This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Are Virtual Machine Monitors Microkernels Done Right? Steven Hand, Andrew Warfield, Keir Fraser, Evangelos Kotsovinos, Dan Magenheimer † University of Cambridge Computer Laboratory † HP Labs, Fort Collins, USA 1 Introduction At the last HotOS, Mendel Rosenblum gave an ‘outra- geous’ opinion that the academic obsession with micro- kernels during the past two decades produced many pub- lications but little impact. He argued that virtual machine monitors (VMMs) had had considerably more practical uptake, despite—or perhaps due to—being principally developed by industry. In this paper, we investigate this claim in light of our experiences in developing the Xen  virtual machine monitor. We argue that modern VMMs present a practi- cal platform which allows the development and deploy- ment of innovative systems research: in essence, VMMs are microkernels done right. We first compare and contrast the architectural purity of microkernels with the pragmatic design of VMMs. In Section 3, we discuss several technical characteristics of microkernels that have proven, in our experience, to be incompatible with effective VMM design. Rob Pike has irreverently suggested that “systems soft- ware research is irrelevant”, implying that academic sys- tems research has negligible impact outside the univer- sity. In Section 4, we claim that VMMs provide a plat- form on which innovative systems research ideas can be developed and deployed. We believe that providing a common framework for hosting novel systems will in- crease the penetration and relevance of systems research. 2 Motivation and μ History Microkernels and virtual machine monitors are both well explored areas of operating systems research dat- ing back more than twenty years. Both areas have fo- cused on a refactoring of systems into isolated compo- nents that communicate across well-defined, typically narrow interfaces. Despite considerable structural sim- ilarities, the two research areas are remarkable in their differences: Microkernels received considerable atten- tion from academic researchers through the eighties and nineties, while VMM research has largely been the baili- wick of industrial research. 2.1 Microkernels: Noble Idealism The most prolific academic microkernel ever developed was probably Mach . A major research project at CMU, Mach’s beginnings were in the Rochester Intel- ligent Gateway (RIG)  followed by the Accent ker- nel . The key motivation to all of these systems was that the OS be “communication oriented”; that they have rigid, message-based interfaces between system compo- nents. Many of the abstractions used in Mach and later systems appeared initially in the RIG, including that of the port . However, the communications orientation of these systems originally intended to allow the distribu- tion of system components across a set of dissimilar physical hosts....
View Full Document
- Fall '11
- Operating Systems, virtual machine, Microkernels, Operating Systems Principles