Abstract :
Computer systems frequently store data that are critical to the organization that owns the system. Any unplanned breach of service can have disastrous effects. The system cannot be expected to continue operating after fires, floods, and earthquakes. Upgrades to air conditioning, power, and basic system software may all require that a system be taken out of service temporarily. To support availability despite these outages, Tandem has introduced the Replicated Data Facility (RDF). RDF extends the component replication strategy to allow a redundant copy of an entire system. Since the second system typically is at a remote site, it will generally be unimpaired. RDF allows an organization to maintain a geographically remote backup system with an up-to-date copy of the database. Should the first system fail, this second system can rapidly take over the workload, minimizing downtime. RDF is an extension of Tandem´s Transaction Monitoring Facility (TMF). TMF provides transaction atomicity and transaction, system, and media recovery. TMF supports distributed transactions in a network of systems using a presumed-abort two-phase commit protocol.<>
Keywords :
computer facilities; computer installation; disasters; distributed databases; fault tolerant computing; redundancy; transaction processing; Replicated Data Facility; Tandem; Transaction Monitoring Facility; air conditioning; basic system software; breach of service; distributed transactions; downtime; geographically remote backup system; presumed-abort two-phase commit protocol; remote site; up-to-date copy; Air conditioning; Condition monitoring; Earthquakes; Fires; Floods; Protocols; Remote monitoring; Resource description framework; System software; Transaction databases;
