The Wikibon consultancy says a developing Server SAN form of storage needs to be tracked by IT people. The authors bring together the concepts of virtual SANs and converged server/storage hardware and software systems with a server SAN being a collection of servers whose direct-attached storage (DAS) is virtualised by software …
that first sentence was bloody hard to follow.
Like blowing up a balloon and not stopping until it went bang
That guy must be japanese!
Reminds me of the good old times when students unsullied by technology asked me to please indicate the whereabouts of the server's office.
And I thought I come from the IT sector
This is the first IT related article I do not even want to understand.
So basically their suggestion for the future of the computing industry is to create a "system" which contains all of the elements currently in use in the industry into a single unit, and comes preloaded with appropriate software.
I wonder how much they got paid to come up with that?
"The system will understand performance characteristics such that there do not need to be knobs or an understanding of RAID types."
I thought the storage admins were the knobs? Looking forward to getting rid of them....
Aspirations, but not much else
This deserves a new category beyond mere vapourware, Perhaps aspirationware or some such term. It's essentially a lost of idealised requirements, almost all of which have been around for years. Any number of attempts have been made at abstracted, virtualised storage models, and it's proved to be an extremely difficult nut to crack. Merely chucking into the mix a few bits of technology which may, or may not be appropriate does not make for a solution to the problem. At the heart of things is the extremely difficult issue of producing a generalised, robust configuration model which addresses the relationship between storage and application and how it's accessed. At the moment, there are multiple paradigms and levels of abstraction of entities from blocks up to objects, and it's not clear that there's any obvious single model.
When there is a product in the offing, then let us know. As it is, this is not much more than another hyped buzz-phrase,
Re: Aspirations, but not much else
Steven Jones, you are absolutely right. We need a neologism for IT related neologism. If only we could have a way to suggest some, and then vote to elect the ultimate one!
Didn't EMC already buy this?
Sounds almost exactly like an Isilon to me, fits all of the requirements of the wikibon people, perhaps they should look into it?
Here's a tip, Wikibon folks: color is good -- when it clarifies things. That second chart just looks like someone threw up all over your first chart.
you're not all invited to the party
Chris - thank you for sharing our research and continuing the conversation.
First of all, we would say that while VSAs (and derivatives of them) have a place in the market, they do not fit into our Server SAN definition. They don't meet the very-scalable architecture requirement, nor the performance (able to utilize flash) characteristics that we set forth for this space.
There are plenty of scalable storage platforms (EqualLogic, Isilon and others) that also don't fit into this category since this category includes compute as well as storage.
VSAs and other scalable architectures still have plenty of utility in the marketplace, they just don't fit in our new category. We appreciate all feedback and apologize if my choice of color didn't thrill.
Why do these people exist?
I have not commented in a long time, but this is the biggest load of fucking bullshit I have read in any IT publication. Generally Chris's Storage articles are pretty good, not quite up to the former TPM's *Nix ones but not far off.
Please tell me you wrote this up for a giggle? Most of it doesn't even make any sense!
Try a more credible source source than this pointless outfit (wikipedia them...) next time, or don't post the story at all.
Server + SAN in one = DAS + LVM?
Or at least that is what it sounds like to me - consolidating a SAN and a VM host. Maybe this is a "new paradigm" for the virtualization hype types, but having read the article twice, the only thing that is "new" is that the disks are pre-configured (handy for those only up to the task of accessing them by clicking on pretty icons in the GUI).
Is this seriously intended to be seen as fundamentally different to running a Xen/KVM VM host full of disks, running FreeNAS in dom0?
A distributed system where the directly attached storage on each server is abstracted and presented as a single resource to all of the servers to leverage without relying on a single SAN link?
I think they called that Hadoop.
After review, HP StoreVirtual VSA and Maxta have been added to the Wikibon report. A standard VMware VSA would fall outside of our criteria as mentioned above.
Missing the z axis.
Cost. The performance vs functionality comparison is almost pointless without the relative cost being considered.
What's the infographics counterpart to "plot vomit"?
So not unlike a *nix server with NFS then
- Leaked screenshots show next Windows kernel to be a perfect 10
- Amazon warming up 'cheapo web video' cannon to SINK Netflix
- Something for the Weekend, Sir? I need a password to BRAKE? What? No! STOP! Aaaargh!
- Episode 13 BOFH: WHERE did this 'fax-enabled' printer UPGRADE come from?
- Vulture at the Wheel Ford's B-Max: Fiesta-based runaround that goes THUNK