Abstract-We present MemX -a distributed system that virtualizes cluster-wide memory to support data-intensive and large memory workloads in virtual machines (VMs). MemX provides a number of benefits in virtualized settings: (1) VM workloads that access large datasets can perform lowlatency I/O over virtualized cluster-wide memory; (2) VMs can transparently execute very large memory applications that require more memory than physical DRAM present in the host machine; (3) MemX reduces the effective memory usage of the cluster by de-duplicating pages that have identical content; (4) existing applications do not require any modifications to benefit from MemX such as the use of special APIs, libraries, recompilation, or relinking; and (5) MemX supports live migration of large-footprint VMs by eliminating the need to migrate part of their memory footprint resident on other nodes. Detailed evaluations of our MemX prototype show that large dataset applications and multiple concurrent VMs achieve significant performance improvements using MemX compared against virtualized local and iSCSI disks.

In the technique known as network coordinates, the network latency between nodes is modeled as the distance between points in a metric space. Actual network latencies, however, exhibit numerous triangle inequality violations, which result in significant error between the actual latency and the distance as determined by the network coordinates. In this work, we show how graph clustering techniques can be used to find regions of the network space that show low triangle inequality violation within the region. By using techniques to increase the relative edge density in these regions, we improve the accuracy of network coordinates in these regions. We reduce the relative error within a cluster by 15% on average for the Meridian dataset, and by 7% over all; when compared to a single spring relaxation over the whole network.