Computational Science
Definition: multidisciplinary field that uses advanced computing capabilities to understand and solve complex problems. Need advanced computing capabilities.
To access advance computing capabilities, one can:
Goals of Distributed Systems
Goals of DS: 1. Access to resources
Goals of DS: 2. Hiding distribution: Forms of transparency:
Access (differences in data representation),
Location (where a resource id located),
Migration (moved to another location),
Relocation (moved to other location while in use),
Replication (replicated),
Concurrency (shared by several competitive users),
Failure (failure and recovery of a resource),
Persistence (software is in memory or disk).
Goals of DS: 3. Fault-tolerance: dependability definition
Goals of DS: 3. Fault-tolerance: dependability attributes
availability (readiness of a system to accept and process requests),
reliability (providing uninterrupted service),
safety (provide service without the occurrence of catastrophic failures),
confidentiality (preventing unauthorized access to and/or handling of information),
integrity (the absence of improper alterations of information and system state),
maintainability (support of fault removal and service restoration, as well as undergo change),
security (combination of availability, integrity, and confidentiality for authorized users)
Goals of DS: 4. Scalability: three dimensions
size (new users and resource can be added),
geography and distribution (users and resources can be anywhere in the world),
administration (ease of management not dependent on system size).
Goals of DS: 4. Scalability: Threats
Centralized services, data, and algorithms
Taxonomy of Distributed Systems
Multiprocessor systems (shared memory x private memory, bus-based x switch-based)
Multicomputer systems
Network operating systems
Middleware-based operating systems