Stop buying storage from EMC!
Stop buying storage from NetApp!
Stop buying storage from HP!
Stop buying storage from HDS!
Stop buying storage from IBM!
Stop buying storage from Dell!
Stop buying storage from Oracle!
Start buying storage from PC World!
Start buying storage from Best Buy!
Start buying storage from Ebay!
Start buying storage from the scruffy PC shop on the high street!
[there are other storage vendors available]
Why? To be honest, it seems that most people would better doing this than buying storage from any vendor who might have added some useful features in the past five years. Because some of the conversations I’ve had recently suggest that people still don’t trust:
Deduplication
Thin provisioning
Wide Striping
Automated Storage Tiering
Compression
Snapshots etc, etc, etc
All of these features are useful and effective; if you are not using them, you might be better simply buying hard-disks and directly attaching them. It’s time to move on and get with the program.
Now all of these features do need some thought and planning to get the best out of them but the really great thing is the vendors have worked really hard to make their arrays easier to configure and manage. This gives you time to start thinking about how best to use these new features, stop being a LUN-monkey and add value to your organisation.
Your life will be more interesting and your work will become less of repetitive, soul-sapping drain.
the sad fact remains that storage vendors are still charging around 7k per TB for the use of their “secret sauce”
that cost model is not sustainable, and as contrarian companies such as xiotech, nimble, etc. come in with competent systems that perform just as well if not better than the big boys, those in the storage realm that want value and performance will simply stop buying from the far too expensive array vendors.
I don’t especially care who you buy your storage arrays from but it’s time to move on and use the features provided. And if you are paying for the special sauce, don’t just scrape it off and leave it on the side.
We had to turn down two vendors because the “secret sauce” cost about what you described there and we weren’t in a position to use those features. (Admittedly, we only had about 20 TB of requirement, and it was mostly for QA VMs, so it didn’t need to be the most stable.) The software cost more than the disk, thanks for breaking it down for us vendor! In the end, we choose a small shop with a virtualized disk array that cost 1/3 of the others (http://www.amsstorage.com/html/tigerhawk.html)
Buying commodity hard drives from the corner shop doesn’t work at scale.
Then people say… oh but that’s what Facebook and Google do (commodity disks)… yes but they also have lots of clever programmers on staff to integrate them into something that looks like a product.
What sort of product? Oh that’s right.. a product like the ones the big boys sell.
The real issue holding clients back is fear… clients who actually dive in and start using thin provisioning or wide striping or snapshots or automated data placement almost universally find great business value in it (yes… business value), but they need good IT people and good business partners and good system integraters to help them use it. Mmm hang on.. those three categories are not going to be found at the corner shop either.
Days of hardware resource silos are over. Whether it is compute or storage. All the features mentioned are best implemented in software which is agnostic to the underlying hardware.
Parallel databases & scale out filesystems + commodity hardware = The future.
Storage vendors & virtualisation(VMware) vendors hey days are numbered.
Hi Martin
It’s not just storage. There are a raft of other examples where IT shops don’t exploit features that would make their lives so much easier.
I cover Service Providers for EMC across Europe from a technology perspective and you wouldn’t believe the number of organisations that I meet that haven’t significantly changed their operating practices by introducing new technology features in the last 5 years. A few examples not on your list:
Tapeless backup
RISC to commodity CISC transition
Virtualisation of large scale enterprise applications
Automation and/or Orchestration of repetitive IT tasks
Migration of management functions (backup, performance management, security) into the hypervisor
I have my own views as to why some of these (plus the storage features you mentioned above) don’t get adopted, and I’m exactly not whiter than white myself, having being a technology laggard on some of the ones I’ve mentioned in prior IT operations leadership roles…
Ah! Finally someone who’s getting it! In any case, the end result is going to be much less complicated, much less expensive, much more reliable and (yes!) less power hungry! If you need something to attach these plain vanilla SATA disks to, I suggest plain vanilla x86 hardware boxes, connected by plain vanilla gigabit ethernet. Tied together by smart distributed software. NO provisioning. NO install. Near-zero administration.
That’s how Google does it. That’s how we do it (shameless plug). One order of magnitude lower TCO. Two orders of magnitude lower misery.
Can anyone explain why this seems so hard a message to understand?