Embedded Java Virtual Machine Technology

As is known to all that the embedded JAVA technology has been widely applied to cell phones, PDA and other consumer devices nowadays in China. However, the copyright of KVM (Kilo Java virtual machine) — the core of this technology — is dominantly controlled by foreign companies, whereas relevant commercial products with domestic copyright are scarce. In addition to this, KVM implement technology still remains to be further developed and studied. Due to the above-mentioned factors. this paper is devoted to an in-depth research on the key technology of KVM so as to pave the way for the development of high performance KVM with our own copyright.First, based on the analysis of the framework of J2ME, the framework's configuration layer and profile layer are introduced. Second, this paper centers on the study and analysis of KVM as the core of CLDC's framework in J2ME The characteristics of KVM and the consumer devices with KVM are expatiated as well as the inner architecture of KVM. What's more, the four core components of KVM are illuminated, namely, the structure of inside Java virtual machine, multithreading and synchronization, and details of the implementation of the interpreter and garbage collection.As for garbage collector, one of the core modules of KVM, this paper gives a new algorithm of GC based on the present algorithm that the KVM model is using after in-depth study, analysis and comparison of the current primary algorithm of GC. Based on current algorithm in KVM model, incremental collection is introduced into the improved algorithm, and this improved scheme could avoid long "pause" due to real full collection caused by former algorithm to large extent, it makes VM more suitable for application requirement of process ability-limited consumer devices: a new solution of object management is presented, it makes object management more rational and more suitable for embedded application. At the end of paper, a svstem prototype of improved algorithm is presented. After analysis the result of test. validity and feasibility of improved algorithm are proved.