Sacrificing true distribution for gaining access efficiency of replicated shared objects
Henning Pagnia, Oliver Theel · 2002
True distribution, in general, is a useful property. Nevertheless, problem domains exist, like the management of replicated shared objects in heterogeneous networks, where assumptions fundamental to the justification of truly distributed algorithms simply do not hold any longer. The paper demonstrates how the true distribution property can be sacrificed in exchange for improved access efficiency. It proposes a novel replication control scheme called Second Chance Protocol, which sacrifices true distribution.