You may remember my recent piece discussing how you can use IBM Spectrum Protect’s new Stgpool Conversion tool to extract longer life out of your legacy storage. If you don’t remember it, then I recommend you go here to see what the fuss is about.
Following on from this, I retired to the laboratory, rearranged the test tubes, and set to thinking about what new features of Spectrum Protect I should delve into next.
One of the more recent developments has been the ability to store data in an object storage pool, located in the public cloud. This originally allowed the use of Object Storage in IBM Softlayer (recently renamed Bluemix), but the latest release adds support for S3 storage provided by Amazon Web Services (AWS).
This capability can free customers who are suffering from the seemingly endless storage capacity paradox. The more storage you have, the more storage you need.
Why not move towards a usage based storage platform, and let somebody else worry about the provisioning? This is far more efficient than over provisioning to accommodate storage growth for the next three or five years.
In addition, depending on the architecture of that storage platform, you may have also managed to remove the question about data off-siting. There’s no need to dump a copy of the data to tape to store offsite, if it’s already being kept offsite by your cloud partner. This capability forms the basis of many Disaster Recovery-as-a-service (DRaaS) solutions.
My first hunch was that customers who are considering using cloud storage will already have a preference and it might not be based around performance or cost. For some customers, this decision will be based purely on the vendor that you already have a working relationship with. That’s valid. A customer who is already using Amazon storage is going to need a compelling reason to add IBM Cloud Object Storage into the mix.
“But,” I hypothesized, “For customers who have not yet decided on a cloud object storage platform, wouldn’t it be interesting to see if one of the vendors performs better than the other? That would help people to make a more informed choice.”
I set to. Armed with a couple of Spectrum Protect servers and a blissfully unpopulated spreadsheet, I started sending workloads to the various clouds, testing different scenarios and recording data.
I need to carry out some more tests to confirm my results. So far, there’s not a lot to choose between them although one has been performing more effectively than the other in most of the tests. I hope to have the results published within the next month, but it will leave me with a gap in my schedule, and once I’ve tidied up the Bunsen burner, I will be looking again for inspiration.
Have you already made the jump to cloud storage? If so, I’d be interested to know if your experiences are similar to mine. Leave a comment to let me know about your cloud storage experiences.