The Cloud`s the Limit With Anything as a Service
Last September, Shepherd Data Services, a Minneapolis-based provider of discovery services for law firms, was asked to convert 14GB worth of Lotus Notes data into TIFF files in just four days. It was an imposing task that was critical to a client’s civil litigation case.
Wade Peterson, Shepherd Data’s director of IT, mobilized a few of his staff and used the firm’s internal electronic discovery software to sift through the data. But it soon became apparent that the scope of the project was more than 10 times what was expected—a whopping 160GB comprising 3.2 million pages—and would require 20 people to work on it around the clock.
Earlier in the year, Peterson had signed the company up for a subscription to IPro Tech’s eCapture Web-based discovery software, and he and Christine Chalstrom, Shepherd Data’s founder and president, agreed that this assignment was the perfect test for this service. Peterson and a small team uploaded the data into eCapture, and the software began churning through the data much faster than the in-house system could—and with far fewer employee resources needed. When the deadline arrived four days later, the discovery effort was complete, and the resulting TIFF files could be retrieved online with a simple log-in.
“It felt like a whole data center was at our disposal,” says Chalstrom. “This allows us to send information to the cloud, get it processed and have our customers access it there. Before, we’d get data in, we’d process it, shove it into an external hard drive, take it to our customer’s data center and then upload it there.”
Ah, the cloud. It’s almost impossible to have a conversation with any business or IT decision maker without the topic of cloud computing surfacing. Cloud computing comes in an assortment of flavors, all of them labeled with a clunky acronym. There’s software as a service (SaaS), platform as a service (PaaS), infrastructure as a service (IaaS) and communication as a service (CaaS). It’s easy to see this trend expanding beyond the capabilities of the alphabet as companies avail themselves of ever more specialized services, such as discovery as a service or video as a service, which haven’t been labeled with acronyms—yet.
All together, the market for cloud services totaled $46.4 billion in 2008, and that figure is expected to grow to $56.3 billion in 2009, predicts IT research firm Gartner. Such growth, especially in this slow economy, would seem to indicate that any business function that’s technology-enabled is a candidate to be offered as a service. Hence, the all-consuming moniker, XaaS, or anything as a service. What started with the migrations of early adopters toward SaaS applications offered by Salesforce.com, NetSuite and the like has become a fast-maturing IT strategy that has IT and business leaders considering any potential service offerings when evaluating new technologies.
And while cloud computing’s bread and butter has been small companies, that’s been changing. “All these systems appeal very much to the small to midsize companies, where you don’t have a mature IT organization,” says Anne Thomas Manes, vice president and research director for IT advisory firm Burton Group, based in Midvale, Utah. “But large organizations also are very interested in the potential of the cloud.”
In some cases, things have moved well past the interest stage. “Cloud computing has changed the way I—and my peers—think about hosting decisions,” says Mike McGarry, CTO for $9.9 billion financial services giant Genworth Financial, based in Richmond, Va. “In the past, our bias was to buy a package first, and if we couldn’t find the package, then we’d build it in-house. Now, our first choice is, hey, can we do it in the cloud?”
Financial services firms have a well-established, and reasonable, resistance to XaaS offerings, citing security and privacy concerns. But even they are finding niche offerings that can help.
That’s the case with Genworth. Formerly GE Financial Insurance, the company was spun off from GE and renamed in 2004, but it had the option of having GE Asset Management continue to manage its IT systems. The executive team decided Genworth should manage IT itself, but it lacked an IT team. So the firm turned to Oracle’s on-demand applications to handle its financial and human resources functions, which contained low-risk data that didn’t cause privacy or competitive concerns.
In 2006, Genworth turned its attention to the more sensitive issue of managing its investment portfolio. The company needed a system in place quickly, and given its experience as a cloud service subscriber, it didn’t take long to zero in on the idea of an XaaS option. That led to Aladdin, a service offered by investment management firm BlackRock.
The Aladdin subscription went live in 2007 with a risk-reporting module that provides Genworth with holding-specific reports that help it manage its portfolio. Last year, it added trading of corporate bonds and other asset classes, as well as integration of the third-party investment managers it works with. If this sounds like a risky thing to place in the cloud, consider that the Aladdin service is not a “multitenant” offering, or one in which individual servers support multiple clients. Genworth has its own dedicated instance of the service.
“It’s not as if every client is running on the same database,” says Kevin Gordon, CIO of corporate systems. “There’s no oppor-tunity for commingling of our data with our competitors’.”
The company declined to discuss financial specifics of its deployments, but McGarry and Gordon have a new appreciation for XaaS that’s likely to change their thinking going forward. For instance, over the past several years, the company has relied on a grid computing strategy to support its actuarial business, harvesting CPU cycles on its thousands of servers and desktops for use when analyzing myriad potential business and economic scenarios. Sometimes, however, that’s not enough. “As we look for additional compute power beyond what we can do with harvesting,” Gordon says, “we’re looking at cloud computing to see if we can supplement that during spikes in demand.”
Managing Spikes in Demand
Few things bring more dramatic spikes in demand than online video, and companies that consider online video either a core component of their business or a growing part of their marketing mix are making the most of whatever services can help. In the case of the Museum of Broadcast Communications in Chicago, a storage-as-a-service vendor’s ability to improve the performance of the museum’s video streams has been a lifesaver.
Since it opened in 1987, the nonprofit museum has had a mission to preserve historic television programs in a way that would provide easy access. By 2004, the museum had digitally encoded 6,500 shows, and founder and President Bruce DuMont decided to upload that digitized content to the museum’s Website, where it would be available for free streaming.
After several weeks, the museum started getting irate e-mails from users having buffering problems. In February 2008, all the programs were pulled from the site. “My desire to share this content with as many people as possible worldwide was stymied by the technological challenges of what we had,” DuMont recalls.
Fortunately, one of those early users was Chris Gladwin, CEO of Cleversafe, a startup that provides “dispersed” storage as a service, in which data is sliced into unusable packets and then reassembled on the fly to respond to user requests. Gladwin felt this technology was the ideal tonic for the museum, and with his staff’s help, the museum began uploading Windows Media Video files to Cleversafe’s dispersal network, while its high-resolution master files remained on its own network.
The impact was immediate and dramatic, with users reporting a 180-degree change in their experiences. Since then, the roster of registered users has swelled more than 40 percent, to more than 20,000. In the months since it began working with Cleversafe, the museum has uploaded more than 10,000 hours of content, or more than 15 percent of its library.
In the case of Kohler in Kohler, Wis., storage as a service wasn’t enough to support its fast-growing video initiatives. With business units offering everything from the company’s kitchen and bathroom fixtures to golf resorts, engines and generators, Kohler has more than 50 Websites on which video is a growing component. Rather than investing in a comprehensive video-delivery infrastructure, the firm decided a few years ago to work with Brightcove, which provides end-to-end video services.
Doing so required just a six-figure annual investment—one-tenth of what it would have cost to build and maintain that video infrastructure, estimates John Engberg, manager for global media and Web development at Kohler. It also kept a time-consuming obligation off of Engberg’s team’s plate. “Having all that in the cloud, off of our servers, without having to worry about response times worked for us,” he says.
In 2008, visitors to Kohler sites watched 4 million videos on topics such as faucet installation, golf course design and home décor. Engberg says the firm is very happy with the quality of the video streaming customers are encountering. Recently, Kohler tapped Brightcove to support a campaign to reduce water usage. After viewing a brief video about the benefits of saving water, site visitors are prompted to take a quiz about their own water-use habits. As part of the campaign, Kohler is donating $500,000 worth of water-saving products to Habitat for Humanity, and it will add an additional $1 for each quiz completed.
Kohler’s effort demonstrates a model for future cloud adopters: tapping the cost-efficiency of the cloud to create inexpensive, temporary public-education and fundraising campaigns.
Counting the Costs
Environmental causes aside, Burton Group’s Manes cautions companies to be certain of the financial benefits of cloud services before leaping into them. She says companies often fail to consider long-term lease commitments on data center hardware that cloud services will replace. Also, they may not be sure of the ongoing costs of the applications that run on those servers.
“Very few organizations know how much they’re spending per year on an application,” says Manes. “So how do you justify shipping something into the cloud? If you don’t know what it costs now, how do you know what you’re going to save?”
Most of the latest hype around cloud-based services has focused on meeting spikes in computing demand and establishing or extending an infrastructure at a moment’s notice without investing in hardware. One of the most compelling use cases for the cloud is in ensuring that businesses can stay up and running during crises. Yet, such infrastructure services are part of a submarket that’s expected to account for just $3.2 billion, or 6 percent, of the cloud market in 2009, Gartner predicts. Still, it’s a market that large IT vendors appear to be counting on to grow.
No one has to sell Jessica Carroll on the virtue of cloud-based infrastructure. Carroll, director of IT for the United States Golf Association, which hosts the U.S. Open, was already looking into upgrading the USGA’s antiquated tape-backup system when the phone infrastructure went down for a couple of days last summer. The outage barely registered as a blip, with employees switching seamlessly to cell phones, but she realized that if an outage hit the e-mail system, the outcry would be immediate.
Carroll decided it was time to add e-mail continuity to the USGA’s disaster recovery mix. But investing in a new data center, or purchasing the 70 servers needed for the existing data center, was prohibitively expensive. Instead, last fall, the USGA contracted with IBM to subscribe to a disaster recovery service, which allows it to back up 500GB of data at any given moment, and to an e-mail continuity service that essentially mirrors the USGA’s e-mail system in an IBM data center.
The recovery system was fully in place in March. The redundant e-mail environment was waiting for staff training—followed by live tests that replicate system outages—before going live.
USGA had been using Microsoft’s Live Meeting environment for three years and also used ePath Learning to host online training sessions. So it’s not surprising that Carroll joined the growing ranks of IT folks who are increasingly asking themselves, “Can’t we do this in the cloud?”
“I feel strongly that looking at cloud computing is a must,” says Carroll. “There’s a right time and a wrong time—not everything fits. But when you start talking about projects, you have to ask the question” about whether to use cloud computing. More often than you probably think, the answer will be “yes.”
Value Before Hype
Hype notwithstanding, just because a company finds a few logical uses for software as a service doesn’t mean it’s a fit for a more aggressive cloud computing strategy—yet. TiVo, the popular digital video recording pioneer, took its first tentative steps into the cloud when it began subscribing to an on-demand human resources application last fall. Recently, it added a subscription to an on-demand CRM app.
But those are probably the two most proven categories of SaaS, and they represent the extent of TiVo’s foray into the cloud—at least for now, says Richard Rothschild, senior director of IT. Rothschild says he’s taken a close look at Amazon Web Services’ S3 storage offering, but in Tivo’s case, “it doesn’t make sense, even economically.” That’s because TiVo built its infrastructure to contend with the company’s spikes in demand. “We’re not going to hit a capacity problem,” says Rothschild.
Still, Rothschild is keeping an open mind. He’s got a close eye on the Google App Engine platform as a service, for instance, in case the numbers start to make sense. “The moment I think it’s a better value for TiVo, I’m going to push for it,” he says.
The Real Cost of the Cloud
Cloud computing offers some real cost benefits, such as eliminating the need to spend capital dollars on infrastructure and only having to pay for what you use, reports Tony Treccapelli, the head of New York-based Alvarez and Marsal’s National Information Technology Solutions team. “This is especially beneficial for smaller companies and businesses that have been divested,” he says.
“The fact that costs aren’t fixed gives these organizations more flexibility to run their businesses. These firms aren’t in a position to develop their own infrastructure and, even if they were, they don’t have the staff necessary to manage it.”
However, Treccapelli adds that over the long term, it may be less expensive for some organizations—especially larger ones—to invest in their own infrastructure. “We estimate that the break-even point on these managed services is approximately three to five years for midsize companies and less time for larger enterprises,” he says. Treccapelli predicts it will be “well into the future before cloud computing becomes pervasive for large companies.”
What will bring more large enterprises into the fold? “Providers will have to increase both their user base and the number of applications that are cloud-ready,” Treccapelli says. “That will bring down the costs and make cloud computing more competitive for existing companies that already have an infrastructure in place.”