Future Grid network technology will face the problem of the efficient provisioning of services to clients by a scalable and dynamic resource allocation (matching) mechanism.
The objective of CATNETS is to determine the applicability of a decentralized economic self-organization mechanism for resource allocation in application layer networks (ALN), by
- producing a 'proof-of-concept' prototype in a real ALN and
- by evaluating its performance against existing resource brokerage approaches in a simulated ALN.
The term ALN integrates different Internet overlay network approaches, like Grid and P2P systems. The allocation of resources in these networks (e.g. matching of demand and supply, deployment of service instances, and service discovery) can principally be conducted in a centralized (e.g. using resource brokers) or in a decentralized fashion (e.g. using self-organizing mechanisms). Centralized approaches reach their limits with increasing network size and growing numbers of elements; self-organizing approaches thus gain attention, e.g. in IBM's Autonomic Computing initiative. The CATNETS project investigates a 'free market' economic self-organization approach, the 'Catallaxy' by Friedrich A. von Hayek, as the basis for self-organizing resource allocation in ALNs.
A preliminary evaluation of 'Catallactic' mechanisms in the FET assessment project CATNET (IST-2001-34030) by simulation has shown positive results, upon which the CATNETS project will build. The performance measurements of the 'proof-of-concept' implementation will be compared against the simulation results, with the goal of being able to make a substantiated statement on the applicability of economic self-organization as a major component of ALN networks. A positive evaluation of the Catallactic approach would have a high potential impact, with new possibilities for resource brokering in future ALN, and maybe for self-organization in computing in general.
Funding SchemeSTREP - Specific Targeted Research Project
CF24 0DE Cardiff