Quantum modeling of distributed object computing

Abstract

Distributed object computing is rapidly gaining notice in the computer engineering community. In response, productized support for these systems is increasing. However, few tools exist for the design and modeling of distributed object systems. Clearly, large scale and mission critical systems are most affected by this deficiency. To attain a level of tractability in large scale design and verification, dynamic models are essential. This paper introduces two modeling components. First, the LCN (Loosely Coupled Network) model supports the definition of real-world network topologies and multicomputer systems. Second, the DCO (Distributed Cooperative Objects) model provides a system-independent method of representing complex, object oriented computing structures. These components are assigned random variables (e.g., bandwidths, error rates, compute loads, data volumes, etc.) that provide a basis for dynamic behavior. Through surjective mapping of DCO software onto LCN hardware, the complete OSM (Object System Mapping) model is created. It is shown that OSM systems can model all modes of distributed object behavior found in the current state of Object Request Broker (ORB) technology while under the constraints of modern network structures. The paper concludes with a qualitative analysis of these behaviors and sets the foundation for a distributed object simulation facility.