Search

The problem of garbage collection in a single-failure fault-tolerant home-based lazy release consistency (HLRC) distributed shared memory (DSM) system, based on independent checkpointing and logging, is discussed. A solution is proposed that uses laziness in garbage collection and exploits consistency constraints of the HLRC memory model for low overhead and scalability.

This article explores the topic of recovery-oriented computing. In this article, the authors address fast recovery, even when a machine's operating system crashes or hangs, through a novel exploitation of backdoor hardware mechanisms intended for system management operations. The authors developed an approach designed to recover session state from a crashed machine and pass it to another node to continue service sessions seamlessly from the user's perspectives. Their lightweight, application-independent system provides both failure detection and recovery, for use with complex, multi-tier Internet services. The core of the system is the Backdoors (BD) architecture, which uses commodity programmable network interface cards with specialized firmware and operating system extensions to provide remote access to lightweight application and operating system state in a machine's memory without relying on its operating system or processors. Using BD, machines in an Internet server cluster can cooperatively observe each other's health, detect failures, and take over client sessions from failed nodes. In this article, the authors describe the BD architecture and their operating system extensions for monitoring and recovery of service sessions. INSETS: Backdoors Architecture;Log-Based Continuation Box Synchronization.