Dalvik bytecode interpreter for 16-bit machines

Abstract

The use of High Level Language Virtual Machines in embedded and-or low-resource systems (henceforth referred to simply as VMs) has gained ground in the past decade owing to various factors, among them the portability of VM code and the appeal of being able to abstract the writing of code for such machines away from the often unfamiliar and widely varying architecture of the target devices. Previous work has focused on Java bytecode. The motivation for the use of Dalvik rather than Java bytecode for such applications is first and foremost its tendency towards smaller size as well as it being engineered with lower resource environments with finite power sources in mind. The principal challenge involved in developing such an artefact is naturally the lower resources available. SRAM memory in particular is very scarce (in the single digits of kilobytes) and thus great care must be taken to minimize memory overhead wherever possible. The continuing development of this project can be followed at the repository specified below.