{"title":"Anytime Available Single IP Address Cluster","authors":"H. Fujita, Y. Ishikawa","doi":"10.1109/HASE.2011.53","DOIUrl":null,"url":null,"abstract":"In this paper, we present SSPA (Speculative SYN Packet Acceptance), a mechanism designed to make a broadcast-based single IP address cluster highly available. SSPA ensures that at least one node always responds to new TCP connection requests from clients with little delay, especially when a node fails. With SSPA, during a node failure, another node, snooping packets from the client, recognizes that no one is taking care of the request, and then replies to the client. This significantly reduces the delays experienced by clients even during a node crash, while keeping the overhead of node monitoring low. A fault injection test using an embedded computer cluster shows that SSPA allows a cluster to respond to clients within 400ms during a node crash.","PeriodicalId":403140,"journal":{"name":"2011 IEEE 13th International Symposium on High-Assurance Systems Engineering","volume":"27 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2011-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2011 IEEE 13th International Symposium on High-Assurance Systems Engineering","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/HASE.2011.53","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
In this paper, we present SSPA (Speculative SYN Packet Acceptance), a mechanism designed to make a broadcast-based single IP address cluster highly available. SSPA ensures that at least one node always responds to new TCP connection requests from clients with little delay, especially when a node fails. With SSPA, during a node failure, another node, snooping packets from the client, recognizes that no one is taking care of the request, and then replies to the client. This significantly reduces the delays experienced by clients even during a node crash, while keeping the overhead of node monitoring low. A fault injection test using an embedded computer cluster shows that SSPA allows a cluster to respond to clients within 400ms during a node crash.
京公网安备 11010802042870号
京ICP备2023020795号-1
分享
求助内容：
应助结果提醒方式：
扫码关注我们
