Advance high performance cluster architecture

Abstract

A computer cluster is a group of loosely coupled computers that work together closely so that in many respects they can be viewed as though they are a single computer. Clusters are commonly, but not always, connected through fast local area networks. Clusters are usually deployed to improve speed and/or reliability over that provided by a single computer, while typically being much more cost-effective than single computers of comparable speed or reliability. Traditional cluster architecture consists of a server and some number of clients connected to each other via LAN cable. Everything takes place through this LAN cable. Here, everything refers to the broadcasting of messages, acknowledgements of these messages back to server and the data transfer. The route to every operation is a single LAN cable port. In Advanced High Performance Cluster Architecture (HPCA), we are divided these tasks and they are done via different ports. For broadcasting messages, we are going to use serial port. For taking acknowledgements from multiple clients in parallel, we are going to use parallel port. For data transfer between the machines of cluster, we are going to use TCP/IP. As we know, the computations are faster in C++ than in Java and coding power of Java is more than that of C++. Thus, we are going to write coding part in Java and computational part is given to C++. For maintaining connection between C++ and Java, the technology called JNI (Java Native Interface) is used. Because of all these modifications in the trivial cluster architecture, we make an Advanced High Performance Cluster Architecture (HPCA).