Despite the promise of cloud, are we treating virtual servers like physical ones?

RightScale had some great data about usage of Amazon EC2 recently that described how cloud computing is evolving, or at least how their portion of that business is progressing. At first glance, it certainly sounds as if things are maturing nicely.

However, a couple things they reported caused me to question whether this trend is as rosy as it seems initially, or if IT is actually falling into a bit of a trap in the way it's starting to use the public cloud. I’ll explain:

Cloud servers are increasing in quantity, getting bigger, and living longer, but…

The RightScale data showed that comparing June 2009 with June 2010, there are now more customers using their service and each of those customers are launching more and more EC2 servers. (I did see a contradictory comment about this from Antonio Piraino of Tier1 Research, but I’ll take the RightScale info at face value for the moment.)

Not only have the number of cloud customers increased, but customers are also using bigger servers (12% used “extra large” server sizes last June, jumping up to 56% this June) and using those servers longer (3.3% of servers were running after 30 days in June 2009, 6.3% did so this June).

CTO Thorsten von Eicken acknowledged in his post that “of course this is not an exact science because some production server arrays grow and shrink on a daily basis and some test servers are left running all the time.” However, he concluded that there is a “clear trend that shows a continued move of business critical computing to the cloud.”

These data points, and the commentary around them, were interesting enough to catch the attention of folks like analyst James Staten from Forrester and CNET blogger James Urquhart on Twitter, and Ben Kepes from GigaOM picked it up as well. IDC analyst Matt Eastwood "knowing a thing or two about the server market" (as he said) was intrigued by the thread about the growing & aging of cloud servers, too, noting that average sales prices (ASPs) are rising.

Matt's comments especially got me thinking about about what parallels the usage of cloud servers might have with the way the on-premise, physical server market progressed. If people are starting to use cloud servers longer, perhaps IT is doing what they do on physical boxes inside their four walls -- moving more constant, permanent workloads to those servers.

Sounds like proof that cloud computing is gaining traction, right? Sure, but it cause me to ask this question:

As cloud computing matures, will "rented" server usage in the cloud start to follow the usage pattern of "owned," on-premise server usage?

And, more specifically:

Despite all the promises of cloud computing, are we actually just treating virtual servers in the cloud like physical ones? Are we simply using cloud computing as another type of static outsourcing?

One potential explanation for the RightScale numbers is that we are simply in the early stages of this market and we in IT operations are doing what we know best in this new environment. In other words, now that some companies have tried using the public cloud (in this particular case, Amazon EC2) for short-term testing and development projects, they’ve moved some more “production”-style workloads to the cloud. They’re transplanting what they know into a new environment that on the surface seems to be cheaper.

These production apps, instead of being the apps that folks such as Joe Weinman from AT&T described in his Cloudonomics posts as being ideal for the cloud because of their highly variable usage patterns, have very steady demand. This, after all, matches the increase in longer-running servers that von Eicken wrote about.

And that seems like a bad thing to me.

Why?

Because moving applications that have relatively steady, consistent workloads to the cloud means that customers are missing one of the most important benefits of cloud computing: elasticity.

Elasticity is the component that makes a cloud architecture fundamentally different from just an outsourced application. It is also the component of the cloud computing concept that can have the most profound economic effect on an IT budget and, in the end, a company’s business. If you only pay for what you use and can handle big swings in demand by having additional compute resources automatically provisioned when required and decommissioned when not, you don’t need those resources sitting around doing nothing the rest of the time. Regardless of whether they are on-premise or in the cloud.

In fact, this ability to automatically add and subtract the computing resources that an application needs has been a bit of a Holy Grail for a while. It’s at the heart of Gartner’s real-time infrastructure concept and other descriptions of how infrastructure is evolving to more closely match your business.

Except that maybe the data say that it isn’t what’s actually happening.

Falling into a V=P trap?

My advice for companies trying out cloud-based services of any sort is to think about what they want out of this. Don’t fall into a V=P trap: that is, don’t think of virtual servers and physical servers the same way.

Separating servers from hardware by making them virtual, and then relocating them anywhere and everywhere into the cloud gives you new possibilities. The time, effort, and knowledge it’s going to take to simply outsource an application may seem worth it in the short term, but many of the public cloud’s benefits are simply not going to materialize if you stop there. Lower cost is probably one of those. Over the long haul a steady-state app may not actually benefit from using a public cloud. The math is the math: be sure you’ve figured out your reasoning and end game before agreeing to pay month after month after month.

Instead, I’d suggest looking for applications in need of different requirements, things you could not get from the part of your infrastructure that's siloed and static today. Even if it is being run by someone else. Definitely take a peek at the math that Joe Weinman did on the industry’s behalf or other sources as you are deciding.

Of course, who am I to argue with what customers are actually doing?

It may turn out that people aren’t actually moving production or constant-workload apps at all. There may be an as-yet-undescribed reason for what RightScale’s data show, or a still-to-be-explained use case that we’re missing.

And if there is, I'm eager to hear it. We should all be flexible and, well, elastic enough to accept that explanation, too.

Cloud, the economy, and short-term thinking: highlights from Gartner's Data Center Conference

There was a lot to take in at the Gartner Data Center Conference this year, and unfortunately I had a very abbreviated amount of time actually on the ground in Vegas to do it. What was the most notable thing? In my opinion, the headline was how much Gartner has strengthened its feelings about cloud computing in 12 months.

And not any 12 months, mind you, but 12 months darkened by the worse economic downturn of our lives. Of course, they did talk plenty about the impact the economy has had, especially on the IT operations psyche and approach, but cloud computing won out, hands down.

From what I gathered, Gartner now sees cloud computing (especially private clouds) and the real-time infrastructures that support them as something big customers should take a look at.

This certainly isn’t an earth-shattering change. Last year at this time, Tom Bittman’s keynote was about how the future of IT looked a lot like a private cloud. This year, the keynotes and sessions throughout the week definitely gave the impression that cloud computing has evolved and improved this past year, and Gartner has moved its thinking as well. The cloud computing discussion felt more mainstream, more central to many of the presentations. It certainly played a starring role in more of them.

Private clouds received quite a bit of positive airtime, and a near-unanimous vote of confidence from Gartner analysts – something that I didn’t feel was true at last year’s event. This year, Bittman was one of a larger chorus, rather than one of the few vocal Gartner private cloud advocates.

Dave Cappuccio’s keynote on “Infrastructure & Operations: Charting a Course for the Coming Decade” emphasized Gartner’s enthusiasm around private clouds with this prediction: of the money being spent on cloud computing over the next 3-5 years, “we think private cloud services will be 70-80% of the investment.” That sounds pretty serious to me.

Cloud computing focus soars, even though budgets don’t

The audience in the room for the opening keynotes (polled as usual with handheld devices they scattered throughout the room) considered cloud computing to be the 3rd biggest data center challenge for 2010 – in a world in which they also expect flat IT operations budgets. The No. 1 challenge was “space, power, & cooling” and No. 2 was that perennial issue, “linking IT & business.” In fact, 6% of people in the room put cloud computing as their top funded item for next year (compared with 9% for virtualization). That’s impressive.

Gartner believes the bad economy is forcing short-term thinking on infrastructure & operations projects

But don't think that just because cloud computing progressed that there's nothing dramatically different from last year. There is: the impact of the economy on the behavior of IT operations.

One of the most depressing quotes of the conference was not from Jeffrey Garten of the Yale School of Management (though I thought it might be). He actually had a pretty interesting talk -- in plain English -- about what has gone well and what has gone badly in the attempts to stave off a full global financial meltdown that went into high gear at the end of 2008.

Instead, it was how Donna Scott and Ronni Colville characterized IT’s response.

Is the best offense is a good defense? Really?

Because of the bad economy, according to Scott and Colville, any talk about data center or infrastructure “transformation” is right out the window. Instead, they said, the kinder, gentler term “restructuring” is in vogue. I guess that descriptor implies short-term ROI, which is at least pragmatic, since projects with long-term pay-outs are off the table completely. Organizations, said Scott, are instead switching to fast (OK, rushed) IT restructuring moves now.

So, hooray for the focus on ROI, but the shorter time horizon isn’t necessarily good news. It means less thoughtful analysis and planning, and a focus only on what can be done immediately. The unfortunate result? “A lot of optimization decisions are made in silos,” Scott said, when in fact sharing those efforts (things like moving to shared IT services), could have a much broader impact. Gee, thanks, bad economy.

In fact, many of the comments I heard from analysts gave the distinct impression that IT is (and should be) on the defensive. “Promises of IT transformation,” said Milind Govekar during his talk on IT metrics, “are taking a back seat to immediate goals of cost optimization, increasing quality or reducing risk.” Govekar noted that, in fact, “cloud hype” is a part of the problem; it “threatens IT budgets and control.”

Comments like these from Govekar and others felt a bit like the analysts were hearing (and in some cases recommending) that IT should circle the wagons, implying that cloud computing was threatening the status quo in IT. Well, if all of its promises come true, it certainly could.

But not if IT operations makes the first move. How? I liked the bit of advice I saw come from Carl Claunch’s keynote: pilot everything. Private clouds, public clouds, hybrid clouds. If you want to know how to do it, the only way to learn is to get started.

For the record, I seem to recall recounting some very similar advice from last year’s conference. Nevertheless, from what I saw and heard this year, Gartner’s research and the evidence from attendees show increased interest and adoption in cloud computing and real-time infrastructures, despite hiccups, bumps, and bruises from the Great Recession.

That in and of itself is impressive.

Why Star Trek can teach us a thing or two about internal cloud computing

In honor of the opening of the new Star Trek (sort-of reboot) movie, the slightly whimsical blog topic for today is a bizarre little connection I noticed between what our Cassatt Data Center Dialog blog normally talks about and how the Starship Enterprise manages to "boldly go." Yes, even if you don't know a tribble from a warp core breach, I'm saying there just might be something to be learned about internal cloud computing from Star Trek.

And if not, at least this post is a good excuse to use the word fizzbin in a sentence in a work setting.

I realize that I am reaching a bit here, but I figured if Star Trek can continue to be a cultural phenomenon after 40 years, it's worth thinking about why. It certainly is right up there with Star Wars and Dilbert as one of IT folks' all-time favorite entertainment (and quote) sources. And since even Wil Wheaton (aka @wilw on Twitter and wunderkind Wesley Crusher on The Next Generation) has given the new movie a big thumbs up, I'm feeling like we're poised for a full-fledged Trekfest for the next few weeks. I might as well join in.

Actually, technology copies Star Trek

Newsweek used their May 4 cover story to ask what continues to make Star Trek so compelling. Sure, green-skinned Orion slave girls have their appeal to a certain audience. But, the Trek ratio of technobabble to green aliens heavily favors the gibberish. One of the Newsweek articles was by former Star Trek: The Next Generation writer Leonard Mlodinow who, after a trial by fire on the show's writing staff, figured out that the point of the show was not to see how you could "put real science in the science fiction," as he was attempting.

Instead, Star Trek lets you imagine something new that exists in the future, as a target for us to strive for. "I had had it backward," wrote Mlodinow. "The fun in Star Trek didn't come from copying science but from having science copy it." Captain James T. Kirk would certainly have recognized all those Motorola RAZRs had one of his slingshots around the sun landed him in 2006, for example.

Which brings me to cloud computing.

The Enterprise's internal cloud: now that's a real-time infrastructure

It seems to me that the minds that came up with -- and got us comfortable with how to use -- the phaser, universal translator, replicator, holodeck, and transporter, also came up with on-premise cloud computing.

Here's why I say that: in the 79 episodes of the original '60s series, and even in the Next Generation version from the '90s, a very seamless computing infrastructure ran the Enterprise. Their futuristic, on-board data center (in addition to inspiring Google, it seems) provided Kirk or Picard's crews with endless, on-demand data; instantaneous holographic simulations despite a few missing variables; and even complex calculations sufficient to bring whales, some very odd wardrobe choices, and at least one hideous toupee 300 years into the future in a stolen Klingon Bird of Prey.

Any real-world IT person would notice that the computing infrastructure and underlying architecture that enabled all this wasn't really mentioned that much. It was asked to do a whole lot (including speech recognition of nervous Scottish engineers), without ever interrupting the flow of any given episode with lame interludes where we had to watch the red-shirted Enterprise IT staff (you had to know that pun was coming) provision servers to each new task that Spock or Data threw at the system.

Of course, most IT folks have been tuning out the representations of "realistic" computing in TV and movies since Matthew Broderick used his home modem to kick off a quick game of "global thermonuclear war' in the '80s. So, I forgive you if you didn't spend much time thinking about this. But I digress.

Sure, there were a few Star Trek computer security problems that occasionally locked the crew on course for the Andromeda galaxy or kept them out of their own bridge, but they rarely had what you or I would call IT operations issues. I recall a few minor run-ins with Iconian probes and sentient nanites, but there really was no drama about the ship's computing infrastructure.

And that's exactly how it should be. This is what we're all working toward. Everyone gets the computing power they need, when they need it. (Which is usually right as the Borg are bearing down on them, blabbering on about resistance being futile and all.) The crew had computer resources allocated to whatever priorities they requested without people manually messing with the IT operations. Reads just like one of our Cassatt white papers. Or an explanation of Gartner's real-time infrastructure concept. Hmmm.

Now, I always had the impression that the Enterprise was actually a giant flying mainframe, which may be the case, but the idea of internal cloud computing has its roots there anyway. You have policy-based automation carving up and allocating available resources based upon priorities ("We're venting plasma down here!") and service levels you set ("I'm a doctor, not a bricklayer!").

Now that's an IT future filled with Roddenberryesque hope and optimism. And one that may sound far-fetched, but as data centers continue their path toward greater and greater complexity, we need computers to do what they are good at and help us mere mortals manage things. We need automation and policy management to take things beyond what we can handle in our own brains (or even Spock's brain) and take things where no one has gone before.

Ahem.

OK, so even if you aren't buying any of this so far, here's another thing to think about. Is there another technology more in need of a futuristic proof point, more in need of a little validation? Fast-forwarding to the 23rd century to see how this cloud computing stuff all works out -- and gets used -- is great fun for two reasons.

One, with the Enterprise, we get to see an organization whose "business" is seamlessly supported by their "IT" -- a situation that most 21st century folks would indeed qualify as science fiction. But it's something to aim for.

And, two, we get to fast-forward through all the cloud computing hype to the point in time where it's just about making stuff work. We can skip both the market noise and the chaotic mess that's sometimes supporting today's applications -- and get on with the living long and prospering.

Anyway, enjoy the movie. I expect a flurry of 120-minute "team lunches" at cineplexes just down the road from IT departments around the planet. I'll be the guy in the costume.

P.S. Just in case you think I’m a little too immersed in this ‘Star Trek’ thing, be thankful that at least I'm not doing Klingon opera like these guys (thanks, @digiphile).

(Update: There was a fun front-page San Francisco Chronicle article over the weekend by Benny Evangelista that recounted a number of the technologies that 'Star Trek' has inspired over the years, some successfully, some not so much. Two of note: a sort of medical tricorder that doesn't require needles to learn things about your blood and a universal translator-type device in use in Iraq since 2003. I'm still waiting for the transporter, myself.)

The 'Bio Data Center': working with Bull on data center efficiency in Europe

Back in November, I mentioned that we had been working with some big customers to improve the way they run many of the largest data centers in North America. With Tuesday's partnership announcement with Bull, Cassatt is expanding that focus to Europe as well.

If you didn't see the announcement, here's a very short recap: Bull has been working on a strategy to help its European customer base improve the efficiency of their data centers. They call it the "Bio Data Center." Sounds a bit esoteric, but the concept is very similar to what Gartner calls a real-time infrastructure.

In short, they are working on ways to improve IT operations processes, data center architecture, and power consumption. Which, of course, sounded very familiar to a lot of what we're trying to do.

So, after a bit of investigation (including some pretty serious technical evaluations), Bull is putting Cassatt squarely in the center of their efforts. The joint work will give European organizations a way to get access to the Cassatt Active Response software without having to wait for us to build out a European field organization. (Having lived in London for 3 years while with BEA, I'm hoping that a Cassatt presence in Europe comes sooner rather than later, but I'll be patient for the moment.)

The partnership means Bull will distribute our software in Europe and other select geographies. Our software, as a reminder, uses policy-based management to control and optimize the diverse data center resources that support your applications (hardware, software, virtual machines, networks) without requiring you to change any of the pieces.

The idea is to take incremental steps that drastically cut your operations costs (our models show most folks can cut 30% pretty easily for starters) with the goal of simultaneously making your infrastructure able to match your business much more effectively.

To help get started, customers in Europe will be able to turn to Bull for strategic guidance and implementation help. We're already working with them on detailed training so their people on the ground in France, Germany, the U.K., and other parts of Europe can be as up-to-speed on our stuff as they are on the issues that data center managers in Europe are facing.

That's something Bull is bringing to the party: direct experience in those European data centers, which I always found to be at different stages of IT evolution than their American counterparts.

Here are a couple examples of what Bull and Cassatt see "over there":

- A much more activist role of government in creating "green tape" (as the 451 Group calls it) -- government regulations around energy efficiency in data centers
- A more pervasive adoption of methodical, structured processes like ITIL. Many data center managers over there make us over here look like cowboys shooting from the hip.
- A slower adoption and healthy skepticism of cloud computing
- The recession has not hit markets like France with the immediacy and profound impact on IT budgets that it has here in the States. Yet.

We look forward to collaborating with Bull in the coming months and years. I’ll post updates as our work across The Pond progresses. If you’re a European organization interested in finding out more, drop us a line (info@cassatt.com).

'Real-time infrastructure': where do you start?

I saw a lot of coverage last week on Tom Bittman's keynote at the Gartner Data Center Conference in Vegas about how IT infrastructure is going to evolve (summary: it's looking very cloud-like). Only a few of the articles (like Derrick Harris' blog) picked up some of the angles Donna Scott presented the next day to another packed room -- more of an "OK, I'm interested, now how do I get started?" session about creating a real-time infrastructure. Given all the hype about how companies should be creating next-generation architectures and the like, I thought her focus on what's being done and what could be done today was worthy of some highlights.

First off, Donna Scott gave credit where credit was due: a big driver toward real-time infrastructure (RTI) has been (and will continue to be) virtualization. It's the first step out of the moving airplane, hopefully with the parachute on. Virtualization allows IT operations to see that it's possible to de-couple hardware and software, to do things differently, and to get seriously beneficial results.

Cost wasn't the main driver for these initial steps toward a real-time infrastructure, according to the in-room poll that Donna took. Thirty-five percent said the reason was agility instead. But it's not yet full-steam ahead to get that agility benefit, either. "There's no such thing as data-center-wide RTI," Scott said. "It only exists in pockets at this point."

So what's holding RTI up?

The two highest results from her polling told a tale that we here at Cassatt have heard before: it's not the technology that's the problem. Management process maturity (29%) and organizational/cultural barriers (28%) were holding organizations back from adopting RTI instead. In fact, in Scott's poll "unproven technology" was the reason holding back only 10% of the audience. To put this in context from our own experience, Cassatt's largest customers have needed the most assistance with handling internal politics, organizational objections, and modifications to the way they are managing their IT operations as they move through an implementation. Some of this assistance comes from us, but a lot of the broader change-management expertise can come from the partners we work with. Donna summed it up nicely: "Don't ignore the people and cultural issues."

Donna's polling also found that organizations in the room had started server virtualization projects as their initial foray into creating a real-time infrastructure (56% chose this option). Makes sense. Three other areas that got some interest as RTI starting points:

1. Disaster recovery -- sharing and re-configuration (11%)
2. Loosely coupled server high-availability -- replacement of failed nodes (7%)
3. Dynamic capacity expansion-- typically for specific applications (7%)

Scott pointed to a way to get the "best of both worlds" and begin making steps toward RTI: combine server virtualization and a dynamic resource broker (what she called a "service governor"). In fact, her assumption is that "there will be many service governors in your organization," often including specialized ones that know specific types of environments very well.

So where do you start?

Donna went on to provide one of the first categorizations of service governors that's been provided to date. Her slide listed application-centric service governors, infrastructure-centric service governors, and other flavors as well. Who's providing them? Vendors like IBM, CA, DataSynapse, VMware, and, yes, Cassatt, too. She put us in a couple of categories, in fact.

"All vendors have something to offer," Scott said, but each has its own perspective on how to deliver value for customers. (Cassatt's take, by the way, is to be as vendor-neutral as possible, while providing dynamic control over the broadest set of hardware, software, and virtualization as we can.) "Most of the innovation in this space," Scott said, "is happening with the smaller suppliers."

She wrapped up her comments with a bit of a warning and a challenge: some organizations are eager to give real-time infrastructure a shot (and these are obviously the type that are Cassatt customers), figuring that the resulting rewards will easily outstrip the risks.

"RTI is not for the faint of heart," said Scott. "This is really cutting-edge stuff, so you could be breaking new ground."

I suppose it goes without saying, but if you're interested in someone to help break that new ground with you, I think we could recommend that someone.

From Gartner's Data Center Conf: Tom Bittman says the future looks a lot like a private cloud

Here's an amusing tidbit that gives you some indication what Gartner's Tom Bittman thinks the future holds for IT infrastructure and operations: he had so many "cloud" images in his keynote PowerPoint presentation here in Vegas on Tuesday that he actually had to go out and take photos of clouds himself. He needed to have enough distinctly different images that simply using clip art was getting Gartner into copyright issues (lawyers do care where you grab clip art from, I guess).

Bittman's presentation and comments kicked off the 27th annual Gartner Data Center Conference. Since Cassatt is here as a silver sponsor anyway, I thought I'd take the opportunity to give you a quick run-down of some of the highlights as they happen. Or at least as soon after they happen as is reasonable in a town like Las Vegas.

Back to Tom's opening comments: in laying out the future of infrastructure and operations, he described them as the "engine of cloud computing." In other words, pretty central to what is happening in this space. Cloud computing as provided by independent vendors today or as delivered by you inside your own data center all require an evolution in IT operations & management. Here are some interesting comments he made on-stage:

"The future of infrastructure and operations looks an awful lot like a private cloud." With the advent of cloud-style computing, "it's now about who can share things best," according to Bittman. But sharing is not something IT -- or end users, for that matter -- have been used to doing. Or have liked doing. Certainly virtualization has gone a long way toward getting people to think about their IT resources in a new way. In fact, he pointed out how interrelated virtualization, cloud computing, and Gartner's concept of a "real-time infrastructure" actually are. That real-time infrastructure story is one that Bittman, Donna Scott, and the Gartner folks blazed the trail with seven years ago and Tom noted that examples are starting to appear. Google and Amazon have some of this figured out. Tom gave them credit for understanding how to do shareable infrastructure, but not for understanding quality of service requirements for applications anywhere close to the way the IT people in his audience do.

"If you fully utilize your own equipment, Amazon [EC2] will cost you twice as much." Meaning this: the cloud computing benefit that is most interesting is not price. In most enterprises, if you are doing a good job of running your IT ops, you can do it for less than Amazon can. Instead, Bittman contended, it is the low barrier to entry that's so intriguing. And the concept of elasticity. The cloud providers have a lot of work to do managing elasticity, though, and it's one of the reasons Cassatt talks to organizations about using our software to do that -- for their internal resources.

"[In cloud computing] you don't ask how it's done, you ask for a result." This is one of the beauties of cloud computing, inside or outside your four walls. It's about masking the underlying infrastructure from you in a way that's financially beneficial and very flexible. In discussing this, Tom explained the new "meta operating system" required in the data center (or in your cloud service provider's data center) to make this all possible. He talked about VMware's future vision of something like this announced at VMworld (Ken Oestreich provided a good commentary on the pros/cons of that) and Microsoft's newly announced Azure.

But, Bittman said, "a meta operating system is not enough. We need to have something that can automate this to service levels. In cloud computing, today this is very simplistic. There's very little you can control." In fact, this is a sign of the immaturity of the cloud offerings.

The thing that can fill this gap is what Tom and Gartner call a "service governor" (and, is the category in which Donna Scott placed Cassatt Active Response in her Wednesday RTI presentation). A "service governor is what's going to take advantage of the meta OS -- whatever's under the covers -- and make it do useful work," Bittman said.

Bittman also offered a warning. The path that cloud computing doesn't want to follow is that of client/server. People went around central IT and the result was skyrocketing costs and little integration. Failures will be rampant with cloud computing, he said, unless IT is involved, no matter how well-meaning those do-it-yourselfers from the business side of the house are.

His advice: have a plan. Be proactive when it comes to cloud computing, but have a blueprint about where you're going, especially in these economic times. Or, if I may add my own suggestion, at least hedge your bets by starting a side business selling cloud photography to IT departments, analysts, and vendors. We could even use a few ourselves.

Next up: more highlights from Vegas, baby.