GLVM is the system-level execution architecture we are developing across participating TVMs. It defines a shared logical model above the individual bytecode engines.