The Grid Computing Competence Center (GC3) at the University of Zurich (UZH) together with FGCZ (Functional Genomics Center Zurich), and SystemsX.ch will present at the International Supercomputing Conference 2013 (ISC13) a paper introducing the software framework VM-MAD.
The paper –authored by the entire VM-MAD team, including GC3’s Tyanko Aleksiev, Sergio Maffioletti, and Riccardo Murri– presents the motivation for starting VM-MAD, the reasoning behind the architectural choices and the implementation of the VM-MAD Orchestrator component, and shows some preliminary data.
The aim of VM-MAD is to provide mechanisms for running scientific applications on virtualized infrastructures (clouds). In detail, VM-MAD consists of the following components:
- a repository of VM images and associated metadata;
- a Python library to control the full lifetime of a VM instance on different cloud backends;
- an “orchestrator” component to extend SGE/OGS clusters with virtualized compute nodes;
- command-line tools to access the basic functionality of the above components.
Abstract (full article»):
The availability of powerful computing hardware in IaaS clouds makes cloud computing attractive also for computational workloads that were up to now almost exclusively run on HPC clusters.
In this paper we present the VM-MAD Orchestrator software: an open source framework for cloudbursting Linux-based HPC clusters into IaaS clouds but also computational grids. The Orchestrator is completely modular, allowing flexible configurations of cloudbursting policies. It can be used with any batch system or cloud infrastructure, dynamically extending the cluster when needed. A distinctive feature of our framework is that the policies can be tested and tuned in a simulation mode based on historical or synthetic cluster accounting data.
In the paper we also describe how the VM-MAD Orchestrator was used in a production environment at the FGCZ to speed up the analysis of mass spectrometry-based protein data by cloudbursting to the Amazon EC2. The advantages of this hybrid system are shown with a large evaluation run using about hundred large EC2 nodes.