We propose a mathematical model for scalable and accurate prediction of reliability of atomic web services based on K-means clustering algorithm. The model we propose, CLUS (pdf, pptx), estimates the service reliability for an ongoing request by considering its similarity to prior requests according to user-, service- and environment-specific parameters.
In order to evaluate our model, we conducted experiments on services deployed in different regions of the Amazon EC2 cloud.
We provide:
(1) Measured reliability data that was collected from distributed agents on the Amazon cloud.
(2) ZIP file containing CLUS evaluation application along with all the necessary input files, implemented in Microsoft Visual Studio 2010. in C#.
Note:
1.) Folder input contains file data.txt containing measured reliability data,
2.) Folder densities contains the available data distributions for each density considered in the evaluation,
3.) Folder tables contains reliability prediction for the entire dataset for each considered approach and
4.) Folder output contains evaluation results files.
Data Usage and Acknowledgement
The data provided was collected in series of experiments conducted on the Amazon EC2 cloud. If you are using this data in a paper, please send an e-mail with the paper reference to the authors and we will add it to this page.
If you use these data in your work, please acknowledge the authors.
@inproceedings{Silic:2013:PAW:2491411.2491424,
author = {Silic, Marin and Delac, Goran and Srbljic, Sinisa},
title = {Prediction of Atomic Web Services Reliability Based on K-means Clustering},
booktitle = {Proceedings of the 2013 9th Joint Meeting on Foundations of Software Engineering},
series = {ESEC/FSE 2013},
year = {2013},
isbn = {978-1-4503-2237-9},
location = {Saint Petersburg, Russia},
pages = {70–80},
numpages = {11},
url = {http://doi.acm.org/http://dx.doi.org/10.1145/2491411.2491424},
doi = {http://dx.doi.org/10.1145/2491411.2491424},
acmid = {2491424},
publisher = {ACM},
address = {New York, NY, USA},
keywords = {Cloud computing, K-means clustering, Prediction model, Reliability, Web services},
}
