The feasibility of distributed computing is well proven by the tremendous success of distributed systems in the past two decades. However, advantages of distributed systems and computer networks do not come for free. The design of such systems is quite complex, in part due to unpredictable faults an
On the computational power of self-stabilizing systems
โ Scribed by James Abello; Shlomi Dolev
- Publisher
- Elsevier Science
- Year
- 1997
- Tongue
- English
- Weight
- 859 KB
- Volume
- 182
- Category
- Article
- ISSN
- 0304-3975
No coin nor oath required. For personal study only.
โฆ Synopsis
The computational power of self-stabilizing distributed systems is examined. Assuming availability of any number of processors, each with (small) constant size memory we show that any computable problem can be realized in a self-stabilizing fashion.
The result is derived by presenting a distributed system which tolerates transient faults and simulates the execution of a Turing machine. The total amount of memory required by the distributed system is equal to the memory used by the Turing machine (up to a constant factor).
๐ SIMILAR VOLUMES
It is shown that non-returning parallel communicating grammar systems with -free contextfree components can generate any recursively enumerable language. Since it was proven that such systems can be simulated by returning PC grammar systems with context-free components, the result extends to those.
The theory of computational complexity and certain explicitly-stated hypotheses imply limitations on the information processing power of biological systems. Parallelism, special purpose organization, and analog mechanisms may provide speedup critical for life processes, but have little power in the
We present a relation between the sets accepted by two-way pushdown automata and certain tape complexity classes of off-line Turing machines. Specifically, let L be a language accepted by a nondeterministic off-line Turing machine T. Let T have a t-symbol storage-tape alphabet. If for all but a fini