At VMWorld this year the talk was around hybrid and private cloud computing. Private cloud computing has the same basic characteristics as public clouds, such as Amazon Web Services (AWS) and Rackspace, and include resource pooling, multitenancy, auto-provisioning, etc., but they exist within private data centers and not on the open Internet.
Lately, there has been interest in leveraging hybrid clouds, or an architecture that allows you to leverage a combination of private and public clouds, sharing the processing load inside and outside of the firewall. The idea is that you can have your private cake and eat it too, allowing certain processes to take advantage of public cloud services during times when the on-premise systems are saturated. Actually, it’s a very clever architecture in theory, and it’s becoming so in practice as well.
Often overlooked is the need for integration as part of these hybrid clouds, or the ability to join the private with the public, sharing data as required to support the business processes. As an old integration guy these jumped on my radar screen rather quickly, but only a few others have married the concepts of hybrid cloud and integration.
The best thinking on this is from my old buddy, Joe McKendrick, who clearly sees the issue around the integration requirements of hybrid clouds, as well as from a great piece by Lori MacVittie. Clearly, Joe, Lori, and myself are kindred spirits on this issue. Lori puts it best: “Simply provisioning the resources in a public environment isn’t enough,” she says. “It must be tied back into the infrastructure and the delivery process. It must be joined to the existing resources so that it appears a seamless extension of the corporate compute resource pool. This process requires integration into existing infrastructure architecture.”
Data integration, for instance, does not occur automatically. You have to consider common sets of semantics that exist between private and public clouds, and then the mechanisms leveraged to synchronize the data. Best practices in this area are still emerging, but the good news is that well-matured data integration technology currently exists. The trick is to figure out the new approaches for your particular problem domain.
The core concept is that you’re extending your private cloud infrastructure out to public cloud providers, and you need to create a solution where the links between the private and public clouds appear to be seamless. While many different approaches are emerging, typically you’re talking directly to a database on the private side and to a database API on the public side. The integration technology sitting in the middle is able to account for semantic differences, as well as work around the different protocols and interface standards.
When selecting a data integration solution, such as those provided by Pervasive, you need to consider a few other issues beyond the data integration mechanisms, such as latency and out-of-the-box support for the clouds you’re leveraging.
Latency refers to the integration technology’s ability to transfer data in a timely manner. This is critical, considering that your cloud solution on both the private and public sides of things should work and play well together without waiting for one side or the other to catch-up.
Out of the box support means that your integration solution supports the public cloud provider you’re looking to leverage, such as Saleforce.com or Amazon Web Services (AWS). The support should be both stated and provable, including out-of-the-box access to all major cloud APIs, and not require a tremendous amount of customization.
The fact of the matter is that integration becomes even more important as cloud computing becomes a larger part of the enterprise and government agency. The core idea, and the value, is pooling resources to provide much more cost effective and green computing platforms. While distribution, such as leveraging a hybrid cloud computing solution, does have great value, enterprise architects and other IT leaders need to be diligent around the fundamental integration requirements.
The action you need to take from all this is to include integration as a part of your client computing strategy that takes its rightful place in the path to cloud computing.