Theme: analysis works on the creating of cloud date centers. Plan: Plan: Cloud User’s Expectations

Indeed, if a VDI user in France is traveling to the US, he will most probably suffer from an increasing latency which will affect the quality of its experience, making the service sometimes unusable.

• Indeed, if a VDI user in France is traveling to the US, he will most probably suffer from an increasing latency which will affect the quality of its experience, making the service sometimes unusable.
• Even if a cloud provider has one data center or is part of a colocation center, he needs to ensure network resiliency to its customers by offering them a multi­homed/distributed architecture. For instance, if a network link is cut and the cloud provider doesn’t have another backup link, the customer won’t be able to access its service anymore, leading to a monetary compensation for the disruption period. A cloud provider should be able to prevent such failovers and put multi-homed services at their customers’ disposal.
• To build stringent cloud architectures at minimal deployment costs, cloud providers are virtualizing their infrastructure with a small software overhead. This amount is negligible due to the fact that servers are becoming faster and less resource greedy. We highlight in the following the benefits that a provider acquire by virtualizing its infrastructure:

Hardware Cost: cloud providers can typically save much more money by virtualizing their resources. Instead of using multiple physical servers for

• Hardware Cost: cloud providers can typically save much more money by virtualizing their resources. Instead of using multiple physical servers for
• Energy Consumption: virtualization reduces the electricity consumption by half. Migrating physical servers to virtual machines and consolidating them into few physical servers lower the cooling costs in the data center.
• Disaster Recovery: fast recovery from a natural disaster or hardware fail­ures is one of the biggest advantages of virtualization. In fact, VMs can be migrated from a failing data center to another one instantaneously with almost no downtime at all. Thanks to hardware abstraction, it is no longer needed to keep the same replicated hardware environment in two different locations. Nowadays, most of virtualization platform provide a software that can help recovering from disaster failovers.