I end up hearing many storage administrators vent about old data. In some cases, it’s old HSM data. In other cases, it’s data that should be deleted but out of an abundance of caution no one actually deletes it. The only option has been to buy more FICON-connected disk or tape, which fills up HSM or backup environments that are already full.
However, two brand-new IBM products are changing the rules:
These products offer five new targets to send your old data to cloud object storage. The first product basically creates an HSM ML3 in the cloud for you, and the second clones traditional tape IO to cloud.
This is where Hybrid Cloud is going to start making sense. All storage boils down to speed and cost, and because network speeds (and costs) are permitting enterprise levels of storage transfer, the mainframe storage world is opening up. The days of choosing only between disk and tape are ending. With a Hybrid Cloud strategy, you can design a storage footprint that meets your exact needs of speed and cost, blending public and private cloud storage options of all types. Obviously, this doesn’t make sense for OLTP data, but the sweet spot for moving old data to cloud has completely opened up.
We’re not just talking about one vendor or one class of storage here. Suppose I needed to have a legacy application that writes tape send the data to a public cloud S3 bucket? Done. Or, suppose I wanted to design a new disaster recovery scenario that exploits IBM Cloud Object Storage to make my data recoverable at several sites? Or, suppose my IT shop has their own HDS HCP infrastructure? I can reliably move my old data to be someone else’s problem, while maintaining control!
Thanks to Hybrid Cloud, the mainframe data world is going to be a much more interesting place very soon.