the storage anarchist

1.032: xiv closeout - buy one, get one free!

Is XIV destined to be the Ginsu knife of storage?

An email came across my desk yesterday noting a competitive situation where IBM was bidding XIV against a CLARiiON CX4. The XIV account team must have known their backs were against the wall, what with the better capacity utilization and power efficiency of the CLARiiON.

Not to mention a few other XIV deficiencies that I've pointed out before.

But things must be getting really desperate over there in Moshe-land, because this deal offered the customer TWO (2) XIV gen 2 arrays for the price of ONE!

Just in time for Christmas!

Now, I'll admit that IBM has come up with a creative way to get around the incredibly high probability of data loss with XIV's RAID-X architecture...simply mirror everything on XIV "A" over to XIV "B", and then whenever "A" suffers the data loss from the inevitable double-drive-in-two-separate-drive-modules failure, you can simply swap over to the intact data on "B" and carry on.

Of course, that does mean you'd be using 360TB of raw storage to get only 80TB of usable storage. Heck, even NetApp offers better utilization than that! (not much better, mind you, but better than 22.2%). Not to mention that you'd be paying for something like FOUR TIMES the power you'd need for 80TB usable of CX4 storage. Like I said before, you have to ask yourself: How much does a "free" XIV really cost?

But hey, if that's what it takes to get an XIV up to a respectable 4 or 5 9's of availability, who am I to quibble?

Now, to be fair, I'm not sure if IBM is offering these deals in the US - the bid I saw was admittedly out of China. Why there and not in the US or EMEA? Maybe IBM is interested in helping out the ailing Chinese economy, instead of its own back yard here in the US. I've also heard that the factory output of working XIV boxes is rather limited, and that for some reason or another, IBM hasn't been able to build enough systems to sell one to EMC.

So maybe this isn't an economic conspiracy on IBM's part. Could it be that the abject lack of any traction in the US has pushed IBM into an inventory sell-off in China? With all those XIV demo systems being returned by the banks and telcos as they tighten their belts, what better to do than dump them on the poor, unsuspecting Chinese? 
 

1.031: meet the symmerator

Ever wonder what happens to old Symmetrix arrays?

Frequently customers will "cascade" them as they age: as new Symms are purchased for production deployment, the older arrays that they replace often become the BC/DR "target" at the other end of the SRDF link.

And the old SRDF target? Well, some simply get returned at the end of the lease term, while others get traded in for credit towards a shiny new Symmetrix.

Some old Symms, of course, meet with a somewhat less fortunate demise:

How sad....

But at least one old Symmetrix 8000-series array has found happiness in a totally new retirement career...
 

1.030: flash as cache - really?

Over the past week or so, Robin Harris, Chuck Hollis and Stephen Foskett each discussed the "appropriate" use of flash technology going forward. Chuck comes down pretty solidly in the "best as persistent storage" camp, while Robin seems more aligned with FusionIO and the "flash as cache" side of the argument, while Stephen seems content to accept that flash will appear at both the initiator and the target sides of the I/O conversation.

I myself tend to agree with Stephen.

BUT!

(You know there's always a BUT! with me).

A few things have been nagging at me about this whole flash-as-cache discourse. The first is centered around the fact that it takes longer to WRITE a block to NAND flash than it does to READ it.

Unlike traditional SDRAM where reads and writes complete at the same speed, with NAND even if your flash controller is smart enough to asynchronously pre-erase blocks, it still takes longer to perform a write than a read. And if you take the time to verify the accuracy of the write, it gets even worse.

So my question is, since it takes longer to write than to read a NAND block, and every read hit required at least one prior write:

What read hit ratios and repetitive reads of a block
are required to overcome the NAND write penalty?

inquiring minds want to know...

1.029: atmos. with, and without, the sphere

I'm just back from vacation cruising several Italian, French and Spanish ports aboard the Wind Star on the Mediterranean with my wife. It was a relaxing, multi-cultural Adventures Afloat trip arranged by her employer (Elderhostel), a not-for-profit who specialize in educational travel and learning opportunities. With a foundational belief that learning is an integral part of a healthy and fulfilling life, the organization offers its unique Adventures in Lifelong Learning to anyone who is interested - at an exceptional value! So, if you're looking for a travel programme with more than just the usual tourist trap visits, I encourage you to visit their web site and/or order their free catalog.

Oh, and don't let the name fool you: participation is quite diverse, and you'd better be in good shape or you might just get left behind.

Anyway, being on such a trip with my wife, I wisely avoided all things work for the duration.

Preserving the atmosphere, you might say.

But so much has gone on in the past couple of weeks, I thought I'd take a stab at connecting some of the key sights from my cruise with a few of the more interesting events of the past week or so.

So let's have a little fun. Shall we?