Abstract—Autonomous mobile programs (AMPs) have been proposed for load management in dynamic networks. An AMP is aware of its resource needs and periodically seeks a better location in the network to reduce execution time. AMPs have previously been measured using mobile Java Voyager on local area networks (LANs). We have constructed a simulation model of AMPs and reproduced 4 sets of experiments on homogeneous networks, i.e. networks where all locations have the same processor speed, and 2 sets of experiments on heterogeneous networks with collection of large and small AMPs. The results show that simulated collections of AMPs obtain similar balanced states to those reached in the real experiments, and have only minor differences from real experimental results. The simulation model gives an opportunity to explore the greedy effect that can be observed in the real experiments. This gives us confidence to apply the simulation model for further investigation of AMP behaviour, including behaviours on wide area networks. I.

Abstract—Distributed load managers exhibit thrashing where tasks are repeatedly moved between locations due to incomplete global load information. This paper shows that systems of Autonomous Mobile Programs (AMPs) exhibit the same behaviour, identifying two types of redundant movement and terming them greedy effects. AMPs are unusual in that, in place of some external load management system, each AMP periodically recalculates network and program parameters and may independently move to a better execution environment. Load management emerges from the behaviour of collections of AMPs. The paper explores the extent of greedy effects by simulation, and then proposes negotiating AMPs (NAMPs) to ameliorate the problem. We present the design of AMPs with a competitive negotiation scheme (cNAMPs), and compare their performance with AMPs by simulation. Keywords-mobile computation; load balancing; autonomous mobile program; greedy effect; mobile agent; scheduling; workflow management. I.

To share resources in open distributed systems we have developed autonomous mobile programs, which periodically use a cost model to decide where to execute in a network. In addition self-aware mobile coordination for common patterns of computation over collections are encapsulated by autonomous mobility skeletons. Performance can be improved if an autonomous mobile program can predict the cost of the entire program rather than a single iteration. We propose a cost calculus that calculate the costs for the remainder of a computation at arbitrary program points. We extend our autonomous mobility skeleton cost models to be parametrised on the cost of the remainder of the program, and propose costed autonomous mobility skeletons. An automatic cost analyser which implements the calculus has been built in Jocaml, which outputs Jocaml programs with higher-order functions replaced by costed autonomous mobility skeletons. 1

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: