Alternative, Low Power, Computer Architectures -- Jonathan Gluck (Garrod)

Jonathan Gluck and Charlie Garrod

In 2006, computer data centers represented 1.5% of the total consumption of electrical power in the U.S.. In a study conducted in 2011, it was found that these data centers comprised a full 2% of the total worldwide consumption of electricity. This growing power usage has lead to research into alternative, low power, computer architectures. We examine, FAWN, one of these computer architectures and attempt to build modern data structures using these simple systems. We implemented lists, sets, and maps on the existing architecture of FAWN-KV, a distributed hash table, and gauged the effects these higher level data structures had on the system. We have found that these higher level data structures can improve overall efficiency of the system, at small cost.