Why hybrid storage is DOA

My prediction: hybrid storage is dying.

The cost delta between disk and flash is rapidly diminishing. The cost curve on enterprise-grade flash is much steeper than disk, with prices set to decline at 40% per year for flash, versus 13% for disk. The value that hybrid storage vendors are capturing today will decline, and so will their market segments.

flash cost curve

Flash (SSD) storage becoming cheaper relative to disk (HDD)

In 3-5 years, I don’t think anyone is going to be buying new disk (hybrid or otherwise) for high performance workloads; and I think 10K drives will be entirely absent from datacenter within 10 years. The cost curve on electrical performance (NAND) is much steeper than mechanical performance (disk). You’ll still have scale-out disk.

Holy Shit: AWS hits $8B run-rate

Bubble or not, you can’t ignore a $8B company growing at 78% yoy and generating 25% operating margins (cash flow is lower given capital expenditure requirements).

AWS, Azure, and Google Cloud 2015 Q3 earnings results.

AWS, Azure, and Google Cloud 2015 Q3 earnings results.

And what happened to developers’ love of Google? It’s a good platform with strong data infrastructure tools (databases, data analytics). But Amazon running away with fickle developer mindshare, and clearly not catering as lowest cost vendor anymore.

I think Azure is different market segment, but also doing well.