A September 2001 article in InternetWeek named Exodus Communications, EDS, and IBM as the "big three" of the U.S. hosting market, and Gartner analysts predict a compound annual growth rate of forty percent through 2005 for the Web hosting market. Yet Exodus filed Chapter 11 on Sept. 26, leaving its customers in a quandary.
TechRepublic interviewed Gartner analyst Ted Chamberlin and EDS vice president for Hosting Services, Steve Lapekas, regarding the fallout from Exodus' exit and the emerging landscape of hosting services. We asked them what the future holds for Exodus and its customers and who will emerge as the new leaders in the hosting space.
What happened to Exodus?
In "Hosting services: The price is right for enterprises," Chamberlin and Gartner analyst James Browning explained that colocation providers, like Exodus, Colo.com, and Level 3 Communications, built redundant facilities near network backbone connections, which required costly real estate. The plan was to be close to their customer bases, consisting mostly of dot coms and Internet-focused companies who lacked the capital to build their own facilities. In the Web-hosting boom of early 2000, the data centers were leasing 50 to 70 percent of their space to businesses at a rate of $100 to $130 per square foot. When the U.S. economy began to sour, however, these "pure play" hosters were hit hard, as those prices fell from 40 to 60 percent below peak rates. Additionally, the client base was disappearing as those businesses fell victim to the economic decline.
Chamberlin said the explanation for Exodus filing Chapter 11 is simple: too much debt combined with poor utilization rates on the company's data centers.
Lapekas agreed that poor utilization was a big reason for Exodus' downfall. He compared his company to Exodus, saying that while EDS has 12 data centers with 1.4 million square feet that's filled near its limits, Exodus has 44 data centers with 5.6 million square feet—but much of it is vacant. Despite its estimated 14 to 15 percent market share, some analysts estimate Exodus' data centers are filled to less than 50 percent of their capacity.
Additionally, Lapekas said Exodus had to build those data centers at a cost of $10 to $30 million apiece, whereas EDS only had to add the Internet functionalities to its existing data centers.
"To be in this business, just supplying floor space, rack [servers], and connectivity is not profitable," Lapekas said. "You have to fill your data center to the very limits and be pushing mass amounts of data."
Is it time to throw in the towel?
After talking to some of Exodus' larger clients during the course of his work, Chamberlin said, "There is no panic," regarding Exodus' ability to continue providing services. However, he said some customers are leaving to find hosters with better financial footing and stability.
Chamberlin said he thinks filing Chapter 11 was an appropriate decision for Exodus, as it may allow them time to reduce their debt and realign their strategies.
In a Gartner report, Chamberlin said certain segments in the hosting arena, such as managed and professional services, would continue steady growth, while the colocation and simple hosting segments will see lethargic growth, or even decline. In response, "large network service providers who have been slow to acquire the requisite skills and resources to provide day-to-day server management will look to acquire the smaller pure-play hosting providers." As an example, Chamberlin cited British telecom Cable & Wireless' purchase of Digital Island.
Rumors began circulating about Exodus being purchased by another entity around Sept. 4 when the company ousted CEO Ellen Hancock in favor of William Krause, ex-chairman of 3Com and a member of Exodus' board of directors. Some predict that a Telco like AT&T or Sprint might be interested in purchasing Exodus to acquire its data centers. Lapekas said this might be the case because telcos will be looking to increase their offerings to compete in the hosting sector.
"The telcos, though a very established sector, definitely have their limitations when it comes to anything above the operating system," Lapekas said. "They run the central office and they run the connectivity piece very well, but as complexity grows, they're definitely challenged. I think you'll see this from the acquisition trail."
One rumor that circulated was that EDS was considering purchasing Exodus. Lapekas said that while EDS does consider such acquisition opportunities from time to time, the company always does extensive due diligence.
"As far as my personal opinion, [Exodus has] significant debt, and I think that's why you haven't seen any major moves by anybody," Lapekas said. "There's no question that they have a very substantial client base. Some of it would fit into EDS's model very well, and some of it wouldn't."
Whose future looks bright in the hosting space?
Gartner's last "Magic Quadrant"—an evaluation of ability to execute and completeness of vision—for the North American Web hosting market was published in January 2001. It showed IBM and Digex as the industry leaders. Chamberlin said he and other analysts are working on a new Magic Quadrant that's likely to show IBM and Digex remaining as leaders but with most others losing ground. One exception, he said, was up-and-coming player Loudcloud, a company that came in "just at the right time."
How has the landscape of hosting been changed?
Chamberlin said Exodus' demise, in addition to the economic downturn, has caused a shift in the way companies will choose their hosting and Web services providers.
"They used to make decisions based on who could get you up and running the fastest, but now they're forced to consider the financial stability of the company," Chamberlin said.
Any new players hoping to grab a piece of the hosting pie will have to have substantial financial backing and a recognizable name for "mind share," Chamberlin said. For example, IBM is a proven player in the hosting space because it has the financial support to sustain its hosting services arm until it's profitable, and the company has a name that decision makers have learned to trust.
Lapekas said he believes the hosting market is viable but that industry leaders will be those with more holistic Web service offerings.
"The IDCs of the world are valuing the market at around $16 billion by 2004," Lapekas said. "We still think servers need to be racked, there's still a need for the data center type of environments, it's just that the managed services piece is probably the most important."
Protecting your business from bankrupt hosters
Exodus customers, as well as customers of other struggling hosting firms, have to make tough decisions about how to handle their data and equipment.
"When you have a server environment, you pick a home for it, and then you have a situation like Exodus—it’s very challenging for the client," Lapekas said. "You can't pack it up, put it in the back of your car, and drive down the street."
While those decisions have to be made on a case-by-case basis, Gartner advises that companies in need of Web hosting services should educate themselves about the boundaries and implications of service-level agreements (SLAs), which are written to protect both the hoster and the client. "SLAs that promise 100 percent overall uptime or have unreasonable limitations on the payout of penalties are examples of SLA metrics that are used to limit liability on the part of the provider."
Additionally, Gartner recommends choosing providers who have six- to twelve-month cash reserves and target dates for reaching profitability. Those who have taken a hard line toward reducing their burn rate will be more likely to deliver reliable service during the coming year.
What are your concerns?
How do you believe the hosting landscape will change over the next year? What precautions will you take to protect your business from a Web hoster gone bankrupt? Send us an e-mail or post your comment below.