PROJECT TITLE :

Deadline-Aware Fast One-to-Many Bulk Transfers over Inter-Datacenter Networks

ABSTRACT:

An ever-increasing number of cloud services are being run on a global scale. In order to increase both the quality and reliability of the service, its data is frequently replicated across a variety of datacenters located in different geographic locations. This type of replication results in a large number of one-to-many bulk data transfers taking place across inter-datacenter networks, going from one datacenter to a large number of receiver datacenters. These data transfers are typically required to be finished within the deadlines that have been set in order to guarantee the services that are provided to the end users. In spite of the exponential rise in the demand for data, relatively little research has been conducted on the topic of guaranteeing deadlines for one-to-many data transfers, which is the focus of this paper. This article suggests a centralized admission control that is combined with a scheduling algorithm that is given the name deAdline-Guaranteed transfEr (AGE) for the purpose of guaranteeing the deadline of data transfers that have been admitted and making effective use of the available network capacity. The primary concept here is to allow for flexible selection of the source datacenter for receiver datacenters and to make it possible for the remaining receivers to obtain a replica either from the original source or from the other receivers that have already obtained a copy of the data. This will allow the most efficient use of resources. AGE increases the number of deadlines that are met by maximizing the number of transfers that take place after jointly allocating the source for receivers as well as the bandwidth and routing paths for each data transfer. According to the results of our simulations, when compared to the current state-of-the-art, AGE not only guarantees the deadline for up to 70 percent more transfers, but it also achieves at least two times higher network throughput and reduces the completion time by up to 80 percent.