EMC blogger Storagezilla posted an interesting Flash animated video this morning about Maui, titled CloudFellas, in a post that has since been whacked. In the original post, ‘zilla alluded to ‘getting too far out in front of the boss’, so maybe that’s what happened (the post has been deleted from Google’s cache as well).

The video showed fun little animations about the spread of data to points around the world, and gave the example of a movie project where dailies from the set have to be sent to production houses for editing, then from the production houses back to the studio for vetting, and eventually out to movie theaters for distribution. Connecting multinational islands of data seemed to be a theme, as was scalability to petabytes, even exabytes.

This is important because EMC has yet to formally tell us just what Maui actually does. When Hulk/Maui were first discussed during EMC’s Innovation day last fall, it was assumed that Maui was the file system for Hulk’s hardware. But it turns out Hulk is shipping with Ibrix as its front-end file system, and according rumors that were going around about Maui at EMC World in May, Maui is instead a layer of software that sits above local storage pools, which could serve as a global data repository for multinational companies, tying multiple data centers together.

This even jibes with the codename - Maui is an island in Hawaii as well as the name of the Hawaiian god that raised the Hawaiian islands from the sea. Raising islands (of storage), joining them together in a chain… 

Then there’s ‘zilla’s comment this morning in his original post: “the internal cloud currently stretches from the east coast of America right into China.” He also mentioned that the business plan is executing to schedule, which would mean Hulk and Maui will both be formally introduced in the third quarter.

NetApp has a strong working relationship with VMware, despite counting VMware’s parent EMC as its storage archrival. So NetApp CEO Dan Warmenhoven didn’t know what to make of the news when he heard VMware swapped CEOs from founder Diane Greene for former Microsoft exec Paul Maritz last month.

“When I first read the news about the change I was a bit shocked,” Warmenhoven said during NetApp’s earnings conference call Wednesday.

The next surprise for Warmenhoven came when he received a phone call from Maritz hours later. “I really want to thank Paul,” Warmenhoven said. “He placed a call to me before 1 p.m. the day he was announced as the CEO, and I’ve got to imagine that was a very busy day for him. When we did connect, he actually reaffirmed every part of the relationship we had prior, and even took some very visible actions to strengthen the relationship so we were very, very pleased. I think it’s going to be terrific.”

Warmenhoven called Maritz’s reaching out “very pragmatic,” considering there is a large pipeline of customers looking to implement VMware and NetApp storage. And with Microsoft entering the server virtualization market, VMware needs all the friends it can get. “He’s facing some significant competition coming up on the horizon, and he’s not about to jeopardize any close relationships he has,” Warmenhoven said.

Maritz has been doing a lot of reaching out in his early days as VMware CEO. Besides talking to VMware storage and server partners, he had to apologize to customers this week for a VMware bug that locked up their servers.

“You need HOW MUCH for storage?!” That question has been heard by many of us currently submitting budgets for the next calendar year, quickly followed by “Are you SURE you need that much disk? Didn’t we just get disk last year? Where did they all go!? I want your house audited. Now!”

Okay, maybe not the audit part, but for most of us, getting the type of disk we need in the quantity we need it is an uphill battle. Add SSD, deduplication, and longer-term retention to the mix, and things are getting a bit hairy with my budgetary requests. I’m at such a point now with a few of my smaller clients, and when they get that “you’re crazy” look, I bring up the chargeback model.

I think I just heard a collective sigh from the interwebs.

I understand both sides of the chargeback dilemma: the accounting side, that has to somehow keep track of all this without keeping track of all this; and the IT side, that is constantly being painted as the cost center only because no one is taking ownership of their parts of the “plumbing.” People (read departments) will request outrageous resources when they don’t have to directly foot the bill. That part I get, but are they so vehemently against accounting for their infrastructure usage?

In my opinion, chargeback would actually lead to better data management habits — at least in the long term — because if you have to pay for everything out of your own budget, then you’ll be more careful about separating what you need from what you want. How many of our managers and accounting folks have processes in place to account for each department’s use of the “utilities” that make up IT and understand that IT isn’t the root of all expenses?

I had an energetic debate with a co-worker about this very issue. I took the stance that chargeback is the way to go. He offered a more community-oriented accounting method. We went back and forth, point and counterpoint, until concluding that it just depends on what your business environment will support and the level of organization that business has in place.

For instance, if you have a well-organized, project-oriented IT environment, and have a project portfolio ready for sizing, you can plan a community budget very well and effectively fund addition to your infrastructure through a single IT budget. The reality from <i>my</I> experience (read, SMB clients) is that most companies are not so well-organized, don’t have a project portfolio for the next 12 months, and will not be able to identify budgetary requirements for infrastructure improvements.

In these cases, chargeback (or, at the very least, departmental accounting) is key to being able to answer my opening question with confidence.

Traditional SAN storage may be easy to bill for, but what of virtualized storage? Take it a step further, how about Softricity/Microsoft’s Softgrid? (Softricity is the company Microsoft acquired not too long ago that allows for application-level virtualization as opposed to host virtualization.) How do you quantify and itemize a streamed, virtualized application?

Then there’s the question floating just below the surface of the chargeback debate: How do I, as a department, know you are giving me what I’m being “billed” for? That question opens a giant can of worms in my mind (and there are already creepy crawlies up there, no need to add worms to the mix).

The crux of what I’m getting at is: Are we as technologists — and storage pros specifically — asking for too much or too little when it comes to chargeback? Are there still companies out there that don’t see the light when it comes to chargeback and departmental accounting. Should we as storage pros be leading the way for other areas of IT to follow our example?

Last time on As the Benchmark Turns, we saw NetApp pull a fast one on EMC. This week’s episode brings us into the torrid realm of server virtualization and Hyper-V.

It began with a press release from QLogic and Microsoft lauding benchmark results showing “near-native throughput” on Hyper-V hosts attached to storage via Fibre Channel, to the tune of 200,000 IOPS.

Server virtualization analyst Chris Wolf of the Burton Group took the testing methodology to task on his blog:

The press release was careful to state the hypervisor and fibre channel HBA (QLogic 2500 Series 8Gb adapter), but failed to mention the back end storage configuration. I consider this to be an important omission. After some digging around, I was able to find the benchmark results here. If I was watching an Olympic event, this would be the moment where after thinking I witnessed an incredible athletic event, I learned that the athlete tested positive for steroids. Microsoft and QLogic didn’t take a fibre channel disk array and inject it with Stanzanol or rub it with “the clear,” but they did use solid state storage. The storage array used was a Texas Memory RamSan 325 FC storage array. The benchmark that resulted in nearly 200,000 IOPS, as you’ll see from the diagram, ran within 90% of native performance (180,000 IOPS). However, this benchmark used a completely unrealistic block size of 512 bytes (a block size of 8K or 16K would have been more realistic). The benchmark that resulted in close to native throughput (3% performance delta) yielded performance of 120,426 IOPS with an 8KB block size. No other virtualization vendors have published benchmarks using solid state storage, so the QLogic/Hyper-V benchmark, to me, really hasn’t proven anything.

I talked with QLogic’s VP of corporate marketing, Frank Berry, about this yesterday. He said that Wolf had misinterpreted the intention of the testing, which he said was only meant to show the performance of Hyper-V vs. a native Windows server deployment. “Storage performance wasn’t at issue,” he said. At least one of Wolf’s commenters pointed this out, too:

You want to demonstrate the speed of your devices, the (sic) you avoid any other bottlenecks: so you use RamSan. You want to show transitions to and from your VM do not matter, then you use a blocksize that uses a lot of transitions: 512 bytes…

But Wolf points to wording in the QLogic press release, claiming the result ”surpasses the existing benchmark results in the market,” and claiming that ”implies that the Hyper-V/QLogic benchmark has outperformed a comparable VMware benchmark.” Furthermore, he adds:

Too many vendors provide benchmark results that involve running a single VM on a single physical host (I’m assuming that’s the case with the Microsoft/QLogic benchmark). I don’t think you’ll find a VMware benchmark published in the last couple of months that does not include scalability results. If you want to prove the performance of the hypervisor, you have to do so under a real workload. Benchmarking the performance of 1 VM on 1 host does not accurately reflect the scheduling work that the hypervisor needs to do, so to me this is not a true reflection of VM/hypervisor performance. Show me scalability up to 8 VMs and I’m a believer, since consolidation ratios of 8:1 to 12:1 have been pretty typical. When I see benchmarks that are completely absent of any type of real world workload, I’m going to bring attention to them.

And really, why didn’t QLogic mention the full configuration in first reporting the test results? For that matter, why not use regular disk during the testing, since that’s what most customers are going to be using?

On the other hand, QLogic and Microsoft would be far from the first to put their best testing results forward. But does anyone really base decisions around vendor-provided performance benchmarks anyway?