June 19, 2008 | Comments: (0)
Red Hat the latest proof that cloud computing is serious business
You know a technology is getting some traction when imitators jump into the game. That's the case these days with cloud computing. Just this week, Amazon added Red Hat's JBoss to its EC2 cloud computing platform, and an established hosting service -- ServePath -- jumped into the fray with a version of cloud computing called GoGrid.
There's been a spirited argument in IT circles about whether cloud computing is ready for the enterprise. That's still unclear.
But there's little doubt that there's room in smaller companies and departments of enterprises to save time and money by using someone else's infrastructure for some deployments. The New York Times, to cite a fairly well known example, used Amazon Web Services (EC2 and S3) to generate PDFs of 11 million articles in the paper’s archives, a mammoth, but one-time only job.
More recently, ScribbleLive, a blogging platform, faced a different, but conceptually similar, problem. The site typically gets 181,000 page views a day. But during Apple's developer conference, usage spiked to 2.3 million page views. Scribble, a two-person operation, quickly scaled up using GoGrid, and was able to keep running with little or no loss of throughput. The price: $15 for a day of server time, plus bandwidth charges. Hmm. Maybe the folks at Mozilla, who tried to set a world's record for downloads of the new Firefox browser but wound up crashing their site, could have done something similar.
Point-and-click infrastructure
In theory, at least, setting up servers on GoGrid seems almost too easy. Once a client signs up for the service, an IT staffer can point a browser to GoGrid's site and choose a configuration from a variety of pull down menus. GoGrid supports Windows Server 2003, CentOS and Red Hat Enterprise Linux. Windows 2008 support will be available in July. Server images -- of which there are at least a couple of dozen -- include preinstalled software such as Apache, MySQL, PHP and Ruby on Rails.
Another interesting feature on the GoGrid road map is the ability to scale up automatically when demand spikes. For now, IT admins will have to do it manually, an easy enough operation, but the trick is to anticipate that demand and plan accordingly.
ServePath has dedicated approximately 100 servers from its store of several thousand to the project, says co-founder Dave Hecht. As demand grows, the San Francisco-based company will add more servers to the project.
Pricing is flexible; ranging from "pay as you go" at 19 cents per hour for each of up to eight virtual servers, through $2,500 a month for up to 100 servers with an allotment of 30,000 server RAM hours. (That is, a server with say 1GB of memory, will consume one RAM hour while running.) GoGrid claims to be cheaper than Amazon, but I'm not going to touch that one because making a fair comparison is difficult.
JBoss Heads for the cloud
Meanwhile, Red Hat is taking an interesting path toward the cloud. Rather than follow in the footsteps of Microsoft and build out its own infrastructure to host its software -- an expensive and time-consuming proposition -- Red Hat is leveraging Amazon's massive compute infrastructure to host its offerings.
Customers can either license JBoss on EC2 from Amazon and receive a virtual image of the software, or make their own subscription of JBoss available on Amazon's compute cloud. Pretty cool. Easy for developers and good for margins at Red Hat, since incremental costs for adding new customers who use the service will be low.
I'm not suggesting that GoGrid will become a real threat to Amazon. The larger company has serious advantages as a first mover, and even more advantages of scale. Actually, it doesn't matter. There's easily room for both. And more important, the competition between the two will give a lot of users a chance to take cloud computing for a low-risk, real-world test drive.
I welcome your comments, tips and suggestions. Reach me at bill_snyder@infoworld.com.
Posted by Bill Snyder on June 19, 2008 04:00 AM
April 10, 2008 | Comments: (0)
Amazon, better known for peddling books and CDs than selling leading-edge technology, surprised much of the tech world by rolling out, and actually attracting customers to, its version of cloud computing. Sure, there's still plenty of reason to be skeptical -- indeed, cloud computing is one of those technologies that can sound more like a buzzword than a solution to real-world IT problems -- but other players are jumping into the game.
And that's good news for business. Competition will force vendors to keep prices low, improve the technology, and, more importantly, prove once and for all whether cloud computing is more than, well, pie in the sky.
Earlier this week Google launched App Engine, a service to host Web applications. Google App Engine is still in diapers; at the moment, it only supports applications written in Python and only has room for 10,000 developers to use the service. But it's likely to grow. IBM's Blue Cloud will go live later this year, though it's more geared to selling tools to build clouds than to hosting them, and Salesforce.com is beginning to position itself as a cloud vendor with Force.com.
Who's next? EMC.
"We will compete with Amazon," says Joe Tucci, the storage giant's CEO. "There will be an EMC storage cloud; in fact, there already is," he said during a meeting with a small group of reporters at the RSA conference in San Francisco.
EMC's bid for the cloud
When Tucci said that EMC has a storage cloud, he was referring to Mozy, the company's online backup and recovery service, acquired in the buyout of Berkeley Data Systems. Tucci's ambitions go far beyond simple, online backup. "Big companies can't do small things. If I don't see a $1 billion business, I don't go into it," he said.
Like Amazon, EMC's cloud will include both storage and computing resources. The timing of the offering isn't clear, but Tucci, who has broadened EMC's capabilities by buying 37 or so software, document management, and security companies, isn't given to vaporware pronouncements.
EMC's cloud play will focus on businesses through its computing and storage offerings, but the company sees an opening on the consumer front as well.
Driven by the bottomless well of content produced at home by digital cameras, MP3 players, and DVRs, a typical consumer may well have as much as a terabyte to manage before too long.
"If I'm a consumer, I'd rather trust EMC to [secure] my data than myself," says Art Coviello, who manages EMC's security business. That may well be true, but won't IT execs, skittish about losing control of their own data and resources, be a harder sell?
"It's not as if you don't have outsourcing today; it's not as if you don't have a Salesforce.com today," replies Coviello. Fair enough, but rightly or wrongly, many IT execs view those strategies as much less radical than cloud computing.
And remember, we're not just talking about parking data; it has to be securely accessible to a variety of employees, partners, and, in some cases, customers.
Earlier this year, I spoke to Doug Menefee, CIO of the Schumacher Group, which finds temporary staffing for hospital emergency rooms across the country. Menefee is an early, and generally happy, adopter of Salesforce's cloud technology, but he's quick to say, "Single sign-on service and password management were the biggest pain points."
With all due respect to EMC’s execs, it surprised me that they weren’t more specific in their discussion of security risks in the cloud. Tucci acknowledged that IT managers may well keep the most sensitive data at home and out of the cloud, and said “there’s no reason why we can’t create a secure cloud.” That’s nice, but I’d like to hear more. Similarly, Coviello, who has a very thorough knowledge of the threat landscape, was awfully casual, if not dismissive, in answer to questions about the risk to data, saying “I’d never suggest there is a perfect security system.”
I’m not suggesting that EMC doesn’t care about security. Far from it. But making the cloud acceptable to big-time IT is going to take some serious, and very concrete, thinking.
Security aside, there's another business opportunity here as well -- selling hardware and software to enterprises big enough to build and manage their own clouds. That's exactly what IBM will be doing via Blue Cloud, and Tucci figures that EMC will partner with customers in financial services and other big verticals to do that as well.
The move to cloud computing is obviously gaining some traction, but more than one "next big thing" fell flat when it failed the test of real-world use. I'm not making any predictions here, but keeping an eye on this nascent trend is likely a good idea for investors as well as serious consumers of technology.
(Disclosure: I own a small number of shares in EMC)
I welcome your comments, tips and suggestions. Reach me at bill_snyder@infoworld.com.
Posted by Bill Snyder on April 10, 2008 03:00 AM
March 13, 2008 | Comments: (0)
Cloud computing begins to emerge from the haze
I hate technology buzzwords. And when I hear one, my impulse is to bat it away like an annoying mosquito. But before you do the same about one of this year's hot buzzwords -- cloud computing -- give it a little more thought.
Cloud computing, a concept that can be as airy as its name suggests, is piquing the interest of forward-looking IT execs and attracting sizable investments from players like IBM, Amazon, Akamai, Sun, EMC and Salesforce.com. Sure there's a big helping of hype and plenty of reasons to be skeptical, but a growing number of startups -- and a still small number of enterprises -- are moving applications and infrastructure into a third party-provided cloud.
What's driving cloud computing? Out-of-control costs for power, personnel and hardware, plus a shortage of space in datacenters and a desire to speed up and simply network deployment and management. What's enabling it? Nearly unlimited bandwidth, increasingly sophisticated virtualization technologies and multitenant architectures, and the availability of extremely powerful servers.
The emerging resource cloud
Cloud computing is an option when there's a need to complete a resource-heavy project without buying hardware or hiring personnel that won't be needed later. The New York Times, for example, used Amazon Web Services (EC2 and S3) to generate PDFs of 11 million articles in the paper's archives.
"Honestly, I had a couple of moments of panic. I was using some very new and not totally proven pieces of technology on a project that was very high profile and on an inflexible deadline. But clearly it worked out ... I can't imagine how we might have done it without Amazon S3/EC2," Derek Gottfrid, senior software architect for the Times, wrote in his blog.
The Schumacher Group, a Lafayette, La.-based company that supplies temporary emergency room staffers, needed more flexibility to add resources for satellite offices and wanted to decentralize part of its operation as a hedge against hurricane-induced outages. Schumacher combined a custom application with a Salesforce.com CRM application to handle thousands of contracts among his company, the hospitals and the doctors -- and he runs it on the Salesforce infrastructure. The savings in avoided operational and hardware costs were significant.
Rather than buy enough servers and other infrastructure to meet peak needs, Powerset, a startup building a natural language search engine, fills the gap with Amazon's EC2 and pays for the resources as it uses them.
(Read a more detailed account of deployments by Schumacher and other companies.)
So what is cloud computing, anyhow?
Not only is cloud computing a very young technology, it's still very loosely defined. Indeed, it's sometimes hard to distinguish "the cloud" from conventional hosting, software as a service, or grid computing. For many, it simply means "something done outside my walls." What it means in practice is a collection of resources -- applications, platforms, raw computing power and storage, and managed services (like antivirus detection) -- delivered over the Internet.
In a just-published report, however, Forrester Research analyst James Staten lays out a pretty good working definition for technologists: "A pool of abstracted, highly scalable, and managed compute infrastructure capable of hosting end-customer applications and billed by consumption."
Staten, by the way, is not beating the drums for instant adoption. In fact, his report opens with this sentence: "Cloud computing is a new IT outsourcing model that doesn't yet meet the criteria of enterprise IT and isn't supported by most of the key corporate vendors." He does, however, argue that "infrastructure and operations professionals can try to ignore it, as it is just in its infancy, but doing so may be a mistake as cloud computing is looking like a classic disruptive technology."
The Forrester report cites a number of concerns that must be erased, or at least mitigated, before the technology is deemed enterprise-ready. They include worries about stability and security; the lack of big-name players offering clouds (IBM, for the moment at least, is selling tools to create clouds but does not host one); a lack of reference accounts; and thin support by software vendors.
"Given that most clouds are unique infrastructures (and in the case of Amazon, one it won't tell you a lot about), most commercial operating systems and applications are not certified on these platforms. Only XCalibre FlexiScale supports Windows, for example. Given that the infrastructures are virtualized, licensing is another issue," Staten writes.
Dipping a toe in the clouds
Despite all these issues, some enterprises are experimenting with clouds -- with or without the approval of IT management, says Staten. Clouds are being used as a way to test new services and applications and meet spikes in demand.
Cloud computing as a mainstream technology is not just around the corner. It will take a few years, and any number of potential failures, for serious IT users to separate the hype from the useful reality.
But don't be too quick to dismiss it.
Posted by Bill Snyder on March 13, 2008 03:00 AM
