Work on Xen began at the University of Cambridge in April 2002. It was initially developed as part of the Xenoservers project, which aimed to create a “global distributed computing infrastructure.”
Around the same time, grid computing was being advanced as the best way to make use of computing resources that are scattered throughout the world. The original grid proposal cast computer time as a utility, like electricity, which could be obtained from a grid—or network—of collaborating computers. However, subsequent implementations concentrated on virtual organizations: groups of companies and institutions that established possibly complicated relationships of trust, which are enforced by heavyweight public-key cryptography for authentication and authorization.
Xenoservers approached the problem from the opposite direction. Instead of forging trust relationships with service providers, the customer chooses a resource on the open market through a broker known as a XenoCorp. The XenoCorp stores a list of xenoservers—computers offered for lease by third parties—and matches customers with servers, collecting and passing on payment for the utility. Crucially, there is mutual distrust between the customer and the provider: the customer cannot harm the provider’s machine, and the provider cannot tamper with the customer’s job.