Enabling continued operation of IT services and infrastructures during floods and other disasters
Prapaporn Rattanatamrong, Mauricio Tsugawa, Jose A.B. Fortes, Andrea Ammatsun
Abstract
Various organizations increasingly rely on information technology (IT) services and infrastructures for their normal operation and
management. Collections of data from various media including the Internet and interviews with related personnel from the
organizations affected by the 2011 Thailand flood reveal the flood’s impact on severely disrupting IT services and infrastructures.
Enabling continued functionalities of such services and infrastructures during any disastrous events, such as floods or earthquakes,
then becomes crucial in controlling the impact of these events to people’s life and their communities. This project focuses on
leveraging Internet Data Centers (IDCs) as disaster recovery sites, where government and corporate data can be backed up and
organizational services residing in virtual machines (VMs) can be temporarily relocated in order to provide high-availability and
resiliency. From studying the nature of the organizations’ IT services and infrastructures, it is evident that exploiting VM migration to
enable service continuity during disasters is plausible; supporting infrastructures, technologies and virtualization in the studied
organizations exist and are sufficient. The practicality and scalability of existing VM live migration mechanisms in wide area network
setting are currently being studied through software simulation. In addition, we present a formulation of scheduling VM migration
during floods and other disasters; these include selecting which VMs have to be migrated, deciding to where they should be
reallocated, and determining the sequence in which VMs are migrated such that performance and migration cost requirements are
met in the available time frame. This preliminary formulation will then be developed further into a multi-objective optimization problem
formulation whose solutions can be provided using a Genetic algorithm based heuristic approach.