Federal Express faces complex computing problems every day. For example, it constantly needs to determine which driver takes which package to its destination, and often has less than an hour to do so.

In the past, FedEx dealt with this challenge by using ever more powerful hardware, scaling up its systems as necessary. But this approach meant buying more and more expensive hardware. It also reinforced a silo, "server-hugging" mentality in which each application relied on its own system resources.

What FedEx really needed to do was scale horizontally and take advantage, for that one hour, of all the many machines the company already had. With this approach, it did not need to buy more hardware until the aggregate demand exceeded the aggregate existing capacity. It also ended the relentless hardware replacement cycle of typically 3-4 years.

FedEx is widely recognized as one of the world’s most sophisticated IT shops, so it came as quite a surprise that Appistry could so quickly provide an abstraction layer between FedEx’s applications and its hardware that enabled the company to pick and choose hardware resources as necessary in a distributed, horizontally scalable way. In effect, FedEx could now run its systems as if it had a pool of servers that it could access on demand.

Appistry also helped FedEx to simplify its support model. The FedEx internal system is now treated as one entity, even though it consists of many physical servers. This is a good example of how very large organizations can choose to develop their own, private, cloud-like systems.

Find out how to get ready for the cloud