Setting the Stage
I believe everyone will agree that cloud technology has been and will continue to be disruptive to businesses and consumers. Creative people are finding new ways each day to deliver new business services faster and with less investment. Gartner and many other thought leaders at the intersection of business and technology both advocate and recognize that “Cloud First” is a dominant strategy right now for many organizations. See this link to peer inside the crystal ball of Cloud First/Cloud Only forecasts.
I consult directly with CIOs and CTOs each week, and they often start meetings by declaring that their organizations are committed to a “Cloud First” first strategy. However, shortly after they leave the room, I witness the struggle that middle management leadership and internal technology personnel go through while trying to grasp exactly what that means and how they are supposed to change what they do. It’s not just the organization’s technical teams, it’s also Value Added Resellers (VARs) and System Integrators (SIs) who struggle to take their traditional offerings of products and professional services to provide solutions that fit a Cloud First strategy.
What’s the Problem?
The challenge is real. Confusing definitions of what a “cloud” actually is doesn’t help. While I won’t go into a lengthy definition of what is cloud and what isn’t, I’ll reference a diagram from Rackspace which provides a great visual of the overlap between virtualization and cloud. While simple, this diagram sheds some light where much Cloud First confusion exists.
In an attempt to demystify Cloud First adoption challenges, let’s visit the perspective of a CIO. The CIO is focused on enabling the business to deliver services to their customers and subscribe to services which help the organization. They socialize with their peers and are coached to learn that the highest value of cloud services follow this order:
- Business Service Automation
- IT Service Automation
- IT Task Automation
To the CIO, the engagement level of cloud value starts with Software as a Service (SaaS), goes to Platform as a Service (PaaS), and will leverage Infrastructure as a Service (IaaS) as a last resort if the others aren’t available. It’s clear and straightforward from a value perspective. This is why SaaS platforms such as Office 365, SalesForce, Concur, etc. have been successful and seen such a high adoption priority.
Demystifying the Confusion
If it’s so clear, then why the confusion? Keep in mind that the people who implement the capabilities businesses use today typically do so within their on-premise data centers. We have cultivated experts in compute, storage, networking, and virtualization that are good at what they do. When being told to engage “Cloud First,” their natural action is to engage the layer closest to their comfort zone and knowledge base. This same approach is what many VARs/SIs have followed when told to assist customers with cloud adoption.
To further complicate the situation of cloud adoption, system administrators/VARs/SIs decide that, when they do engage IaaS, they do so using the same techniques to provision and manage that they used for on-premise because, to many, the cloud has been overly simplified as merely “someone else’s data center.” This leads to excessive cost, no change in service, and even more frustration between the strategy and the implementation. The organizations who have been advised (by VARs/SIs/other) to move their existing compute workload from on premise to a cloud solution simply to comply with the “cloud mandate” are often disappointed in their cloud adoption and are rightly hesitant to make other additional efforts.
This problem becomes compounded in a world where most organizations lack full-time developers to implement their custom business solutions. The organizations which do employ full-time developers typically have them focus on implementing on-premise solutions. Those which engage in cloud and adopt some PaaS solutions typically still engage compute/storage components (IaaS) and not the more value-added Database as a Service, Machine Learning, Internet of Things (IoT) hubs, etc. that provide exceptional value and cost savings.
Now that we understand the confusion, how do we resolve it? I propose we approach Cloud First with the following methodology:
Start from the top of the value chain
The entire organization should be coached to solve business challenges that rely on technology by starting at the top of the cloud value chain. SaaS solutions typically provide the shortest path to market and represent the lowest cost investment. This is especially true of commodity services which don’t provide much competitive differentiation.
Don’t Fear the PaaS
Much of the competitive differentiation today is going to come from the PaaS level of cloud. Organizations which understand how to leverage DocumentDB, Database as a Service, Machine Learning tools, etc. to provide services through Alexi, Siri, Cortana, et al are going to lead innovation in the marketplace with their offerings. Do these organizations have to invest in development resources? Absolutely. However, given so much of the hard work is already done and available as a service, the time to market has never been faster nor the cost been lower.
Intelligently Engage IaaS
Lastly, if your desired service can’t be met with SaaS or doesn’t fit properly with PaaS, then consider IaaS. However, don’t make it work just because it’s cloud. As most of us already know, the challenge with traditional on-premise applications is that they are designed for vertical scaling. Likewise, most applications are designed for peak workloads and the resources are allocated to have good application performance if that peak is reached. This is one of the reasons that moving on-premise workloads to IaaS is so expensive. You continually pay for resources you’re not using which is contrary to the purpose of the cloud.
Instead, choose workloads which are variable in demand and which can leverage autoscaling technology. Yes, you may have to modify your applications to have session states stored outside of cookies and modify service calls to decouple components. However, the ROI on those modifications is quick (typically less than three months), and then you can architect for minimum workload and allow the scaling and orchestration methods of IaaS to show you cloud’s true value.
For more from Michael Chapman, visit his blog.