الفهرس 
ABSTRACTOnly 14 pages are availabe for public view
 |  | from | 94 |  |  |
| | | from | 94 | | |
Abstract
Cloud services based on datacenter networks are becoming very important. Datacenters host computing and storage resources. These resources are served to customers through a network of datacenters. In such a network, different contents and/or services are replicated over multiple datacenters, so that a user request can be served by any datacenter that supports the specified content and/or service, to optimize availability and performance. Protecting such a network from disasters like earthquakes, hurricanes and terrorist attacks is crucial. In this work, we propose the use of multiple routing configurations (MRC) IP fast reroute recovery process for datacenter disaster recovery. A demonstration of how this recovery scheme can be applied is shown. Also, the impact of the MRC recovery process on the post failure load distribution over network links is discussed and how this impact can be minimized. It is proposed to use a manual technique to minimize the impact of MRC and it is compared to an automatic technique called modified MRC. Both advantages and disadvantages of the two techniques are outlined. It is concluded that a newer technique is needed to overcome the disadvantages.The new technique should not manipulate link weights to achieve good load distribution. To achieve this, a new proposal to implement unequal weight load balance inside open shortest path first (OSPF) routing protocol is made. Two methods of unequal weight load balance are implemented, the regular method and the smart method. It is shown that, the implementation manages to add unequal weight paths to the router forwarding table. A new method that combines the use of unequal weight load balance in OSPF with MRC is proposed to solve the problem of MRC negative impact on the post failure link load distribution in the network.The simulation results show the effectiveness of the proposed model. The smart method is on average of 21% better than the regular method in achieving better load distribution of traffic across network links. A very high load is noted on an average of 17% of the links. The new method succeeds in decreasing this high utilization by redistributing traffic over free network links.