TURBO: an autonomous execution environment with scalability and load balancing features

Abstract

Execution environments and application servers require load balancing features in order to scale properly whilst maintaining high performance and availability, especially when running mission-critical services. Usually it is difficult to apply load balancing to real systems due to their complexity and lack of tools that support their application; moreover their integration requires changes that could break design cohesion or introduce unmanageable complexity. To overcome some of those limitations, we propose a system called "telco uniform request balancer" (TURBO), a distributed execution environment that is natively designed to provide autonomic high scalability and dynamic load balancing. TURBO integrates load balancing behavior effortlessly by self-organization through a collaborative behavior. The system is based on a peer-to-peer variant of the Linda distributed model and a "poll when idle" (PID) approach for load distribution