Presence Cloud Supporting Mobile Presence Services In Large-Scale Social Network Services |
Social networking services on the Internet are growingand increasing numbers of people are using these new ways to communicate andshare information. Many users are communicating with both friends from outsidethe service as well as with people they have only been in contact with througha social networking service. At the same time mobile phones are becoming morepowerful and increasingly offer high speed Internet connectivity. Because ofthis people expect these social networking services to be available on their mobiledevice, as well as on their personal computer. Given the capabilities oftoday’s mobile devices, it is possible to extend the existing phonebook withcapabilities to support a variety of social networking services in addition tothe existing communication options. By integrating the contacts gained from thesocial networking service into the mobile phonebook the user can reach thesecontacts easily. Social network applications are becoming increasingly popularon mobile devices. A mobile presence service is an essential component of asocial network application because it maintains each mobile user’s presenceinformation, such as the current status (online/offline), GPS location andnetwork address, and also updates the user’s online friends with the informationcontinually. If presence updates occur frequently, the enormous number ofmessages distributed by presence servers may lead to a scalability problem in alarge-scale mobile presence service. To address the problem, we proposeefficient and scalable server architecture, called Presence Cloud, whichenables mobile presence services to support large-scale social networkapplications. When a mobile user joins a network, Presence Cloud searches forthe presence of his/her friends and notifies them of his/her arrival. PresenceCloud organizes presence servers into a quorum-based server-to-serverarchitecture for efficient presence searching. It also leverages a directedsearch algorithm and a one-hop caching strategy to achieve small constantsearch latency. We analyze the performance of Presence Cloud in terms of thesearch cost and search satisfaction level. The search cost is defined as thetotal number of messages generated by the presence server when a user arrives;and search satisfaction level is defined as the time it takes to search for thearriving user’s friend list. The results of simulations demonstrate thatPresence Cloud achieves performance gains in the search cost withoutcompromising search satisfaction.