Software aging prediction – a new approach

Shruthi Parashivamurthy, Nagaraj Girish Cholli


To meet the users’ requirements which are very diverse in recent days, computing infrastructure has become complex. An example of one such infrastructure is a cloud-based system. These systems suffer from resource exhaustion in the long run which leads to performance degradation. This phenomenon is called software aging. There is a need to predict software aging to carry out pre-emptive rejuvenation that enhances service availability. Software rejuvenation is the technique that refreshes the system and brings it back to a healthy state. Hence, software aging should be predicted in advance to trigger the rejuvenation process to improve service availability. In this work, the k-nearest neighbor (k-NN) algorithm-based new approach has been used to identify the virtual machine's status, and a prediction of resource exhaustion time has been made. The proposed prediction model uses static thresholding and adaptive thresholding methods. The performance of the algorithms is compared, and it is found that for classification, the k-NN performs comparatively better, i.e., k-NN showed an accuracy of 97.6. In contrast, its counterparts performed with an accuracy of 96.0 (naïve Bayes) and 92.8 (decision tree). The comparison of the proposed work with previous similar works has also been discussed.


algorithm; machine learning; prediction; rejuvenation; software aging;

Full Text:



Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

International Journal of Electrical and Computer Engineering (IJECE)
p-ISSN 2088-8708, e-ISSN 2722-2578

This journal is published by the Institute of Advanced Engineering and Science (IAES) in collaboration with Intelektual Pustaka Media Utama (IPMU).