Large web-scale applications typically use a distributed platform, like clusters of commodity servers, to achieve scalable and low-cost processing. The Map-Reduce framework and its open-source implementation, Hadoop, is commonly used to program these applications. Since these applications scale well with an increased number of servers, the cluster size is an important parameter. Cluster size however is constrained by power consumption. In this paper we present a system that uses low-power CPUs to increase the cluster size in a fixed power budget. Using low-power CPUs leads to the situation where the majority of a server's power is now consumed by the I/O sub-system. To overcome this, we develop a virtualized I/O sub-system where multiple servers share I/O resources. An ASIC based high-bandwidth interconnect fabric, and FPGA based I/O cards implement this virtualized I/O. The resulting system is the first production quality implementation of cluster-in-a-box that uses low-power CPUs. The unique design demonstrates a way to build systems using low-power CPUs, allowing a much larger number of servers in a cluster in the same power envelope. To overcome software inefficiency and increase the utilization of virtualized disk bandwidth, optimizations necessary for the operating system are also discussed. We built hardware based on these ideas and experiments on this system show a 3X average improvement in performance-per-Watt-hour compared to a commodity cluster with the same power budget.

Load latency contributes significantly to execution time. Because most cache accesses hit, cache-hit latency becomes an important component of expected load latency. Most modern microprocessors have base+offset addressing loads; thus effective cache-hit latency includes an addition as well as the RAM access. This paper introduces a new technique used in the UltraSPARC III microprocessor Sum-Addressed Memory (SAM), which performs true addition using the decoder of the RAM array, with very low latency. We compare SAM with other methods for reducing the add part of load latency. These methods include sum-prediction with recovery, and bitwise indexing with duplicate-tolerance. The results demonstrate the superior performance of SAM.