Loading...
Please wait, while we are loading the content...
Similar Documents
Using Remote Procedure Calls Implementing Distributed Algorithm
| Content Provider | Semantic Scholar |
|---|---|
| Author | Murali, Ganta Anusha, K. Shirisha, A. Sravya, S. |
| Copyright Year | 2011 |
| Abstract | Remote Procedure Call (RPC) is a powerful primitive used for communication and synchronization between distributed processes. RPC poses a problem that it reduces the amount of parallelism, because of its synchronous nature. This paper shows how simple processes can be used to find a way of avoiding a difficulty in this problem. The combination of blocking RPC calls and light-weight processes provides both simple semantics and efficient exploitation of parallelism. We will describe how two important classes of algorithms, branch and bound can be run in a parallel way using this RPC. The results of some experiments comparing this algorithms on a single processor discussed. 1.INTRODUCTION: As computing technology advances, by just increasing the speed of the chips it becomes increasingly difficult and expensive to make the computers faster. In copper wire, electrical signals travel at 2/3 the speed of light or about 20 cm/nanosecond, so very fast computers must be very small, which leads to severe heat dissipation problems among other things. The clear solution is to achieve the same computing power as one very fast computer, but at a fraction of the cost by combining large number of moderately fast computers. In distributed systems, organizing multiple processors in different ways have been proposed. At one end of the spectrum are tightly-coupled systems with multiple processors on the same bus and sharing a common memory. At the other end are the loosely-coupled systems consisting of a number of independent computers, each with its own operating system and users, exchanging files and mail over a public data network. In between, are systems consisting of mini or microcomputers communicating over a fast local network and all running a single, system-wide operating system. We have used a system in the latter category as a testbed for the implementation of some distributed algorithms. 1.1 Distributed System: A distributed system consists of different independent computers that communicate with a computer network. To achieve common goals computer interacts each other. Distributed program defined as a computer program that runs in a distributed system and process of writing such programs is called distributed programming. To solve computational problems distributed computing refers to use of distributed system. Within some geographical area networks where individual computers were physically distributed is referred as the word distributed in such as distributed system, distributed programming. The term distributed are used in wider sense, also referring to independent processes that execute on the same computer and to pass the message they intersect each other. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://ijcta.com/documents/volumes/vol2issue6/ijcta2011020603.pdf |
| Alternate Webpage(s) | http://www.ijcta.com/documents/volumes/vol2issue6/ijcta2011020603.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
