Computational Grid as an Appropriate Infrastructure for Ultra Large Scale Software Intensive Systems
Babak Rezaei Rad, Fereidoon Shams Aliee · 2010
Ultra large scale (ULS) systems are future software intensive systems that have billions of lines of code, composed of heterogeneous, changing, inconsistent and independent elements that are dispersed through worldwide global networks. Huge scale of this heterology in addition to distinctive characteristics like decentralization of operation, management, deployment and evolution, continuous evolving, people participation in system operation and normal failures that dominate in ULS systems face them to new challenges that today software architectures and developers have less or no concerns about them. To face these challenges there must be an appropriate infrastructure with enough or at least extendable features and capabilities to support ULS distinctive characteristics. In this paper we propose computational grid as an appropriate infrastructure for ULS systems that not only supports some of ULS characteristics itself, but also can exploit service oriented architecture, fault tolerant approaches, distributed management algorithms and many other mechanisms in order to extend its fundamental capabilities to adapt itself with ULS systems. We explain what kind of grid is appropriate for ULS and expound how it provides solutions for each of the ULS particular characteristics.