Horizons in scientific and distributed computing

The article surveys current technologies relevant to developing scientific applications on globally distributed networks. It reviews some networking and computing technologies that are having a significant impact, real or perceived, in the commercial computing world and might be valuable for future distributed scientific computing. In many cases, an unfortunate combination of technical inbreeding and aggressive marketing has created jargon and hyperbole barriers to understanding. So, presentations of these technologies too often use terminology potentially foreign to scientific computing people. That has been our experience at least. The article's goal is to remove some of these barriers. Developing scientific applications on globally distributed networks requires language support (Java, MPI, OpenMP), mechanisms for managing distributed computations and services (components and agents), and advanced networking technologies (IPv6, ATM).

[1]  David A. Padua,et al.  Cedar Fortran and Its Compiler , 1990, CONPAR.