Abstract
In the future, many more people will use web service in Ubiquitous environments. SMD-cluster, which is one of previous research, is a proposed scalable cluster system to meet the clients' requirement. SMD-cluster has a highly scalable architecture with multiple front-end nodes. However, SMD-cluster cannot provide continuous web service to clients, even though all real servers are alive when a front-end node fails. It causes considerable damage to the system, which provides mission-critical services such as E-business, Internet banking, and home shopping. To overcome this problem, we propose an efficient service recovery method in SMD-cluster. Service recovery information is maintained at each real server to minimize the overhead on front-end nodes. Waste of network usage is minimized by reducing communication messages among nodes in a cluster. We propose fast service recovery using a contentaware routing mechanism, which considers the specific content type of a request, to give priority to mission critical services.
| Original language | English |
|---|---|
| Pages (from-to) | 297-306 |
| Number of pages | 10 |
| Journal | Lecture Notes in Computer Science |
| Volume | 3421 |
| Issue number | II |
| DOIs | |
| Publication status | Published - 2005 |
| Event | Networking - ICN 2005 - Reunion Island, France Duration: 2005 Apr 17 → 2005 Apr 21 |
ASJC Scopus subject areas
- Theoretical Computer Science
- General Computer Science