Joshua LeVasseur

Pre-virtualization: Soft layering for virtual machines

Despite its current popularity, para-virtualization has an enormous cost. Its deviation from the platform architecture abandons many of the benefits of traditional virtualization: stable and well-defined platform interfaces, hypervisor neutrality, operating system neutrality, and upgrade neutrality - in sum, modularity. Additionally, para-virtualization has a significant engineering cost. These limitations are accepted as inevitable for significantly better performance, and for the ability to provide virtualization-like behavior on non-virtualizable hardware such as times86. Virtualization and its modularity solve many systems problems, and when combined with the performance of para-virtualization become even more compelling. We show how to achieve both together. We still modify the guest operating system, but according to a set of design principles that avoids lock-in, which we call soft layering. Additionally, our approach is highly automated and thus reduces the implementation and maintenance burden of paravirtualization, which is especially useful for enabling obsoleted operating systems. We demonstrate soft layering on times86 and itanium: we can load a single Linux binary on a variety of hypervisors (and thus substitute virtual machine environments and their enhancements), while achieving essentially the same performance as para-virtualization with less effort.

Case study: visualizing Internet resources

The goal is to improve the ability of people from all walks of life and interests to access, search, and use the information distributed in Internet resources. The process of interacting with information resources starts with browsing, continues with digesting and assimilating pieces of information, terminates with generation of new information, and begins anew with analysis of pre-existing and new information. Our approach is user-centric-taking users needs into account by allowing them to interact with the information contained in large arrays of documents. The visualization process is an integral part of the overall process. We have covered three related categories in this methodology. The first one is browsing through the World-Wide Web (WWW) hyperspace without becoming lost, based on a visual representation of the hyperspace hierarchical structure (hyperspace view). The second category is overcoming the rigidity of the WWW by allowing the user to construct interactively and visually a personal hyperspace of information, linking the documents according to the application or problem domain, or to the users own perception, experience, culture, or way of thinking. The third category includes discovery and analysis of new information and relationships in retrieved documents by aggregating relevant information and representing it visually.

A sledgehammer approach to reuse of legacy device drivers

Device drivers account for the majority of an operating systems code base, and reuse of the existing driver infrastructure is a pragmatic requirement of any new OS project. New operating systems should benefit from the existing device driver code base without demanding legacy support from the kernel.Instead of trying to directly integrate existing device drivers we propose a more radical approach. We run the unmodified device driver, with its complete original OS, isolated in a virtual machine. Our flexible approach, requiring only minimal support infrastructure, allows us to run any existing device driver, independently of the OS or driver vendor.

Pre-virtualization: uniting two worlds

Virtual machines are used in an increasingly varied set of application scenarios that favor different trade-offs. The virtual machine (VM) is an attractive solution, since it enables the use of the same operating systems across the scenarios, while permitting substitution of different hypervisors appropriate for the trade-offs. One of these scenarios is server consolidation, where a number of machines are replaced by VMs running on a single physical machine, increasing resource utilization. Another attractive scenario is the use of a VM to add features to an OS that contradict the design of the OS, such as enabling secure computing platforms with strictly controlled information flow. These two scenarios have dramatically different performance versus security trade offs, easily addressed by using different hypervisors.

Dealing with the Data Deluge. Visualization and Data Management

This paper describes solutions to some difficulties with handling, querying, browsing, accessing, and retrieving data and information from large distributed repositories of complex data and information. In one prototype, we integrated the capabilities of a geographic information system, statistical analysis, visualization, graphic user interface, a storage system, and a data base management system. In another system, relevant information in n dimensions could be aggregated and visualized enabling a more rapid understanding of the informational content when the representation is adapted to the particular problem or application. In an enhancement to NCSA Mosaic, the user can modify interactively the otherwise rigid structure of linking pieces of information (hyperlinks) and view the information space and its relationships visually.