IBM Won't Open Up XSM-Replicated Data in i5/OS V5R5
Published: September 25, 2006
by Alex Woodie
Flexibility has been a drawback of IBM's advanced Cross Site Mirroring (XSM) high availability technology. According to IBM officials at the COMMON conference last week, XSM will continue to be limited in its usefulness with the next release of the operating system, which will be called i5/OS V5R5.
XSM, which IBM introduced with OS/400 V5R3 and enhanced with i5/OS V5R4, is a form of clustering used to mirror data and objects from one iSeries or i5 system to another. The technology, which leverages independent auxiliary storage pools (iASPs), has been heralded by OS/400 high availability experts for its capability to shield users and applications from the complexity of clustering for high availability. Since it runs in hardware, it "just works," as the thinking goes.
However, XSM is limited in usefulness because mirrored data can only be used for disaster recovery. This is different than traditional mirroring or clustering, where the second set of data can be used for other tasks, such as for running queries or business intelligence workloads. IBM's Steve Finnes, who is business resiliency manager for the System i line, did not say whether this was a technological or business decision.
Regardless, this will have several impacts on iSeries and i5 users and solution providers. First, it will ensure that XSM remains a technology used only by the biggest shops, which can afford to keep their iSeries or i5 resources standing idle in case a disaster strikes. It will also protect IBM's high availability business partners from a potentially disrupting technology. IBM's stated goal is to include more automatic capabilities, which XSM surely is, into its operating systems. In the long run, this benefits the entire ecosystem, even if it disrupts the status quo in the short term.
XSM and Clustering: The Future of OS/400 High Availability