| 摘要 |
Technology for sharing data among multiple virtual machines in a cluster of virtual machines is disclosed. Each virtual machine identifies "managed" objects of an instance of an application running at the virtual machine. Operations performed by an instance of one application which affect the state of managed objects are detected and distributed. Technology is included for distributing signaling between threads on different virtual machines. The technique extends existing language semantics, including "synchronized", "wait" and "notify", "thread.join" and network call methods to an entire cluster of virtual machines.
|
