The load index most frequently used for load balancing in distributed
systems is the job queue length. This work examines some of the implic
ations of scheduling jobs according to an additional abstract dimensio
n attribute called job size. The load balancing algorithm is supported
by a virtual ring structure which organises the network nodes in grou
ps and defines the information-gathering activities to take place with
in and between such groups. A two-phase approach to information gather
ing and decision making is adopted. This enables the selection of jobs
for transfer to be delayed until as close as possible to the moment o
f transfer. The operation of the protocol is described for each phase
and synchronisation of the parallel activities in the virtual rings is
discussed. The schedule length performance of the distributed algorit
hm is examined in a series of closed-system tests.