|
|
 |
| |
|
|
|
A Prediction Recovery Method for Supporting Real-Time Data Services |
|
|
|
PP: 363S-369S |
|
|
|
Author(s) |
|
|
|
Yingyuan Xiao,
Hua Zhang,
Guangquan Xu,
Jingsong Wang,
|
|
|
|
Abstract |
|
|
| Real-time systems, which are often employed to monitor and interact with dynamic environments, are widely applied in time-critical applications, such as autopilot systems, medical patient monitoring, robot navigation, military command and control systems, agile manufacturing, etc. These time-critical applications require real-time systems can provide real-time data services ceaselessly. However, real-time systems cannot completely avoid all kinds of failures, so real-time systems must prepare for possible failures and provide fault tolerance capability. The conventional failure recovery methods cannot guarantee real-time data services available when some data items are damaged by failures. In this paper, we present a novel prediction recovery method through the integration of regression model and grey theory. The prediction recovery method guarantees real-time data services available by means of providing predictive values of damaged data to application activities which have to access these data immediately. Performance test shows that the proposed prediction recovery method can significantly improve the real-time performance. |
|
|
|
|
|
|
|
