Strategic Moves

State-of-the-art high-availability and disaster-recovery solutions are the product of a 10-year strategy at IBM.

State-of-the-art high-availability and disaster-recovery solutions are the product of a 10-year strategy at IBM.
Illustration by Chris Gall

It was more than a decade ago that we sat in a conference room with Bob Gintowt discussing and debating future development investment priorities in the area of availability and recovery. We’d had a fair amount of success with our foray into the world of horizontal capacity growth (a form of clustering) with the OptiConnect solution. It was clear that IBM* processor technology would enable massive server consolidation, which in turn would drive escalating requirements for robust high-availability (HA) solutions. IBM’s larger clients were asking for a “hardware” solution for HA, and the Internet hadn’t even taken off yet. Long story short, IBM agreed on a plan to develop an overall investment strategy for HA that included clustering, and we were off to the races with Snyder taking the lead. Each subsequent release of the operating system introduced building blocks that would eventually enable IBM to deliver a complete solution. The announcements made with IBM i5/OS* V6R1 represent a major part of achieving that goal. That’s not to say IBM’s finished innovating, but the company has arrived at a point where IBM customers can select from a complete set of solution options to address their HA and disaster-recovery (D/R) needs.

IBM’s strategy to build that oft-requested end-to-end solution began with independent auxiliary storage pool (IASP)-based clustering technology. It’s estimated that more than 500 customers around the world have deployed that pre-V6R1 solution. More advancements followed.

Some of IBM’s largest customers moved to the Advanced Copy Services Tool Kit, a Lab Services offering that combines the i5/OS IASP and clustering technologies with the IBM storage-server technology. Customers using a storage-area network (SAN) storage strategy appreciate this approach because it addresses all of their availability needs with minimal day-to-day administrative activity while letting them fully capitalize on their SAN server investment.

Others adpoted cross-site mirroring (XSM). XSM capability is provided by the native i5/OS geographic mirroring or by IBM DS8000* (and IBM DS6000*) storage replication technologies Metro Mirror and Global Mirror. Both geographic mirroring and Metro Mirror are synchronous replications solutions, which is the optimal approach for HA solutions.

Customers continued to adopt new HA technologies piece by piece. Then came V6R1 and the announcement of the High Availability Solutions Manager (HASM). For more information on HASM, see “Announcing IBM High Availability”.

HASM enables i5/OS-based management of an integrated storage cluster or a SAN server-based cluster (see Figure 1, below). V6R1 availability capabilities include:

  • Simplified cluster management with a Web-based browser
  • Synchronization of most non-IASP-capable objects via administrative domain
  • iCluster software replication for geographic dispersion and/or data replication/recovery
  • Metro Mirror and Global Mirror integrated as an extension of the i5/OS HASM/XSM for DS8000 and DS6000 technology-based solutions
  • Significant Fibre Channel performance and capacity improvements that put DS8000 performance on par with i5/OS native disk-storage deployments

If you’re looking for geographic dispersion for a D/R solution, the IBM DS8000 Global Mirror replication solution may fit. Another D/R approach is to use iCluster, the IBM software solution for HA and D/R. iCluster is one of the well-known logical replication solutions well suited for geographic-dispersion implementations as well as for data replication and recovery operations. With HASM, iCluster and DS8000, IBM can now offer the i5/OS customer a full menu of HA and D/R solution choices (Figure 2). Ten years of vision and work were paying off.

Mike Snyder was a senior technical staff member in System i development at IBM.

Steve Finnes is the System i Business Continuity product manager with IBM. Steve can be reached at

comments powered by Disqus



2019 Solutions Edition

A Comprehensive Online Buyer's Guide to Solutions, Services and Education.

Achieving a Resilient Data Center

Former IBM exec Bob Angell shares his insights on improving data-center resiliency

Achieving HA

Using cluster resource groups for high availability

Announcing IBM High Availability

Customers have asked for an end-to-end high-availability solution from IBM, and it delivers with HASM.

IBM Systems Magazine Subscribe Box Read Now Link Subscribe Now Link iPad App Google Play Store
IBMi News Sign Up Today! Past News Letters