- casino online slots


by Administrator on October 4, 2008

SNIA will shortly be submitting both its SMI-S management specification and eXtensible Access Method (XAM) to the various standards groups for adoption as architectural standards. 


Architecture.  What does that mean exactly?  I asked smart guys in the industry to give me a definition, but most say that, whatever architecture is, it is a stretch to regard protocols or methods as “architectures.”  Bottom line:  by calling these things architecture, SNIA is once again converting architecture into marketecture.

In the interest of fairness, and for those of you who don’t know what XAM is, SNIA has a nice flash presentation on it at their website.  Turn down the volume:  the presentation voice is a bit noisy and sounds about as coherent at points as Sarah Palin talking economic policy in a Katie Couric interview.

The show starts with an explanation of a problem:  “Companies store a lot of non-changing data that the SNIA chooses to call reference data.”  OK.

XAM is called a specification of “a standardized way” to store data and metadata to “XAM-conformant” storage systems.  Hmm.

Prepare for a jump:  the objective of XAM is to help companies realize “information independence.”  What the heck does that mean?  The presentation goes on to show how you can be “independent” by using Vignette with EMC Centera, HP IAP, or Sun ST5800 platforms.  I guess this list will grow if more vendors of data movers and storage platforms agree to embrace XAM specs — which has about as much chance of happening as all vendors embracing SMI-S.

Illustrated is a drag and drop file movement into the XAM-compliant repository and a quick retrieval of a document stored in the repository.  These are things I can do with a file system today.  (NOTE TO SELF:  Maybe do a spoof on the XAM flash demo showing how all of these things can be accomplished using any storage and any data mover, even Windows Storage Server.)

Next we see a demo of “HP’s database archiving” product (Outer Bay?) to clip out old data and place it into the XAM conformant repository.  Database performance improves with XAM, the flash claims.  The truth is, any database archiving approach removes stale data from the database, allowing it to cook better — point being, you don’t need XAM to do this.

Next comes a contradiction:  with XAM, you don’t need any third party software.  Then, the animation goes on to show that EMC provides XAM enablement on its Disk Extender product to move data into XAM conformant Centera (isn’t this third party software?).

You can select files for inclusion in certain data movement rules.  Okay.  You can use an SDK to enable your own apps with XAM code.  Okay.  “You can code to the interface without the XAM conformant storage device.”  Hmm.  Doesn’t that kind of defeat the point?  “This reduces the risk and cost of technology obsolescence and vendor lock-in.”  Except that once you write to Centera…

“With XAM’s integrated metadata query, users can easily search for data but without scaling limitations of file systems or file system overhead.”  Hmm.  Of course, this only works in a top to bottom XAM implementation.  The gear on the back end needs to be XAM conformant.

Conclusion:  “As the volume of reference data expands exponentially…” (What is the exponent?  Doesn’t everything change by exponents according to Occam’s Razor?)

“…Only XAM provides an interface standard…” Not true.

“…That eliminates vendor lock-in…” Not true.  Earlier they said this scheme only works with XAM conformant hardware.

“…Achieve application interoperability…”  Zzzzzz.

“…Storage transparency…”  Zzzzzz.

“…And automation for ILM based activities.”  Hmm.

XAM enables app dev to build “robust” applications with “XAM common libraries” (metadata constructs and APIs, assuming that either are embedded in hardware and software.)


Look, my bullshit meter is climbing to orange here.  Everything suggested here, the core value prop, is already available from ECM and archive management software that are not “XAM aware” and it works on any storage platform that connects to any server WITH OR WITHOUT a XAM provider.

The useful stuff might be the metadata constructs underlying XAM, which I will take a look at shortly.  For now, this SNIA development effort has the smell of yet another “elite club” of vendors who want to get their products “XAM certified” as another pointless SNIAesque product discriminator.

We will shortly relaunch with a goal of using real web services standards from W3C, which are already supported by application vendors and by a growing number of management software and storage hardware vendors (starting with ISE from Xiotech), to support policy-based data resource provisioning and protection and to facilitate data movements.  The code is free, the development effort will be minimal, and the application software already talks the talk.

Some fun, eh Bambi?

{ 4 comments… read them below or add one }

Jon October 6, 2008 at 9:15 am


We are being told that this interface is to keep data for generations. I am entirely cynical that the API will last 6 months without alteration, yet alone a generation. Therefore, I’ll need to keep updating my backend storage solution to match the latest API version used by my latest application-that-uses-xam. The API is WAY too complex – it states at the start that it is seeking “simplicity” and a “minimum number of API commands” and then goes on to 178 pages in the architecture document and 154 more pages in the Java document. And it could have gone on much further – the documentation isn’t exactly verbose at describing many of the concepts.

This is going to go through multiple alterations.

Administrator October 6, 2008 at 1:44 pm

Someone smart just observed that the number of pages in storage standards docs is 100s of times greater than the number of pages in the Bible…and look how many wars have stemmed from interpretation of the latter. October 14, 2008 at 2:52 pm

Your comments are unfair:
1) Comparing XAM (a relatively light protocol -it’s its beauty-) and SMI-S (an extremely extensive and heavy specification) is just non-sense.
2) Your digression about the word architecture is just that, a digression. (Thank God, you are not writing protocol specifications…)
3) Rather than concentrating on the XAM SNIA presentation, why don’t you address the specifications themselves: it’s the first version and there is food for constructive criticism there.

Administrator October 24, 2008 at 7:49 pm

Hey rd,

I am delighted to begin a thread here for a more useful discussion. Let’s begin with something like, What problem are you trying to solve?

We’re all ears.

The Management.

Previous post:

Next post: