Summer vacations mean summer travel. That’s a good thing for the travel industry. For some companies, however, like Mobil Travel Guide, which is launching a new travel service later this year, it also means a huge upswing in its need for computing power.
Built using Java and IBM WebSphere and running on an Oracle database, the firm’s Mobil Travel Companion Service will offer drivers sophisticated Web-based travel planning services, as well as an enroute support network for making reservations or getting updated travel information.
Rather than buy servers for the new system, Mobil Travel Guide will tap into Linux “virtual servers” running on zSeries mainframes at IBM datacenters, paying only for the computing power and capacity it requires.
The five-year, multi-million dollar deal will let the company adjust to the huge swings it expects in volume.
“It’s like owning a virtual machine,” says Mobil Travel Guide CIO Paul Mercurio, “which we can shrink or grow as we need. That lets us handle the big summer driving season without more capital investment. It’s a huge advantage economically to be able to shrink it after peak season, so we don’t end up with servers running at one-third capacity.”
IBM’s service is not exactly like buying water or electricity from a local utility company, where a meter measures actual use and customers pay only for what they use. IBM customers purchase processing power in “service units,” a measure that equates to the processing power being utilized. Customers commit to buying a baseline number of service units, which IBM helps them determine based on anticipated demand. Customers can go 10% above the baseline without being charged.
Since the service is based on virtual machines, IBM can easily configure additional virtual servers if a customer wants to grow beyond that, according to Warren Hart, IBM Global Service’s director of e-business hosting.
Viewing computing power as a utility “helps line up your information technology expenses with your business volume,” says Hart. “You only pay for the computing power that you need.”