Page 550 - Cloud computing: From paradigm to operation
P. 550

2                                            Cloud Computing management


                Table I.8 – Data placement policies for data-intensive applications in the inter-cloud environment

             Title                        Data placement policies for data-intensive applications in the inter-cloud
                                          environment
             Description                  Data-intensive applications are widely used in an increasing number of fields,
                                          e.g., high-energy physics, bioinformatics and astronomy. There are several
                                          challenges if the datasets required by this kind of application, including already
                                          existing data, intermediate data and final data, should be placed in the inter-
                                          cloud environment. Therefore, it is essential for federated CSPs to negotiate an
                                          appropriate placement policy for the different datasets. The placement policy
                                          conditions include, but are not limited to, the time cost of data transmission,
                                          storage space, network performance and data dependencies. The following
                                          aspects are described as the illustrations.
                                          –  The transmission of data among different CSPs is inevitable. On one hand,
                                             the data scale is huge and the bandwidth limited; on the other, sometimes,
                                             there are several datasets limited to location in some specific CSP based on
                                             application logic. Consideration of how to reduce the time cost of data
                                             movements is required.
                                          –  There are dependencies among these different datasets placed in different
                                             CSPs. Consideration of how to maintain these dependencies is required in
                                             order to improve calculation efficiency.
             Roles                        CSC, CSP (SaaS), CSP (NaaS, SaaS)
             Figure (optional)

























             Pre-conditions (optional)    The CSPs form an inter-cloud federation pattern.
                                          The datasets need placement in different CSPs.
                                          There are dependencies among these datasets.

             Post-conditions (optional)
             Derived requirements         Dataset placement policies among different CSPs (see clause 8.1)

            I.9     Use case of data regulation across different countries in the inter-cloud environment

            This use case illustrates data regulation across different countries in the inter-cloud. The inter-cloud peering
            pattern used to illustrate the use case is an example only.












            542
   545   546   547   548   549   550   551   552   553   554   555