BEGIN:VCALENDAR
PRODID:-//Microsoft Corporation//Outlook MIMEDIR//EN
VERSION:1.0
BEGIN:VEVENT
DTSTART:20111116T011500Z
DTEND:20111116T030000Z
LOCATION:WSCC North Galleria 2nd/3rd Floors
DESCRIPTION;ENCODING=QUOTED-PRINTABLE:ABSTRACT: Current trends dictate increasing complexity and component counts on supercomputers and mainstream commercial systems alike. This trend exposes weaknesses in the underlying clustering infrastructure needed for continuous availability, maximizing utilization, and efficient administration of such systems. =0A=0AWe propose a step to alleviate this problem by providing a highly scalable clustering infrastructure, based on peer-to-peer technologies, for supporting resiliency-aware applications as well as efficient monitoring and load balancing. =0A=0ASupported services include Membership, supporting resiliency-aware applications, publish-subscribe messaging, Convergecast supporting load balancing schemes, attributes dissemination for propagating slowly changing information, and DHT facilitating services discovery and integration.=0A=0AWe employ a flexible two layer hierarchical topology, comprised of base zones federated by a management zone. Our design aspires to quickly identify membership changes in a scalable environment with minimal overall system disruption, thus enabling efficient Exascale size deployments. We present experimental evaluation taken from an IBM BlueGene/P, demonstrating scalability up to ~256K nodes.
SUMMARY:Scalable Infrastructure to Support Supercomputer Resiliency-Aware Applications and Load Balancing
PRIORITY:3
END:VEVENT
END:VCALENDAR