Multiscale not multicore: efficient heterogeneous cloud computing

  • Authors:
  • Anil Madhavapeddy;Richard Mortier;Jon Crowcroft;Steven Hand

  • Affiliations:
  • University of Cambridge, Cambridge, United Kingdom;University of Nottingham, Nottingham, United Kingdom;University of Cambridge, Cambridge, United Kingdom;University of Cambridge, Cambridge, United Kingdom

  • Venue:
  • Proceedings of the 2010 ACM-BCS Visions of Computer Science Conference
  • Year:
  • 2010

Quantified Score

Hi-index 0.02

Visualization

Abstract

In this paper, we present a vision of the future of heterogeneous cloud computing. Ours is a cleanslate approach, sweeping away decades of accreted system software. We believe the advent of the latest technology discontinuity---the move to the virtual cloud---makes this a necessary step to take, but one capable of delivering significant benefits in the security, reliability and efficiency of our digital infrastructure at all scales. We motivate this vision by presenting two challenges arising in different fields yet with fundamental commonalities best addressed by a unifying software platform supporting devices ranging from virtual servers in the cloud, through desktops, to mobile smartphones. After drawing out this common ground, we describe our solution and its benefits. We then describe the initial steps we have taken toward our solution, the Mirage framework, as well as ongoing future work.