Migrating Traditional DC Application to the Cloud Sample Clauses

Migrating Traditional DC Application to the Cloud. Many of the traditional DC applications described above can be migrated into the consolidated Data Centre environments, some with more ease than the others. On the one hand, cloud DCs pose difficulties for including workloads with real time response time demands, harsh regulated security requirements, or predictable resiliency contracts. On the other hand, modern consolidated DC environments facilitate elasticity, dynamicity, and cost efficiency thus enabling workloads that are impossible in the traditional DCs. Figure 2-8 illustrates how several common IT workloads fare on gain versus pain measures, based on IBM’s research and experience. In general, the workloads that appear in the upper right quadrant have proven to be the best fit for cloud computing. Figure 2-8: Workload affinity for cloud computing based on IBM experience [15] It can be foreseen that DC technologies will evolve in the coming several years to overcome the challenges (security and privacy, trustable performance SLAs and performance isolation, management flexibility and smart orchestration) and to enhance the benefits so that a vast majority of imaginable computational workloads will be sustainable over the shared and federated utility computing environments Here are, for example, DC requirements for some of the traditional workloads discussed in Section 2.3.1:  Engineering Applications require specialized software, are computationally intensive and often long running. These application’s computational demands are bursty – high demand during computation, no demand in between the runs when engineers evaluate results and refine their models. Network bandwidth is required in and out of the data center -- to transfer the model’s data from the engineer’s workstation and to deliver the results back to the engineer. Network latency also plays a significant role -- if the latency is greater than 40 ms between the engineer and the engineering cloud, the graphics will not display smoothly causing loss of productivity and complaints. Engineering workloads can benefit from elasticity in hardware resources allocation, e.g. compute clusters allocated to a task can grow and shrink, and software resources, e.g. flexible allocation of shared licenses to running instances of specialized application.  Scientific Applications are both compute and data intensive, depend on large volume data warehouses allowing parallel low latency access, and consist of large amount of different components and system. One...