One topic I don’t hear a lot about is how you get large of amounts of data into the cloud and between clouds. Maybe the assumption is that enterprises will ultimately migrate to “Everything as a cloud” (EaaC); presentation, computing, AND data in the cloud. Consequently, it is understandable how all the discussion is around SaaS, custom thin clients with PaaS, or applications on IaaS. However, wide enterprise adoption of the cloud will rely on the flexibility to use different cloud technologies and providers for different portions of the application and infrastructure.
In the long run enterprises will need a mix and match strategy driven by the financial, security, and compliance constraints. For instance, cloud architectures should support having your data on site in your data center and the compute and presentation layer in the cloud. Cloud architectures should also support the ability to have your data and your compute on different clouds. A critical requirement of this flexibility is the ability to move potentially large amounts of data to and from clouds, since your data and compute resources may be in different data centers. Ultimately, cost-effective network connections are needed. However, typically the bandwidth needed in-between components is a constant and can only be optimized so much in code. In the end, the only option is to optimize the traffic in the network between the components. This is where WAN optimization technologies and products come in.
WAN optimization technologies and products will use compression, caching, and de-duplication to reduce the actual bits going across the inter cloud network allowing enterprises to deploy smaller bandwidth networks and save operating costs. Without a serious offering of WAN optimization technologies and products large-scale enterprise cloud adoption will fail.