5TB drives

I just read this on The Register. 5TB drives! Can you imagine it! The HDD manufacturers continue to push the envelope even further.

Now I have a concern about drives getting to this size and that's the ability to get data on/off the drive itself. With 73/146/300GB drives, the capacity to response time ratio is still within a tolerance that means adequate random access throughput can be achieved. But with larger drives the number of different concurrent accesses will increase and if response time doesn't decrease then very large HDDs will start to operate like sequential devices.

I think I need an illustration to make my point. Imagine a 73GB drive is receiving 200 random I/Os per second, each with an average 5ms response time. Scale the capacity up to a 5TB drive and that's about 69 times the capacity. The scaled up drive would have to cope with 13800 I/Os a second and provide an average response time of 0.07ms!

Firstly, it is unlikely 5TB drives will be expected to perform like today's 73GB drives but it serves to illustrate that we can't expect to simply consolidate and shrink the number of drives installed into an array. We need something more.

I think we need a more innovative approach to the design of the drive interface. This may simply be shed loads of cache, to improve the overall average response time, or perhaps multiple virtual interfaces per drive or independently mobile read/write heads which don't need to read/write a cylinder at the same time. It could even be drives that dynamically reallocate their data to make read/write quicker (for example, put frequently read/write blocks in the same physical area of the drive).

Who knows what the solution is, but rest assured something needs to happen to make 5TB drives useful devices.

USP-VM

Hitachi has announced (10th September) the availability of a new storage array, the USP-VM. At first glance this appears to be the USP-V equivalent of the NSC55 as it has very similar characteristics in terms of cache cards, FEPs etc. Unfortunately HDS have provided links to specification pages not all of which include the USP-VM references. Bit sloppy that.

I've previously deployed a number of NSC55's and within 6 months wondered whether they were the right decision. They weren't as scalable as I needed and there were a few features (such as BED ports and FED ports sharing the same interface card) which were a bit of a concern (imagine losing a FEP and having to take 1/2 of all your BE paths offline to fix the problem). I'm always reminded of the DMX1000/2000/3000 range when I think of the NSC model as these EMC arrays weren't expandable and of course a DMX1000 quickly filled up....

Hu describes the USP-VM as "Enterprise Modular" in his blog entry. This may be a bit generous as (a) I doubt the USP-VM will be priced as low as modular storage and (b) I don't think it will support the whole range of disks available in a modular array. I say "think" as the link to the capacity page for the USP products doesn't yet include the USP-VM.....