Cloud to Cloud Data-migration Performance


Cloud to Cloud Data-migration Performance

  • 19 Sep 2014
  • cloud, cloud computing, Data migration

In the past few years, different types of clouds have emerged. Organizations and companies are migrating data from their local servers to clouds. Due to the high demand of clouds, today there are different kinds of clouds that emerged from different providers whether it is commercial cloud or open source cloud. These clouds do not exchange information and are incapable of migrating data between them which makes it one of the major obstacles in cloud computing systems. Hence, such obstacle results vendor lock-in and lack of interoperability.

The objective of this Tool is to measure the performance of cloud-to-cloud data migration and data migration between two accounts of same cloud.

In this work, an experimental study was used to show the performance of data migration between clouds. For each test case, an application was developed to carry out the experiment. The results have been achieved by testing the performance of data migration from:

  • AWS to AWS
  • AWS to Azure
  • Azure to Azure
  • Eucalyptus to Eucalyptus

After we carried out the experiment the result showed that it takes roughly the same amount of time while transferring 1 GB data from AWS to AWS, AWS to Azure and Eucalyptus to Eucalyptus. There rest test cases show a big different amount of data migration time.

Finally, after the results we conclude that clouds with the same API have a close data migration time than that of clouds with different API. In addition, Walrus data storage used for Eucalyptus and S3 which is used for Amazon have the same architecture, hence the similarity and closeness of data migration time that occurred between these three cases mentioned above is the result of the similarity of their architecture.