Storage elasticity on IaaS clouds is an important feature for data-intensive workloads: storage requirements can vary greatly during application runtime, making worst-case over-provisioning a poor choice that leads to unnecessarily tied-up storage and extra costs for the user. While the ability to adapt dynamically to storage requirements is thus attractive, how to implement it is not well understood. Current approaches simply rely on users to attach and detach virtual disks to the virtual machine (VM) instances and then manage them manually, thus greatly increasing application complexity while reducing cost efficiency.  This talk focuses on transparent techniques to handle storage elasticity for two dimensions:
(1) storage space; and (2) bandwidth. It introduces several adaptive monitoring and prediction schemes that automatically add and remove virtual disks of different capacities and I/O capabilities in order to get as close as possible to the application requirements in terms of both space and bandwidth. These contributions are discussed in the context of both synthetic benchmarks and real-life applications, where they demonstrate large reductions in cost with minimal impact on performance when compared to over-provisioning.

Bio:
Bogdan Nicolae is a member of the High Performance Systems group of IBM Research Ireland. Before this, he was a postdoc within the Joint Laboratory for Petascale Computing, an initiative of University of Illinois at Urbana-Champaign, the National Center for Supercomputing Applications and INRIA. He holds a PhD from University of Rennes 1, France (2010) and a Dipl. Eng. degree from Politehnica University Bucharest, Romania (2007). His research revolves around scalable storage techniques and fault tolerance for cloud computing and exascale architectures.