In This Chapter
You’ve learned about the current state of IT and the “why” and “what” behind enterprise cloud. You’re an expert on all the reasons enterprise cloud is a great path forward.
Just one problem remains: getting from here to there. That requires a number of activities on your part — changing the way you think about managing IT, ensuring that your staff is ready for the changes, and modifying processes and infrastructure.
So, this chapter starts at the beginning.
Change is hard. Everyone knows that. However, in an industry that often leads change in organizations, there is some irony in how difficult it can be for IT professionals to accept change themselves.
As with so many things, you must adapt or you will be relegated to the dustbin of irrelevancy. Look at how many mainframe operators who failed to stay current didn’t survive the wave of decentralization that defined IT in the 1980s and 1990s.
History teaches that change is inevitable. It’s time for you to adjust your perspective on a number of fronts — beginning right now — so that you stay relevant into the 2020s and beyond.
“If IT isn’t providing it, no one should be doing it.”
Variations of this phrase have been around for a long time. My, how times have changed.
Today’s business units can stand up services with nothing more than a credit card and, frankly, many do so. Services have become vastly more consumable than they were just a few years ago. You can thank the cloud for much of this. Thousands upon thousands of software services are available for your users to quickly and easily stand up and consume.
Further, end‐users have become far more tech‐savvy than they used to be. Luddites no more, many end‐users rival and exceed IT staff technical knowledge in many ways and they are the subject matter experts in their areas.
Users don’t want to be fully reliant on IT anymore, either. They want to do their jobs on their own terms.
This phenomenon has become known as shadow IT because it often happens in the dark corners of the company. End‐user departments build systems that they need because IT either can’t or won’t build them. Or, users perceive IT as being so slow that it makes no sense to engage IT in the first place.
CIOs and other IT leaders see shadow IT as a threat to be abolished, and sometimes they have good reasons:
So, it sounds like you should do everything in your power to stop shadow IT in its tracks, right? Well, not so much.
Shadow IT rises because the organization has a need that is not being met. Whether the need is real or perceived is generally irrelevant. Even if the need is only the perception of a failing, something happened in the organization to cause that perception.
It’s time for CIOs and IT leaders to embrace shadow IT. Find out why the shadow systems were set up, and look for the underlying shortcomings in IT’s services. That may even require IT to extends its portfolio and begin encompassing services that were stood up by end‐users.
In general, IT governance processes should provide support for how shadow services can be brought under at least a semblance of IT management. The governance process should outline how departments stand up such services so that they comply with organizational security and data guidelines.
You can no longer ignore shadow IT, but you can’t run in and put a stop to it, either. Instead, you need to implement constructs to help you discover the true needs of the business and ensure that your environment meets those needs within the confines of organizational policy.
If you’ve worked in IT for any length of time, you’ve probably heard of the 80/20 rule. Also known as the Pareto Principle, the rule states that 80 percent of the IT budget and IT’s efforts go to keeping the lights on, while only 20 percent is dedicated to innovation and propelling the business forward. Okay, if you’re a purist, the original Pareto Principle states that 80 percent of the effect comes from 20 percent of the causes, but the meaning has been extended to IT operations in the way I describe.
As CIOs, other IT leaders, and the executive team look for ways to better address key business problems, the 80 percent of the budget that goes to sunk costs can look enticing. After all, if you can improve efficiency just a little, you can change the ratio to 60/40 or 50/50. Improve efficiency enough, and you can create an IT organization that spends only 20 percent on the basics and 80 percent on value added.
This is where IT needs to head. The 80/20 rule is a remnant of a different time. With business needs changing at a faster pace than ever before, IT needs to reduce the 80 percent figure and focus on revenue generating activities.
By deploying enterprise‐cloud‐enabling hyperconverged infrastructure, IT can begin to shift some of that 80 percent toward other activities. An enterprise cloud infrastructure includes automation capabilities and user self‐service, which helps users reduce their reliance on IT and frees up IT staff to focus on the business. Further, with a revamped economic model that enables just‐in‐time infrastructure and easy scaling, that 80 percent of the IT budget can get even lower.
A common school of thought says that IT departments need to fully embrace bimodal IT, discussed in Chapter 1. Under a bimodal support paradigm, you’d have people supporting legacy environment and a separate group of people supporting modern apps.
The problem is that this kind of support is expensive and inefficient. All you’re doing is patching a symptom rather than addressing a root cause.
Rather than attempt to build structures around different application and infrastructure support models, a more sensible plan is to deploy infrastructure that can support both modes of support.
That’s exactly what you get with an enterprise cloud deployment. You gain an infrastructure model that can support legacy applications as well as modern apps.
Compared to dealing with people, the technology is easy! However, you can’t ignore the need to ensure that your people are prepared for the changes that you need to make to your IT organization and to your infrastructure.
For a long time, businesses have had to hire specialists for each area of the IT infrastructure. As organizations move into the brave new world, IT seems to need an ever‐increasing number of specialists to keep the burgeoning set of resources operational.
Hyperconverged infrastructure driven enterprise cloud systems don’t require specialists to operate. You don’t need to hire expensive storage administrators, virtualization administrators, and systems administrators to maintain your legacy environment. You don’t need to hire cloud experts to deal with your public cloud systems.
With an enterprise cloud foundation, you need IT pros who have a breadth of knowledge, though they don’t require massive depth. These IT generalists are the future of datacenter support. They’ll be in the forefront as organizations seek to simplify a complex morass of technology in the datacenter.
At first, this change can be threatening to existing staff, particularly those who define themselves based on their subject matter expertise. Because every area is still represented in the new paradigm, you can move existing people into new roles that are more general in nature. However, they can also have a more business‐facing component that helps shift IT’s focus from infrastructure to the bottom line.
Beyond people, you also need to rethink how you handle some of your IT processes and your infrastructure.
To find a starting point, consider the current IT replacement cycle. For this scenario, I assume that the organization has a five‐year replacement cycle, a visual depiction of which is shown in Figure 4‐1.
When you buy infrastructure, you probably overbuy, even if you run out of capacity. How can this be possible? In Figure 4‐1, the horizontal line depicts the overall capacity that you’ve purchased. In this context, capacity doesn’t refer only to storage; it also refers to the amount of processing (CPU) power and RAM that you have available.
Most IT departments buy what they think they’ll need for the duration of the replacement cycle. Sometimes the estimate is correct and sometimes it isn’t, but one fact is always true: For some period of time, you won’t even come close to using all the capacity you’ve purchased.
In Figure 4‐1, the diagonal line depicts the actual workload demand for the organization that purchased this infrastructure. The lines intersect in Year 4. The shaded area before Year 4 depicts the “waste” that this organization is suffering from. I refer to this as the zero return on investment zone. More than three years go by before the fictional company grows into what it purchased.
Additionally, the organization did not reach the end of its replacement cycle before running out of capacity. This means the company must make an out‐of‐cycle infrastructure purchase to add capacity.
Figure 4-1: Traditional infrastructure procurement economics is not a viable solution.
With enterprise cloud and hyperconverged infrastructure, you can begin to adopt a just‐in‐time approach to datacenter resources. This method allows you to also adopt cloudlike pay‐as‐you‐go economics. Figure 4‐2 shows what such a scenario might look like. In Year 1, you buy what you need for that year, making sure to keep your purchased capacity just a little ahead of your workload needs.
Under this model, you have no zero ROI zone. You’re effectively using what you’ve purchased. Your upfront economics are far better than they are with traditional infrastructure. In short, you aren’t wasting your capacity.
Notice that your organization didn’t run out of capacity in Year 4. Instead, your company simply added more nodes to its hyperconverged infrastructure‐based enterprise cloud environment. You’ve successfully operationalized the changes you’ve made to the datacenter environment.
Figure 4-2: Hyperconvergence and enterprise cloud can help you to reinvent the IT budget.
With enterprise cloud based on hyperconverged infrastructure, you no longer have to worry about trying to manage resources separately. Instead, resources are aggregated and scaled together in a somewhat linear manner. Hyperconverged infrastructure vendors typically make it possible for end‐users to focus on needed resources as they add new capacity. For example, if you’re running low on storage capacity, your new node can be storage heavy, although it will have CPU and RAM as well.
A disaggregated replacement cycle is far more difficult to manage than one based on aggregation, as is the case with hyperconverged infrastructure. As you seek to scale your enterprise cloud environment, you don’t need to focus on individual resources. You simply worry about the needs of your workloads, and add nodes as necessary.