A Recovery Model for Survivable Distributed Systems through the Use of Virtualization

The interest in system survivability under a variety of attacks, faults or accidents has been on the rise in recent years and survivability is receiving increasing attention as a key property of the mission critical system. Distributed systems are vital for all and it is a critical task to make such systems secure and survive. Survivability in the presence of attacks or failures depends on three key system capabilities: resistance, recognition, and recovery. Virtualization at levels (system, storage, and network) became important as a way to improve the system security, reliability, availability and survivability, reduce costs, and provide greater flexibility. In this paper, we focus attention on continued service and recovery issue. We introduce a survivability framework for distributed systems through the use of virtualization technology and software rejuvenation methodology. We present a recovery model and evaluate the steady-state system availability and survivability based on the familiar Markovian analysis through SHARPE tools.