A Token Based Approach Detecting Downtime in Distributed Application Servers or Network Elements
This paper defines and outlines and proposes a heuristic method to detect unavailability on a practical service created on the Internet. It uses a token that is passed between all the nodes. The algorithm is under trial in a set of servers spanning over multiple administrative domains. This paper confirms that using a simple token passed through the contributing applications servers, is able to detect unavailability without intruding or changing the service that one wants to monitor.
KeywordsApplication availability Web Services
- 1.George, P.: Improving Distributed Application Reliability with End-to-End Datapath Tracing, PhD at Electrical Engineering and Computer Sciences, University of California, Berkeley. Technical Report No. UCB/EECS-2008-68, May 22 (2008) http://www.eecs.berkeley.edu/Pubs/TechRpts/2008/EECS-2008-68.html
- 2.Bouricius, W.G., Carter, W.C., Schneider, P.R.: Reliability Modeling Techniques for Self-Repairing Computer Systems. In: Proceedings 24th National Conference ACM (1969)Google Scholar
- 4.Java virtual machine, http://java.sun.com
- 5.Tomcat, http://tomcat.apache.org/
- 6.TCPDUMP tool, http://en.wikipedia.org/wiki/Tcpdump