Method and apparatus for write-back caching with minimal interrupts

作者: Kit M. Chow , Michael W. Meyer , Gary L. Boggs , Keith P. Muller

DOI:

关键词: Computer hardwareComputer data storageVolatile memoryFloppy diskNode (networking)Protocol (object-oriented programming)InterruptComputer networkComputer science

摘要: A method and apparatus for write-back caching in a data storage processing system has been described. The comprises the steps of receiving write request including from compute node first I/O node, forwarding to second sending an acknowledgment message after is received by node. After written into non-volatile purge or command sent volatile memory In one embodiment, not until receives request, which case, same interrupt as request. each with means that other Each also back directly, without through data. result protocol reduces number interrupts required store data, while still implementing improve speed turnaround. invention can be described terms program device, such hard disk, floppy CD, tangibly embodies instructions stored thereon performing practice invention.

参考文章(15)
James J. Wolff, Distributed I/O store ,(1998)
Stephen J. Sicola, Susan G. Elkington, Richard F. Lary, Paul Guttormson, Michael D. Walker, System for data replication using redundant pairs of storage controllers, fibre channel fabrics and links therebetween ,(2000)
Bruno Alterescu, Natan Vishlitzky, Daniel Castel, Moshe Yanai, Gadi Shklarsky, Yuval Ofek, Asynchronous remote data mirroring system ,(2000)
Roger Alan Fleming, Fault-tolerant processing method ,(1997)
Thomas F. Fava, Randy Fuller, Joseph M. Keith, Simultaneous, mirror write cache ,(1997)