.png){kind=icon}
Return
Pre-allocation based flash crowd mitigation algorithm for large-scale content delivery system
.png){kind=icon}
PRE
AI
DOI：10.1007/s12083-014-0272-4.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
A burst of requests may starve the upload capacity of server and degrade the Quality of Service (QoS) when a popular or time-critical content is released in content delivery system. With an enormous volume of popular content available in the network, as well as the growing size of online users, the previous improvement of download protocol is far from enough to handle the flash crowd in large-scale system. In this paper, a novel flash crowd mitigation algorithm called Peer Assisted Pre-allocation (PAPA) is proposed from a new perspective, in which proactive strategies are provided to handle sudden workload of the system. Through numerical simulation and practical experiment in download platform of Tencent, it is proved that benefited from the collaboration of servers and helpers, PAPA is able to save the peak server bandwidth by over 30 % and thus greatly reduce the operating costs of system.
Keywords:
Content delivery
Flash crowd
Pre-allocation
Fluid model
Journal
IF:
2.6
Papers:
2.2K
Citations:
2.9K
Organization
View more organizations
Cited Papers
Potential bile acid metabolites. 19. The epimeric 3α,6,7β-trihydroxy-and 3α,6,7β,12α-tetrahydroxy-5α-cholanoic acids.png)
Steroids
SteroidsIF0
PREAI
ShareSave View more