Using object oriented analysis and design to study the SSCL SDC computing system

Conference · Wed Jan 01 00:00:00 EST 1992

OSTI ID:5760345

Kubena, G T; Liao, K ^[1]; Palounek, P T ^[2]; Day, C ^[3]

---

1. International Business Machines Corp., Houston, TX (United States)
2. Los Alamos National Lab., NM (United States)
3. Lawrence Berkeley Lab., CA (United States)

A joint study between the Computer Working Group of the SSC Solenoidal Detector Collaboration (SDC) and IBM's Federal Sector Division is focusing on the use of Object Oriented Analysis and Design on the SDC Offline Computing System. One key challenge of the analysis is to discover an efficient way to store and subsequently retrieve raw and reconstructed event data, estimated to be 2 petabytes per year. The Object Oriented approach being used during the analysis and early design is intended to yield a smooth transition to detailed design, prototyping and implementation. The object oriented approach is used as a subprocess of a larger process used by IBM FSD, i.e., a systematic approach to architecting and integrating large complex systems. A description of the overall process and early results are described in a study report produced jointly by the SDC and IBM FSD. The overall process focuses on requirements analysis, operational concept development, problem domain decomposition, development and selection of candidate architectures, automated performance modeling and software architecture. This paper will focus primarily on software architecture. The high level software architecture is viewed as a layered stack consisting of: system services, common physics application framework and unique physics applications. Object oriented analysis is being used to investigate the data storage and management of the event data. An object hierarchy is being created and operational concept scenarios are being used to validate the design. Several database prototypes can then be developed, e.g. object oriented or relational, to prove the concept. The object oriented development is fundamentally different from traditional approaches to design, such as those based exclusively on data flow. Object oriented decomposition more closely models a person's perception of reality, hence the developed system is more understandable, extensible, and maintainable.

• Research Organization: Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
• Sponsoring Organization: USDOE; USDOE, Washington, DC (United States)
• DOE Contract Number: W-7405-ENG-36
• OSTI ID: 5760345
• Report Number(s): LA-UR-92-67; CONF-920172-2; ON: DE92007718
• Resource Relation: Conference: 2. international workshop on software engineering, AI, and neural nets for high energy and nuclear physics, L'Agelonde (France), 13-18 Jan 1992
• Country of Publication: United States
• Language: English

Related Subjects

43 PARTICLE ACCELERATORS
46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND TECHNOLOGY
99 GENERAL AND MISCELLANEOUS//MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE
DATA ACQUISITION SYSTEMS
COMPUTER ARCHITECTURE
SHOWER COUNTERS
SUPERCONDUCTING SUPER COLLIDER
DATA BASE MANAGEMENT
DESIGN
PROGRAMMING
MANAGEMENT
MEASURING INSTRUMENTS
RADIATION DETECTORS
STORAGE RINGS
430303* - Particle Accelerators- Experimental Facilities & Equipment
440104 - Radiation Instrumentation- High Energy Physics Instrumentation
990200 - Mathematics & Computers
990301 - Information Handling- Data Handling- (1992-)