: |
|
|
Summary on Grant Application Form |
|
Parallel transaction processing is of such commercial significance that is has become one of the main applications of emerging highly-parallel computer systems. At the same time, parallel transaction processors are also the subject of active research: the field is young and many important challenges remain.The PARADE project takes a new approach to building a parallel transaction processor, by constructing it in a purely-functional language. It has been recognised for some time that functional languages are suitable for querying and modelling databases, but previous work has been limited in nature and updating has been almost exclusively sequential. Our objective is to implement a functional transaction-processing system on a leading-edge commercial parallel system, to study its behaviour, and to compare its performance with standard database software (Oracle) running on the same system.Progress: In close collaboration with the AQUA project (also at Glasgow), we have now almost completed the implementation of the initial version of our parallel runtime system for the Glasgow Haskell compiler (GHC). This is implemented above the portable PVM message-passing library, but is not in any way dependent on that library. Our overall design was published at the 1994 International Workshop on the Implementation of Functional Programming Languages. In order to test our ideas about data distribution in general parallel systems, we have implemented a fairly accurate parallel machine simulator, GranSim, using GHC as a basis. The simulator allows us to study many aspects of a parallel machine design, including the effect of varying communications latency on overall system performance. GranSim provides useful design input for the rest of the PARADE project. In particular, we have already confirmed the importance of task migration for our model of dynamic load management. Gransim has also demonstrated that it is more efficient to use synchronous rather than asynchronous messages when fetching data on a demand-driven basis. We now intend to enhance the simulator to incorporate the more sophisticated data fetching and distribution techniques we have designed for PARADE transaction processing. A description of this system will be published in the 1995 conference on High Performance Functional Computer Systems (HPFC 95). The Haskell 1.3 I/O design includes direct disk-access features which we will need for database programming, and which were previously rarely found in pure functional languages. The I/O design is endorsed by the principal Haskell compiler writers and by an international committee of prominent functional programmers. Within this framework, we have also prototyped ideas for incremental I/O, which we believe will be necessary for PARADE. We have been conducting detailed research into conventional concurrent transaction processing systems, attending relevant conferences, and organising research visits. We have been collaborating with colleagues in Austria, Germany and Greece, with a view to exploring Geographical Information Systems as an applications area. We actively maintain links with more conventional database researchers at our university. In particular, we are organising a joint one-week mini-workshop on parallel functional database technology. We have invited colleagues from both the UK and abroad to attend this workshop.
|
|
Key Findings |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Potential use in non-academic contexts |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Impacts |
|
| Description |
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk |
| Summary |
|
| Date Materialised |
|
|
|
|
Sectors submitted by the Researcher |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
| Project URL: |
|
| Further Information: |
|
| Organisation Website: |
http://www.gla.ac.uk |