Amazon at Your Service

By Edward Cone  |  Posted 01-07-2007

Amazon at Your Service

Thousands of businesses from around the world outsource their e-mail management chores to Webmail.us Inc., a Blacksburg, Va., company that deals with everything from maintaining mail programs to filtering spam and squishing viruses for clients. Webmail.us runs its own operations on a couple of hundred servers at data centers in Virginia and Texas owned by Rackspace Managed Hosting, based in San Antonio, Texas. But when Webmail.us needed to add short-term storage capacity and build more redundancy into its backup of primary data earlier this year, it turned to another vendor: Amazon.com Inc.

Amazon's new Simple Storage Service, or S3, has lived up to its name, says Bill Boebel, Webmail.us's cofounder and chief technology officer, even though the numerous small files typical of e-mail create inefficiencies that conventional backup systems don't handle very well, and Webmail.us is sending S3 more than a terabyte of data each week. "Amazon let us move faster than anything else out there," says Boebel. Working with a traditional hoster might have meant writing a custom application to accept the e-mail data, but Amazon required only a relatively simple interface to communicate with its new customer. "They provided back-end storage that we could build on top of," adds Boebel.

To host the development effort required to build the new storage interface, Webmail.us used another recently introduced Amazon service, a pay-per-usage computing environment called the Elastic Compute Cloud, or EC2, and the company continues to use EC2 for computing tasks related to storage backup. Says Webmail.us chief executive Pat Matthews, "We love Rackspace; they are our biggest strategic partner. But Amazon met our needs for this job, and we cut our data backup costs 75 percent overnight." S3 costs 15 cents per gigabyte of storage per month; Amazon's posted prices for EC2 start at ten cents per hour.

Quicker, easier and cheaper—that's the message Amazon wants companies to hear about its new offerings as it challenges the hosting industry's long-dominant pricing and service models. The e-commerce pioneer is renting out the muscular infrastructure it has developed to power its own business over the last eleven years (sales for the third quarter of 2006: $2.3 billion), in order to provide storage, computing and other services to all comers.

If the strategy pays off, Amazon will have realized a dream cherished by many CIOs: taking its own internal technology to market as a commercial product. The direct financial return to Amazon in the near term is likely to be minimal, as CEO Jeff Bezos told BusinessWeek in November. But there is value in running the company's data centers and related businesses at levels closer to full capacity, and in positioning Amazon as a leading provider of Internet services to other companies.

"Amazon is a technology company," says Adam Selipsky, vice president of product management and developer relations for Amazon Web Services, the unit of the Seattle retailer that is responsible for the new offerings. "We built up this platform, this series of technologies. Now we are allowing other companies to run inside Amazon data centers, so we can fully pass along the economies enabled by our platform, whether to one-man developers or large corporations."

Amazon's services go beyond the bounds of traditional hosting and what is sometimes called utility computing, says Selipsky, because of the wealth of engineering and experience the company brings to its customers. Need storage capacity, or a few dozen servers to test a new application or power your business? Just access Amazon via the Internet and pay as you go. No contracts, no guesswork on capacity, no renegotiation to accommodate more or less usage than planned. This is hardware as a service, with all kinds of expertise baked in.

"It's not just about a whole bunch of capacity," says Selipsky. "We are giving customers the ability to work at Web scale"—that is, to handle large amounts of online data from diverse sources in a near-real-time environment. Says Selipsky: "We have been doing this a long time, so you don't have to make the same mistakes we did. Amazon.com is one of the true Web-scale applications out there, and it is non-trivial to build something like that." Even if a company has the capital and engineering to create such a platform itself, he says, dealing with multiple versions of software, legacy hardware and so on, make it a complicated and daunting task. "We take out the complexity, you pay as you go, and that turns out to be very appealing," adds Selipsky.

It is an approach that will change the hosting business, says Daniel Golding, a vice president with Tier 1 Research, a Minneapolis-based firm that covers technology, Internet and telecom companies. "The long-term contracts that have defined the industry are dinosaurs," he says. "Hardware as a service is the way of the future. Managed hosting vendors will bill on usage, or customers will pay to guarantee a certain amount of computing capacity. This whole idea—the fast-food approach—is going to spread throughout the industry."

But Amazon may not end up being the big beneficiary of the trend it is helping to launch. Golding doubts that enterprise customers will be comfortable dealing with a retailer that lacks experience serving large business customers to host their core applications and information (large corporations, including Microsoft Corp. and Xerox Corp., have been early customers of Amazon's services). S3's streamlined storage could make it attractive, he says, "at least until there's more competition using the pay-by-the-drink model." Golding expects that more established players, such as IBM Corp. and Savvis Inc., will eventually enter the market, directly or as hosts to smaller service providers.

"When the big hosters divest the idea of locking up revenue with long-term commits, the shift is inevitable," Golding says. "The larger story is what this means for the industry, not what it means for Amazon. Frankly, I'm not even sure they should be doing this from the point of view of their own business." Wall Street, already nervous about Amazon's thin margins and high price-to-earnings multiple, did not reward the company for its move; shares are down about 20 percent from the start of 2006, and almost half the sell-side analysts following the company have neutral to negative outlooks on the stock.

Laurie McCabe, an analyst with AMI Partners Inc., agrees that hosting companies will have to pay attention to the Amazon model. "The bottom line is that a lot of the bigger names have given lip service to on-demand service, but the pricing has not been quite so utility-like," she says. "They will have to keep close tabs on this." While Amazon may end up as something of a niche player, providing computing power for software developers and startups, she says, "there's a lot of money in that kind of development." And, she adds, Amazon's expertise in offline services like fulfillment may give it some competitive advantages against more traditional hosting vendors, who lack that kind of business-process experience.

But Amazon has bigger plans for its services. So far, says Selipsky, demand for services is varied, and stronger than expected. "The sheer variety of what people are doing is surprising," he says. "We're building these things very horizontally, so they are broadly applicable." At least some early customers believe the services will succeed in a big way. "Amazon specializes in high-volume, low-margin businesses," says Barney Pell, founder and CEO of San Francisco-based Powerset Inc., a much-buzzed-about new search-engine company that uses both EC2 and S3. "They've been very responsive to our needs, and I think they will win over a lot of people."

From Retail to Services

From Retail to Services

Amazon's Web services unit was created in 2002 to help developers build applications that allowed the company and affiliated vendors work more smoothly and sell more product. "At first, we primarily exposed our product information and data to our affiliates and associates," says Selipsky. "It was really a productivity tool for our associate base, and it unleashed an explosion of innovation, with people creating a lot of cool applications."

Over time, Amazon has gone from selling books and other products to opening up its platform to merchants who sell their own stuff, to enabling a developer community. The next move seemed natural—turning the guts of the $2 billion system—especially since Amazon, like many companies, uses only a small fraction of its computing capacity at a given moment. "We saw we could help people, and we saw a potentially interesting business for us," says Selipsky. "We thought, maybe there's other stuff Amazon can open up. What else could we expose or build in terms of Web services that would help make developers' lives easier?"

S3 was launched in March 2006, with demand strong from the outset. When a beta version of EC2 went live in August, says Selipsky, "I don't think we had fully thought through its utility for testing software. If you want 50 or 100 boxes pointing at an application and pounding on it, that's really expensive to set up if you only need it for an hour. Instead, you can use EC2 and feel like you have the scale of Amazon. It's disruptive to the way developers currently develop applications, when and how much they make capital expenditures, the level of quality and how they scale."

A third service, called Mechanical Turk (named for an 18th-century chess-playing contraption with a human concealed inside), combines computing power with networks of real, live people who are paid to do the kind of work machines can't yet do well, such as recognizing inappropriate content in images or transcribing audio. A company that needs to post hundreds of photos to its Web site, for example, or identify key portions of a podcast, might use the service. "It's a marketplace for intellectual capital, and we're the marketplace for the workforce" says Selipsky. "Pieces of work get posted onto Mechanical Turk, and people get paid online by the entity that requests their help. We get a percentage of the fee paid." (Amazon takes a 10 percent commission.) A fourth service released in 2006, Amazon Simple Queue Service (SQS), allows developers to transfer data efficiently between application components.

Simplicity and ease-of-use are a big part of what Amazon is selling. Users program to Amazon's services by means of Application Programming Interfaces (APIs) the company makes available, so some expertise is required, but nothing exotic. That's exactly how Amazon designed things, says spokesman Drew Herdener. "All of the services are intentionally primitive and feature-poor—'close to the metal,' we say." Amazon is close-mouthed about the details of its architecture. The company runs all its operations on grids of commodity servers at multiple data centers across the U.S., and uses service-oriented architecture to insulate outside users from the workings of the underlying systems.

And more services are on the way. "As with most good ideas, we come at things from two different directions: What do people need, and what can we do well?" says Selipsky. "We look internally and ask, 'what do we have built, or what can we build with our world-class engineering talent that would justify creating a business?' "

Amazon prohibits illegal activity on its servers, and not much else. "Our terms of use are broad and there are no contracts, so if you want to stop paying you just take your data out of storage," says Selipsky.

In the larger view, says Selipsky, Amazon's moves go beyond Web services to a "long shift—from proprietary ownership of everything, to opening up and focusing your efforts on things that really add value to your customers, whether you are a big company or a small one. The entire open-source software movement is part of the same phenomenon. The emphasis now is on nimbleness and speed, grabbing things other people have done."

Turning On the Tap

Turning On the Tap

Powerset, the search-engine company using EC2 and S3, dreams of challenging established competitors such as Google Inc. by offering searches that use natural language, rather than stilted phrases and imprecise key words. It is a task that involves intense computational capacity. The young company, which has raised $12.5 million in funding, wants "to avoid spending resources on anything else than our core business," says CEO Barney Pell. Time was also an issue for Powerset. "There is a risk that building your own infrastructure can take longer than you thought," says Pell. Instead, he outsourced most of his computing and storage jobs to Amazon.

Amazon was significantly cheaper than other large data-center companies Powerset investigated, says Pell. "A number of the big players charge about one dollar per CPU hour—we are paying a fraction of that." Powerset might have gotten its costs down as low as five cents to seven cents per hour by building its own systems, says Pell, without factoring in development or human-support costs. "If we purchased our own equipment and amortized the cost over three years, the cost per CPU hour would be competitive. But it's easy to write it down on an accounting basis, and you still have to spend the cash. This is like taking the amortized number without spending the cash. Amazon does the investment for you. It's a fantastic value proposition for us," Pell says.

Beyond the price, the payment model is a critical part of the equation, says Pell. "Elasticity is important. We don't know how fast demand for our service will grow, and we don't know how many queries we will need to support each second," he says. "Users come in bursts—there are sudden spikes of traffic and interest. We can model demands and the needs of hardware, and we're certainly going to be wrong. If we overestimate peak capacity, we spend too much money; if we underestimate, we will be turning away users or giving them a suboptimal experience. But a system that identifies bursts of traffic and lights up second and third strings of servers solves that problem."

Powerset plans to launch for public use in 2007, and is using EC2 for its intensive offline activities: the background work of reading, processing and indexing Web pages that underlies search. "The computers doing that work could almost be buried underground and not even connected," says Pell. Still to be decided: whether Powerset will rely on Amazon for the "run-time" application that powers actual searches. "We are investigating if we can make serving user queries work on EC2—if our technology works in their infrastructure," says Pell. "That could be another massive game changer for us."

In many ways, the user experience with these services is different from a traditional hosting relationship. Webmail.us executives have never visited an Amazon facility, for example. "It's all self-service," says Boebel. "We started playing around with it, then engaged with one of their business development people. They worked with us pretty closely." S3 involves no client software and no hardware on the user's site. "There is just some simple code to execute on your server, and you're done. You could set up a server to back up to S3 in an hour and never touch it again," he says. Unlike traditional rack rental, says Boebel, Amazon has intelligence built into its network of servers. "The logic is there to handle failure. You can't just lease a server to do that."

Another difference: Amazon doesn't offer service-level agreements that guarantee the amount of time services will be available. The company argues that its own reliance on the platform is enough be give outside users confidence. "Amazon is a customer of our platform," says Selipsky. "That's important philosophically. We believe if it's good enough for our customers, it's good enough for us to use. We have mission-critical data stored in S3, and we have groups using EC2. It's important that we really understand all the good things and bad things we put out there, and we get incredibly good feedback from our own people."

That's good enough for Webmail.us, up to a point. "I'm very comfortable, because they rely on the same infrastructure for their own business," says Boebel. "If S3 goes down, they're hurting." Webmail.us has not seen any downtime since it went live with the storage service in September 2006; in early January, Amazon S3 did experience its first significant outage. Boebel likes the fact that Amazon stores multiple copies of its data, including at least one copy in a separate data center. But Webmail.us CEO Pat Matthews says that SLAs of the sort offered by Rackspace matter a lot for his core business, and he expects other companies may find the absence of SLAs more of an issue.

In the long run, though, says Tier 1 Research's Golding: "SLAs are useless. An SLA is not a guarantee of good service, it's a bad rebate. Companies architect a solution as well as they can, up to the limits of the existing technology. In the on-demand world, it makes sense to do away with them. If customers don't have a commitment, and they don't like what they're getting, they will go somewhere else."