Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
SFDCloud: top-k service faults diagnosis in cloud computing
Jia Z., Chen R., Xing X., Xu J., Xie Y. Automated Software Engineering21 (4):461-488,2014.Type:Article
Date Reviewed: Jul 9 2015

Cloud computing is an important service delivery model using the Internet, remote servers, and resource sharing. This model exhibits the key characteristics of agility, cost reduction, and device and location independence, among others. However, despite these benefits, the model has some real challenges in terms of implementing its architecture. One of the important areas of interest in this regard is fault management.

Generally speaking, a fault management system should help detect and isolate the causes of a problem, besides its reporting and correction. The current state of research indicates that an ideal fault diagnosis system should detect the root causes of a problem, isolate them, and correct the malfunctions.

This paper addresses a unique approach in fault diagnosis/management in cloud computing with an automatic diagnosis of the fault’s composition services in the cloud environment. It presents the business process execution language (BPEL) of a service composition, which is converted to a service dependence graph (SDG) model to simplify analysis. It also highlights the method for locating the top-k faulty activities and their root causes. The effectiveness of the method is demonstrated by providing a case study using 2,507 web services and scenarios. In addition, the authors have provided a case study for a travel planning service using BPEL. The results clearly validate the strengths of the proposed approach for fault analysis under a variety of error scenarios. To put the results in proper perspective, the authors provide an overview of related work, service fault diagnosis, and handling techniques like the model-based method (MBD), exception handling, and faults in global service failures; they also compare the method to the proposed SFDCloud method.

Although I this is a well-written paper on fault diagnosis methods in the cloud, this work does not address all the components of an ideal fault management system. The authors have just adopted one of the approaches to address fault handling. Specifically in sections 4 to 6, they demonstrate and validate the SFDCloud method by providing a use case and a large amount of experimental data. Putting this in perspective with a perfect fault diagnosis framework and the work already performed elsewhere, this method seems to have its limitations in finding the exact location of faults, determining the root cause of issues, fault correction, reporting, and corrective actions.

This paper can serve as a strong basis for further research on fault handling. I strongly recommend it to researchers in the cloud community focusing on fault management, diagnosis, and monitoring, to improve service availability, reduce downtime, and restore services quickly.

Reviewer:  Sajjad Khan Review #: CR143594 (1509-0783)
Bookmark and Share
 
Cloud Computing (C.2.4 ... )
 
 
Diagnostics (D.2.5 ... )
 
Would you recommend this review?
yes
no
Other reviews under "Cloud Computing": Date
Cloud security and privacy: an enterprise perspective on risks and compliance
Mather T., Kumaraswamy S., Latif S., O’Reilly Media, Inc., Sebastopol, CA, 2009.  336, Type: Book (9780596802769), Reviews: (1 of 3)
Dec 14 2009
Cloud security and privacy: an enterprise perspective on risks and compliance
Mather T., Kumaraswamy S., Latif S., O’Reilly Media, Inc., Sebastopol, CA, 2009.  336, Type: Book (9780596802769), Reviews: (2 of 3)
Jan 26 2010
Cloud security and privacy: an enterprise perspective on risks and compliance
Mather T., Kumaraswamy S., Latif S., O’Reilly Media, Inc., Sebastopol, CA, 2009.  336, Type: Book (9780596802769), Reviews: (3 of 3)
Mar 18 2010
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy