Cloud hosting is not (really) cloud computing
In a recent customer meeting, several IT staff members were explaining some serious performance problems they were having with a newly deployed IT system. It began with a general discussion among us about the benefits of scalability and agility in cloud computing. Then they dropped a bombshell: “Oh, our application is in the cloud!”
Given our discussion about scalability in the cloud — juxtaposed with a newly deployed IT system that was not scaling even under a normal load — this statement seemed out of place. And here we have the crux of the matter: Why host a brand new IT system in the cloud and not take advantage of a core cloud benefit like scalability?
In the federal government there are really only three possible answers to that question:
First, an agency’s IT organization may be just “checking a box” to comply with the administration’s “cloud-first” policy. Cloud-first, which called on agencies give a priority to cloud-based applications and services, was announced by Jeff Zients in 2010, when he was deputy director of management at the Office of Management and Budget. This policy was later reaffirmed by federal CIO Vivek Kundra’s 25-point plan to reform federal IT and the Federal Cloud Computing Strategy. These are both good policies that may or may not be implemented well depending on an agency’s interpretation of the spirit versus the letter of the policy.
Second, the agency’s IT organization may be leveraging the cloud’s ease of provisioning and cheaper hardware costs. While this is a valid reason to host in the cloud, it is a marginal benefit at best when compared to the vast improvements in application effectiveness, reliability and scalability that are available for cloud-native applications.
Also, anyone who has estimated the cost of developing an IT system knows that labor costs always dwarf the hardware costs. The pecking order is labor costs, software costs and then hardware costs — each differing by multiple orders of magnitude. So customers will pay much more in lost productivity and additional re-engineering labor costs than they would have if they built it right the first time.
Third, the agency’s IT organization may not yet be educated on the distinction between running applications in the cloud and building applications for the cloud.
The term for applications built specifically to run in the cloud is “cloud-native.” As I discuss in my new book, cloud-native applications leverage a set of cloud architectural patterns to deliver scalability, reliability and elasticity. Some of the patterns leveraged are the queue-centric workflow pattern, reliable messaging pattern, scheduler pattern, a foreman-worker pattern and a failure/recovery pattern. If you use a service-oriented architecture, you may have already worked with some of these patterns.
Simply put, a cloud-native application must take advantage of message queues, asynchronous communication and the ability to automatically scale across multiple servers (called horizontal scaling).
The education necessary to create cloud-native applications goes beyond federal IT managers to the IT staff and developers (including the contractor’s developers). Unfortunately, “elasticity of computing resources” is just not in the current vernacular of most software developers.
Thus, IT managers must check if their development staff, including contractor development staff, is up to speed on developing for cloud computing environments (and especially cloud architectural patterns). It is important to note that building these cloud-native applications is applicable to both platform-as-a-service and infrastructure-as-a-service environments.
After reading this, I am confident that any new application that you put “in the cloud” will actually make use “of the cloud” to benefit your end-users with scalability, elasticity and reliability.
Michael C. Daconta ([email protected] or @mdaconta) is the Vice President of Advanced Technology at InCadence Strategic Solutions and the former Metadata Program Manager for the Homeland Security Department. His new book is entitled, The Great Cloud Migration: Your Roadmap to Cloud Computing, Big Data and Linked Data.
Posted by Michael C. Daconta on Feb 19, 2014 at 1:33 PM