« Information Governance Enters The Mainstream | Main | Better Times Ahead For Service Providers? »

October 24, 2008


marc farley

Chuck, I think this is the first step towards storage socialism from IBM!

But seriously, don't you think there might have been NDAs in place?

Free evaluations do end up being purchased. It's an effective strategy that worked very well for EqualLogic and others (including 3PAR). Whether or not IBM is doing this and whether or not they succeed will be interesting to all of us. Thanks for pointing it out.

Chuck Hollis

Hi Marc

Sure -- every vendor does a bit of eval in the process of selling their wares, including EMC.

But usually, it's a precondition for a sale. What I'm seeing is the IBM Santa Claus dropping them off everywhere, with no apparent preconditions or defined expectations.

And I think that's unusual.

Sudhir Brahma

I think there could be some good reasoning behind this exercise of providing free Evals, beyond making a "trap sale" as has been mentioned.
If u notice, the XIV is quite inefficient in offering usable disk space: Typically, with 120TB drive, the user only gets 80TB (about 67%) - the rest of it (40TB)goes into "distributed parity data", which is used by the XIV in rebuilding a failed drive in a RAID...Yes they don’t use standard RAID controllers (which are much more space efficient, but a little slow in rebuilding). They have their own RAID algorithm, to do that. Their "claim to fame" being: "lower rebuilding time for failed drives", which unfortunately is at the cost of usable drive space.....it is probably similar to having a luxury parachute (complete with a bar and refrigerator) occupying more of usable space in the plane...yes the bail-out experience is probably breathtaking!! The industry standard RAID controllers only have normal "parachutes", which just do the job of rebuilding failed drives-plain and simple, no wine and no chilled soda during the drop!
With these Eval machines, IBM is now probably testing waters with these "luxury parachutes" and collecting data on what really matters to the end user and what really happens in a production environment- is it usable space or the rebuilding time? I bet their storage boxes are logging data on the number of drive failures and the rebuild time over time. They will then use that statistics to tell the users how they recovered from drive failures faster than competition and how much "quality time they saved", because of this new RAID algorithm.
What the user may not know is the associated cost:
1. How much storage he lost in the process.

2. During rebuilding, if his entire node was affected in IO response times for good drives . I understand that the entire node is associated in the rebuilding process and IO to good drives could be affected because of this rebuilding exercise.

Chuck Hollis


I don't think your numbers are correct, it's worse, please see Barry's post at thestorageanarchist.typepad.com

Many people have expressed concerns about XIV's parity scheme and rebuild process. One of the first things I see customers do is start pulling drives and see what happens.


Sudhir Brahma

Thank you Chuck..yes you are right. It is 80+ TB out of 180 TB. 120 TB was probably old data and that had about 51 TB usable space. On the matter of usable space, I referred the matter to a friend of mine, associated with Data Centres and here is what he said: "As long as they are SATA, we dont care too much about useable space and yes being close to RAID 1 was OK.". Looks like for critical data, they have this maintenance policy of regularly replacing drives after a given usage duration. Having SATA drives helps in saving.

It seems that for such requirements, the ability to rebuild faster, is key, so the array come back sooner...they recover all the costs doubly from their clients, so they don’t care too much (I wonder with this downturn, how long this policy will last though). The XIV does thin provisioning so that could help him in planning for replacements better. On the matter of double drive faults, he said, the probability of power outages (including failure of back up power ) was far more and recovery is always possible from the remote mirrored copies!!...so I guess there is a requirement for such "luxury parachutes" after all!!

I bet, as we discuss, IBM is slowly and steadily collecting such actual usage data on “what really matters” in data centres, through these eval boxes and building a set of case studies. This information could be vital for their sales teams!


Geoff Mitchell

I think that customers that are careless enough to put production data on an eval units undergoing a proof of concept deserve every invoice they get from IBM!

Chuck, I'm not sure what you're saying exactly.
Are you saying that IBM is giving away XIV's for free, or are they supply them under a try and buy evaluation program?

The former's not going to help sales figures and the latter is not free.

I don't believe anything's really for free, unless being paid for out of someone's marketing bucket, or the customer's paying for it in another way (such as buy a SAN and we'll throw in a free NAS head).
Are they throwing them in with a mainframe or P-series sale, or buy two sharks, get an XIV free scheme?

To answer your questions

1 -- Can any storage product be successful without some sort of clear positioning as to where it fits, and what it does better than other storage products?

Yes. Customers are a canny lot and despite logic, make some very unusual decisions from time to time. They will choose to believe a trusted source over others, despite the preponderance of evidence to the contrary.

2 -- Will the strategy of seeding customer data centers with "free evaluations" at massive scale actually translate into revenue and happy customers

Certainly an installed based will give IBM bragging rights (aka 1000+ installed units). Revenue debatable - those customers that like it will buy more - and happy customers, to be decided.

I've read up a lot on XIV and I still don't get the relevance of this product, either to IBM's portfolio or the storage market in general.

Geoff @ Dell

Chuck Hollis

Hi Geoff

My survey isn't exactly scientific, but it seems to be a lot of "try and buy" without the "buy" part, e.g. no clear T&C on length, price, conditions, etc.

We've seen exactly one legit quote for XIV from IBM, and the pricing was mind-boggling -- more than an EMC DMX, more than an EMC CX, more than just about every storage array out there in terms of price per usable capacity.

It's weird, I tell you ... weird.

Chuck Hollis

This is Chuck

Here's what I can't figure out.

If someone wanted SATA capacity (e.g. 1 TB) with either RAID 6 or RAID 1 protection, you could do a CLARiiON for a whole lot less money,

The CLARiiON would offer more capacity, more performance and better availability for a lower price -- in addition to getting a boatload of features like replication, QoS, spindown, virtual provisioning, security, etc. that XIV doesn't provide.

Same deal would probably exist for other midtier arrays out there.

So -- what's the appeal? Doing something nice for your IBM rep for Christmas?

I just don't get it.

Martin G

Have I been offered an XIV on evaluation? You bet I have; there's been some interesting suggestions as to what I might want to evaluate on it and it's definately production-like data that has been suggested.

Am I going to evaluate one? Not a chance at the moment I'm afraid, so I won't be able to give you the skinny on it.

I'm not a big fan of 'try and buy' schemes anyway; they are extremely dangerous and it is amazing the number of times production data does end up on 'try and buy' kit leaving you with a rapid migration issue if you decide to not buy. Still, I've had a few 'try and buys' where the kit has never been paid for and never removed.

And XIV pricing; I've not seen any yet but I have been told that it will not be knowingly undersold (UK readers will know what I mean).

Sudhir Brahma

Hello Chuck,
I can only tell you what my friend tells me: There are 2 key aspects:
1. Fast rebuild
2. Low cost SATA drives

For some requirements, they replace drives after a stipulated time or usage, irrespective of their current state. These new drives are first tested, before they are put to production use and labeled to be removed after the said duration/usage. It is a process that seems to work well and data uptime is high. When they insert these new drives, they want the rebuilding time to be less, so the array goes on-line soon. This unit seems to address such a requirement. That coupled with the low cost of SATA looks like an appealing proposition for such applications.


Hmmmm, at one company I was with every eval unit we ever got was used with production data. I have had chips taped out on machines way past the eval period, with the vendor wondering where their hardware was(we shipped it back eventually). I suspect the XIV's will get quite a work out, and I suspect IBM will eventually bill for them and then look the other way when the check stays in the mail (bet it makes the books look real good).

Jason Treu

Since you are on this topic, I was wondering if we would get an update on Hulk/Maui and when that would be launched?

You can check out HP's comments:

Happy Early Halloween,

Chuck Hollis

Hi Jason

Unfortunately, I can't really comment on unannounced products, especially their status prior to GA.

Although I'm itching to talk about this one, I am ...

Stay tuned!

-- Chuck


Don't you think you owe the external world a reason Joe Tucci said you'd launch in May (last November)? And in May said you would launch by end of Summer and it still isn't launched yet? One slip anyone can make, but 2 slips and we are counting now several months after that last time your CEO said it was coming.

Stay tuned b/c we have more news coming...soon...and customers too.
Have a good one,

Chuck Hollis

Jason, I bet you were an annoying kid in school.

Just to specifically answer your question, I don't think I owe you anything whatsoever.

If I were you, I'd spend my efforts making HP's storage portfolio competitive, rather than throwing rocks at a market leader via blog comments.

Because you and your company have a lot of work to do, my friend.

-- Chuck

Martin G

I wonder if Chuck will let this through but can I say that if you are expecting an extremely scalable NAS solution from Hulk/Maui, you might be a little disappointed. It's not a PolyServ look-alike. Hopefully that is a vague enough.

Chuck Hollis

Hi Martin -- depends on what you mean by "scalable", doesn't it?

So, any speculation on how many peole won't "get" what these products are all about?

-- Chuck

Trichy Premkumar

mmm... I can see IBM's technique will work very well for building up reference sites - after all, who would buy without a reference site?

The comments to this entry are closed.

Chuck Hollis

  • Chuck Hollis
    SVP, Oracle Converged Infrastructure Systems

    Chuck now works for Oracle, and is now deeply embroiled in IT infrastructure.

    Previously, he was with VMware for 2 years, and EMC for 18 years before that, most of them great.

    He enjoys speaking to customer and industry audiences about a variety of technology topics, and -- of course -- enjoys blogging.

    Chuck lives in Vero Beach, FL with his wife and four dogs when he's not traveling. In his spare time, Chuck is working on his second career as an aging rock musician.

    Warning: do not ever buy him a drink when there is a piano nearby.

    Note: these are my personal views, and aren't reviewed or approved by my employer.
Enter your Email:
Preview | Powered by FeedBlitz

General Housekeeping

  • Frequency of Updates
    I try and write something new 1-2 times per week; less if I'm travelling, more if I'm in the office. Hopefully you'll find the frequency about right!
  • Comments and Feedback
    All courteous comments welcome. TypePad occasionally puts comments into the spam folder, but I'll fish them out. Thanks!