Shahed University

The Parvicursor infrastructure to facilitate the design of Grid and Cloud computing systems

Alireza Poshtkohi | M.B. Ghaznavi-Ghoushchi | Kamyar Saghafi

Date :  2017/02/21
Publish in :    Computing
Link :
Keywords :Distributed systems, Grid computing, Cloud computing, Distributed thread-level parallelism, Secure high-throughput data transfer, .NET Framework, Scalable Internet services

Abstract :
During the past decades, different variants of technology solutions have emerged to eliminate the restrictions on the processing power of computers in solving various problems. Grid and Cloud computing patterns are among the most important of them. In this paper, we introduce a new infrastructure referred to as Parvicursor based on the distributed objects paradigm that can facilitate the construction of scalable and high-performance parallel distributed systems. It proposes several peer-to-peer services to construct scalable distributed system paradigms such as HPC, Grid and Cloud computing. Also, Parvicursor realizes a partial, native, cross-platform, high-performance and C++-based implementation of the .NET ECMA standards. To the best of our knowledge, Parvicursor.NET Framework is the first attempt that allows developers to implement .NET ECMA programs directly in native code. Parvicursor makes use of combining the thread-level parallelism and distributed memory programming models to exploit the strengths of both models in many-core era.