发明申请
US20100325500A1 Detecting and Recovering from Silent Data Errors in Application Cloning Systems 有权
检测和恢复应用克隆系统中的静默数据错误

Detecting and Recovering from Silent Data Errors in Application Cloning Systems
摘要:
A method, system, and article for resolving a silent error is disclosed. A primary program copy runs on a primary host, and a secondary program copy runs on a secondary host. The primary and secondary copies communicate to maintain synchronized execution. A third copy of the data is stored on a storage device as a write operations log and maintained in memory on the primary host while the program is running. The primary copy is synchronized with the secondary copy by computing a first checksum of data on the primary host in response to a read operation local to the primary host, computing a second checksum of data on the secondary host in response to a read operation local to the secondary host, and periodically communicating the first checksum to the secondary host, and resolving any discrepancies between the first and second checksum of data reflecting a silent data error.
信息查询
0/0