Rensselaer Polytechnic Institute researchers have developed a scheme to solve a fundamental difficulty with distributed grid computing: coordinating the efforts of all computers.

The simple solution avoids the need to have a global supervisor, which would introduce scaling problems. Each individual computer makes occasional checks with randomly-chosen others, to ensure it is properly synchronized. The result is a self-stabilising effect on the system as a whole; processors that are ahead with their calculations are asked by others to wait for the next batch of data.