Designing efficient algorithms for distributed systems
Search for efficient algorithms for distributed systems has become an important area of computer science. This research is driven by the need to efficiently process and communicate information generated by the system. In distributed systems, topological information plays an important role in the design of fast algorithms for problems such as routing, broadcasting, and sorting. The central focus of this dissertation is the design and analysis of distributed algorithms for determining topological information in asynchronous communication networks. Specifically, the author presents distributed algorithms for two generic problems: distributed graph problems and network traversal problems. Network location and network recognition are two important graph problems in distributed systems. He presents unified algorithms for locating centers and medians of asynchronous communication networks. Also, he presents both the centralized and decentralized versions of the algorithm. Furthermore, this is the first decentralized algorithm reported in the literature. These results are further extended to weighted networks. In addition, the unified algorithm can also be used to determine other topological parameters such as the diameter, and centroids of distributed networks. Efficient algorithms for problems such as finding shortest paths, centers, and sorting could be designed if the network topology is known a priori.
- Research Organization:
- Louisiana State Univ., Baton Rouge, LA (United States)
- OSTI ID:
- 6091377
- Country of Publication:
- United States
- Language:
- English
Similar Records
Use of efficient broadcast protocols in asynchronous distributed systems. Doctoral thesis
Decentralized Schemes with Overlap for Solving Graph-Structured Optimization Problems