Recently, I was asked to address a group of IT and security professionals from the Department of Defense (DoD) and other agencies in an online presentation. When I was asked what I wanted to present on, I was overwhelmed by all the possibilities; as a self confessed geek at heart my mind was racing with thoughts of VXLAN technology use cases. Obviously, the big “C” came up – that’s right – the Cloud. I’ve struggled with the concept of the Cloud over the last few years. What is it? How does it benefit a DoD/government customer? And seriously, does it save the customer money?
Based on all the questions surrounding the Cloud, I opted to present on this topic, but chose to take a non-traditional approach. I focused on breaking down the essentials of what the Cloud really is and how to leverage it for organizational and monitoring/metric purposes.
I started developing my presentation from one question: Why has there been so little adoption of Cloud technologies in federal government agencies?
An initial response to this question is that there’s been relatively little adoption of Cloud technologies in the government agencies because there are serious security and compliance issues with maintaining data in the public cloud. I believe the reality has much more to do with market competition occurring between a complex systems model and volume operations model. Looking at current data centers, placing a Cloud or additional abstraction model on top creates a competing scenario. The complex systems model attempts to address highly complex problems with highly complex solutions. The volume operations model attempts to address complex data centers through simple, repeatable solutions. Moving to a cloud consumption model removes the complex system model and begins to encroach upon existing volume operations. In the current IT market, a majority of vendors fall into either category or leverage both within respective market segments.
Ok, so what is the Cloud?
The Cloud concept itself is not overly complex; it removes the underlying infrastructure complexity, viewing it as a commodity and migrates towards allowing all workloads to be treated as a commodity. Per the National Institute of Science and Technology (NIST), Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources, for example, networks, servers, storage, applications and services that can be rapidly provisioned and released with minimal management effort. Hint, hint, if you have a hypervisor technology deployed, it’s already doing 95 percent of the NIST definition. Poof, the magic is gone.
From a historical sales and marketing perspective, the federal agency customer has been told to move their workloads – everything from hosted email to data center operations – to Amazon Web Services, Google, Microsoft or any number of other companies who offer this “new” technology. Let’s go back to the NIST definition, if you’re using something like a VMware hypervisor, what would migrating your workload do? The short answer is nothing really and the same issues that the environments face today will still exist.
New technologies go through a double bell curve; a short curve, followed by a gap, which is then followed by a much larger bell curve. The first two years of heavy Cloud marketing provided the initial rise in the bell curve. The early adopters deploy the technology and then comes the gap. During the gap, the more cautious agencies, wait to see how the technology performs. Most importantly, during the gap, all the gloss has worn off of the technology and the real discussion and use cases begin to develop.
Over the course of the next few years, Cloud deployments will either be integrated or they will fall victim to a new technology. Institutionally, there’s always a struggle going on between how to adopt new technologies while managing or integrating the previous generation. In the case of Cloud technologies, it doesn’t help that the old model is tied to a very physical world.
Sure, we have hypervisor technologies and virtual machines, but think about what a Cloud Consumption Model is all about. The Cloud is not just virtualizing your infrastructure; the real value comes in paying for what you use – specifically from a licensing model perspective. For example, does everyone in your organization need Excel? Would it be more cost effective if an organization could pay hourly or daily for the use of Microsoft Word? This is exactly what a Cloud consumption model ultimately enables. This seems like fiscal Nirvana, but doesn’t provide any additional benefit if the agency isn’t organized. The downside is that if an agency is not organized to understand the resources they already use and what they need, they may end up robbing Peter to pay Paul.
The first order of business for any Cloud infrastructure is organization. Start with an inventory of what you currently have and organize those resources first. I recommend organizing based on your business unit functions; such as portal front-end/portal back-end, Exchange, SQL, organize them per “sub-tenants.” Organize them in a fashion that makes sense per the respective functions they provide. How do you organize them? Right now, if you have vCenter deployed, rebuild your folder trees per the above.
In the next post, I’ll review steps that an organization should take when considering a cloud model and questions to ask your vendor.