RESQme studies for SDC regional computing centers
The Technical Design Report for the SDC proposes a model for offline computing which includes a computing center at the SSCL, containing all of the physics data, together with a number of regional computing centers around the world. These regional centers would contain subsets of the data, and would support the daily work of most physicists. For large or unusual requests, where the data are not held at the regional center, the requests would automatically be forwarded to the SSCL. It is assumed that the closeness'' of the regional centers and their reduced demand from fewer users would result in improved system performance. Such a system is too complex to model analytically; simulation is the only viable approach. However, Monte Carlo models built from scratch for complicated systems are very difficult to maintain and hard to modify. Fortunately, we have obtained from IBM a modeling framework, RESQme, explicitly designed for building statistical models of computer systems. This note describes a first pass at modeling the proposed offline system.
- Research Organization:
- Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)
- Sponsoring Organization:
- USDOE; USDOE, Washington, DC (United States)
- DOE Contract Number:
- AC03-76SF00098
- OSTI ID:
- 7368579
- Report Number(s):
- LBL-33540; ON: DE93015230
- Country of Publication:
- United States
- Language:
- English
Similar Records
Using object oriented analysis and design to study the SSCL SDC computing system
Using object oriented analysis and design to study the SSCL SDC computing system
Related Subjects
99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE
DATA PROCESSING
STATISTICAL MODELS
SUPERCONDUCTING SUPER COLLIDER
COMPUTER ARCHITECTURE
COMPUTERS
SHOWER COUNTERS
MATHEMATICAL MODELS
MEASURING INSTRUMENTS
PROCESSING
RADIATION DETECTORS
STORAGE RINGS
430100* - Particle Accelerators- Design
Development
& Operation
990200 - Mathematics & Computers