Micro Virtual Machine: Abstraction, contained

This proposal will address a systemic source of inefficiency in widely used software which leads to many programs running as much as ten times slower and using considerably more energy than necessary, shortening battery life on mobile phones and increasing costs for large server farms. This inefficiency is endemic because it is due to the underlying languages rather than the particular software. We will address this problem by developing a high efficiency substrate on which languages may be built, which we call a micro virtual machine. Our insight is that by providing fundamental services as a common substrate we address key inefficiencies at their source and allow language developers to focus resources on languagespecific concerns.