Computer Laboratory Home Page Search A-Z Directory Help
University of Cambridge Home Computer Laboratory
About XenoServers
Computer Laboratory > Research > XenoServers > XenoServer platform > About XenoServers

 
About the XenoServer platform

Home

About XenoServers

Live information

Deployment status

Downloads

Documentation

People

Contact us

Overview

The XenoServer project is building a public infrastructure for wide-area distributed computing. We envisage a world in which XenoServer execution platforms will be scattered across the globe and available for any member of the public to submit code for execution. The sponsor of the code will be billed for all the resources used or reserved during the course of execution. This will serve to encourage load balancing, limit congestion, and hopefully even make the platform self-financing. More information about the XenoServer project can be found here.

The XenoServer Open Platform

The Xenoserver Open Platform is creating a network of XenoServers, running the Xen virtual machine monitor, which allow registered clients to deploy their - potentially unsafe - computation in exchange for (virtual, at this stage) money. Our paper Controlling the Xenoserver Open Platform at OPENARCH 2003 provides an overview of the platform. Global Public Computing describes the vision and architecture in more detail.

Our research focuses on five key problems which arise when building a global network of Xenoservers that are open to the public.

  • Resource discovery: how does a client of the platform discover which server (or group of servers) to use? How do they describe their requirements to the platform? (See our paper at HPDC 2003 for initial work)

  • Resource management: how can control be exercised, in terms of which clients are able to use which resources and to what degree? How can the policies of the federated entities (e.g. server owners, local network administrators, ISPs, infrastructural authorities, and law enforcement agencies) that may wish to influence such decisions be interleaved? (Papers currently under submission)

  • Service deployment: how can clients be allowed to fully customize the execution environments in which their services are deployed? How can users configure and build their environments once, and then efficiently launch them on large numbers of machines around the world? Our paper "Global-scale service deployment in the XenoServer platform" in the First Workshop on Real, Large Distributed Systems (WORLDS '04) summarizes our work.

  • Reputation management and auditing: in a global deployment, how do clients identify rogue servers and servers identify rogue clients? How do server operators respond if a rogue client uses their machines for nefarious purposes? (See our position papers at the First International Conference on Trust Management, Second IEEE International Workshop on Trust and Privacy in Digital Business), and Second International Conference on Trust Management

  • Distributed storage management: how do clients access the data that they require? How do they locate suitable operating system images and configurations to boot over Xen? (See our paper at the 2002 International Workshop on Peer-to-Peer Computing for our initial work)