Design, implementation and evaluation of ICARE: an efficient recoverable DSM

Abstract

SUMMARYIn the light of the increasing throughput of local area networks, Networks Of Workstations (NOWs) which provide a Distributed Shared Memory (DSM) have become a convenient and cheaper alternative to parallel architectures in the framework of parallel scientific applications. However, the probability that a failure occurs in such a system made up of a large number of components must not be neglected, especially for long-running applications. This paper presents the design, implementation and performance evaluation of ICARE, a page-based recoverable DSM implemented on top of an ATM-based NOW running the CHORUS microkernel. ICARE relies on a Backward Error Recovery (BER) mechanism, and provides a way to combine both efficiency and high-availability. The fact that checkpoints are stored in volatile memory provides a low-cost fault-tolerance mechanism, as well as the opportunity to exploit the symbiotic relationship between the data replication implemented in DSM systems and that needed for fault-tolerance. Furthermore, ICARE efficiently implements transparent process rollback recovery. Performance evaluations show the efficiency of the ICARE prototype that implements the proposed algorithms.

Fault-tolerant distributed shared memory on a broadcast-based architecture

“…Care must be taken that the checkpoint is saved to a storage medium that can be accessed after a failure occurs. In [2], checkpoints are saved in the memory of another node rather than to disk thereby reducing the amount of time that is required to save the checkpoint. We have adopted this approach for our simulations.…”

“…It is not necessary to recreate an entire copy of the memory at each checkpoint interval. Instead the checkpoints can be incrementally updated since it is only necessary to save the data that has been modified since the last checkpoint [2]. After the processors are synchronized in preparation for taking a checkpoint, all of the cache controllers write back any exclusively owned cache blocks.…”

“…Much research is currently being conducted in order to exploit the data replication mechanism already existing in a DSM to reduce or hide the overhead of implementing BER. An example is the ICARE system [2], a software-based recoverable DSM for a network of workstations.…”

Protocols for fault-tolerant distributed-shared-memory on the SOME-bus multiprocessor architecture

“…Saving large amounts of data to disk can be extremely time consuming. Another approach that has been proposed [3] is to utilize the memory of another node to store checkpoints. This approach provides the necessary requirement to tolerate a single node failure, or multiple node failures where all copies of the checkpoint do not reside on the failed nodes.…”

“…Much research is currently being conducted to use these mechanisms for cache coherence for the benefit of fault tolerance in the form of checkpointing [3][4] [5].…”