LCU count mitigation

As we look ahead to future releases of the DS8000 product line, we anticipate that continued increases in both physical capacity and I/O capability will eventually outpace the subsystem’s ability to scale volume addressing under the 255 LCU limitation. With each new hardware generation delivering higher throughput and greater concurrency, the constraint may no longer be driven solely by capacity growth, but also by the need to sustain higher I/O densities within the same LCU ceiling.

In environments such as ours—where subsystems are configured with separate LCUs for Safeguarded Copy (SGC) recovery volumes and each production LCU is twice as large because they contain Global Mirror (GM) journals—the effective LCU consumption is amplified. Specifically, each production LCU requires three LCUs to support these functions.

One potential solution is to eliminate the requirement that Global Mirror journals be defined as CKD volumes and instead allocate GM journal space from a managed pool, similar to the approach used by Safeguarded Copy. Direct access to GM journal volumes is not a requirement for our use cases. While pooled space management could present concerns for some customers, this challenge could be mitigated through a model analogous to SGC, such as the use of a configurable “journal multiplier” to ensure adequate space allocation.

This would ideally make it into the current DS8A00 generation to support our needs