Grid is mostly heterogeneous system where the computation capability of each
node varies. Therefore, load distribution and load balancing are among the most
important issues in Grid. Some techniques have been proposed for dynamic and
adaptive load distribution, such as relocating jobs from a high-loaded node to
a low-loaded node while retaining job execution state based on virtual machine
and thread migration. In this paper, we propose a method for dynamic and
adaptive load distribution between computation nodes using mobile thread
migration which yields lightweight job relocation without unnecessary overhead.
Using an example problem called parallel PrefixSpan, whose computation cost is
absolutely unpredictable and which is used in analysis of amino-acid sequences,
we demonstrate effectiveness of our technique through experiments.