It's surely axiomatic: Oracle will replace the NetApp E Series-sourced 6000 line of storage arrays with the Pillar Axiom. Oracle dislikes OEM supply deals, seeing no good reason to send part of a customer's purchase money to another supplier. The company has bought the Pillar Data startup, which was funded by Oracle CEO Larry …
Call it a Gruntmaster 6000
Dilbert: What kind of product do you see when you imagine a Gruntmaster 6000?
Dilmom: Well it's a stripped-down version of the Gruntmaster 9000 of course. But it's software-upgradeable.
Defeat from the jaws of victory
Sun had a glorious opportunity with the 7000 series, they had a stable OS to base a product on, extensive hardware support and the amazing features of ZFS (file/block checksums, snapshots, ease of adding de-dupe) and they put them together in a product line that promised to beat all comers. Reasonably priced (1TB SATA HDD selling for ~£300, not the £1k+ of others) and with all features included (no license surprises to *use* them, nor to keep them if you expanded the storage size).
And what happened? They screwed up by releasing a product in the 'prototype stage' and not something that was usable for serious deployment. And Oracle has done nothing worthwhile to improve matters since taking over.
Critically the controller 'appliance' software is an utter mess, and years later it is still flaky and serious issues have been ignored. We would not buy it again.
And Pillar, what USP do they have? Smart data deployment on HDD on a usage-basis ('hot' data at disk edge for rapid access, etc) was their selling point, and at one time intelligent reps (who they sacked, just before we tendered, so they screwed up on our chance).
But that is not enough to make it a great product, and more so as flash takes over and the data tiring strategy game has changed. Maybe if Oracle were to combine the best features and fix the ungodly mess of the 7000 series appliance they would have a serious contended,
Oh wait - they dropped SATA disk trays in favour of SAS due to higher selling margins, thus screwing folk like us that have a usage pattern best dealt with by TB-sized HDD.
I Tend to agree...
....the ZFS Appliances did hold a lot of promise but the engineering execution failed to deliver a stable and predictable product. The ones we deployed experienced a lot of downtime and unpredictable behavior making you cross your fingers every time you did anything especially controller fail-over or firmware upgrades.
Nexenta has done a good job at reselling the ZFS appliance based on their own implementation and I've only seen good reviews. I have no personal experience with the ZFS appliance. Based on talks with folks that have, it sounds like the first release was pushed too quickly to market with out enough testing and documentation (for the field reps), with later releases fixing those issues.
We use a mixture of the rebranded HDS arrays, the StorageTek 6000 discussed in this article, and netapps. Never had an issue with the first two. The netapps love to go down. It's so bad, we make the netapp folks come in to do servicing (ie: installs, and fw upgrades) which even they as the makers of the system have problems when doing. Each array will do a job better than the other, and we each will have slightly different stories of how it worked for us. All I want from storage for it to work, not corrupt data, and be fast. Doesn't seem like much to ask.
As far as the 6000 line. Oracle will probably drop it. With how storage is going, you'll either be ok with the server's internal capacity and direct attached, or need something larger that is possibly shared. If Oracle can bill an Axiom array as the later, and tie it with their product line, I'm sure it will sell. I also suspect IBM will also drop their rebrand of the Engenio line at some point.
NetApp likely to keep a foot in the door ....
While you may be right about the eventual fate of the Storagetek/Engenio 6xxx line (as well as the IBM/Engenio DS 5xxx), I doubt either Oracle or IBM plan to drop the more SMB oriented (NetApp/Engenio) Storagetek 2xxx or IBM DS 3xxx line (this has always been the only product line Dell takes from NetApp/Engenio). The margins aren't big enough.
Slow storage news day?
I don't imagine there is much worry in NetApp considering the overall Sun external storage market share position on any of its storage platforms. Sun 7000 included. I'd imagine there was more of the HDS rebadge and old Sun T3's out there than anything.
And the concept that anyone is going to start buying Pillar in enough volume to change the world is just laughable. Out of curiosity has anyone here actually paid cash money for one - I would genuinely be interested? Apart from Oracle shareholders, and they got the company...
Did Oracle's shareholders just pay off Larry's bad investment?
Makes Hurd's alledged fiddling of the expenses look trivial in comparison!
Matt, I can tell from all your post against Oracle that you love slagging off, and you can accuse Oracle of a lot of things but making bad business decisions aint one of them.
There are few companies that can compete with Oracle in making the right commercial decisions (commercially) over and over again.
So that really was lame attempt
RE: @Matt Bryant
"....There are few companies that can compete with Oracle in making the right commercial decisions (commercially) over and over again...." Ah, that did make me laugh! The continued decline in Snoreacle server sales just goes to show us customers don't believe Larry's blather about software and hardware being made as one as beeing somehow better. Having goofed in not offloading the Sun hardware bizz (and Larry did try), he's now swallowed his own bull and bought a failing storage manufacturer to go with the failing server one! He should have bought NetApp, but then that would have not allowed him to recover his lost investment at the expense of the Oracle shareholders.
Bricks and Mortar
Or, should we call that Bricks and Slammers?
The Axiom's birth comes from Speakman's vision and Ellison's pocket, so the Pillar "acquisition" was logical.
Now it's payback time, Larry is going to advocate Oracle software on Oracle Servers and Oracle storage.
The Pillar product is decent, sure it may not have the years of trust we've put in NetApp, EMC, HDS and [various rebranded] LSI products, but the Pillar, oops Orcale Axiom Slammer + Brick architecture makes it a good modular alternative to the usual storage suspects
So, Larry's vision is a happy family of products to keep your infrastructure maximised and your choices minimised, while making Oracle shareholders believe everything is just peachy