There have been dozens, if not hundreds, of articles or blogs about how IT should deal with the organization’s demand for the types of service offerings available in the public cloud. Unfortunately, as enticing as many of these offerings are to the business, they are equally problematic for IT. Regardless of the difficulty, this does not change the fact that the purpose of IT is to support the needs of the business in as effective and efficient a way as possible. So, if the business wants these types of services, it is core to the job of IT to figure out how to provide them rather than trying to block them.
Cloud Offerings vs. Traditional IT
I believe we need to first consider some of the key reasons cloud offerings are so attractive to the business. They are:
- Very fast to provision
- Relatively simple to understand
- Have low upfront costs
- Use subscription based-fees which can be easily terminated
- Fees are often tied to usage
On the other hand, to get these benefits, you do generally have to stay within a well-defined set of choices with fewer options. We might think this loss of flexibility would be too great a cost and certainly not having enough choices can be a problem. Too much choice is also a problem though, so the smaller set of options actually factors into what makes cloud offerings easier to understand.
The next step is to contrast the experience of a cloud offering with the typical experience the business has when working with an internal IT organization. As much as we hate to admit it, IT tends to offer our business users none of the things which attract them to the cloud:
- Provisioning cycles tend to be long
- It is not always clear exactly what is being provided
- Costs are generally up front and unrecoverable
- If there are ongoing costs they are frequently obscurely defined
This contrast is not said to denigrate how traditional IT meets the needs of the business. My purpose is to highlight the business view and emphasize how those needs might be satisfied differently while still covering many of the gaps not immediately obvious to those looking at a cloud offering.
One of the most extreme examples of the differences is comparing the traditional IT server provisioning process to obtaining a server instance via an IaaS offering in the public cloud. Following the typical internal flow (presuming capital funds have already been budgeted for and approved), the requirements must be documented and a proposed configuration reviewed or developed. Once a configuration is agreed upon, equipment must be procured, received, and deployed. All together, these tasks can take weeks or even months from when the request is initially made until the server is actually available for use. Imagine the frustration of waiting for this process to complete when the requester knows they could have a new server available the same day via an IaaS offering. From the IT perspective, the two end results are wildly different, but to the end user both paths got them the capacity they needed. Differences which are significant to one may not be to the other.
The Operation was a Success, but the Patient Died
These are not new ideas. Phrases such as “service-oriented IT” or “servicentric IT” have become common place the last few years and are largely driven by a demand from the business to have the IT experience feel more like what they can get elsewhere. Unfortunately, when IT creates a service catalog in an attempt to respond to this, it generally ends up being nothing more than a somewhat streamlined method of requesting the same types of services already being offered. Some added efficiencies might cut the time to provision a server from five months down to three, but not to the same day kind of timeframe available via IaaS. The nature of the services being offered has not really changed, so even if the effort is successful it fails to address the actual shift in what the business needs and wants.
To become truly “service-oriented,” IT needs to make a fundamental shift in how it satisfies the needs of the business. IT must become “cloud-like” in what it offers. By doing so, IT can offer the type of experience the business demands while avoiding the myriad problems associated with having them go to the public cloud directly. But, what does it mean to be “cloud-like”? This type of offering must give the business those key factors which attracted them to the cloud in the first place: they must be very fast to provision, relatively simple to understand, have low upfront costs, and ongoing subscription based fees which can be easily terminated and are often tied to usage.
This is not just an evolutionary shift which can be accomplished with a few tweaks to how IT and the business it serves interact. It is a revolutionary change and requires that the entire relationship be completely redefined. In order to provide “cloud-like” offerings, IT must become like a cloud provider and the nature of the relationship with the business must become similar to what they would experience from a cloud provider. In being “cloud provider-like,” IT will have to act and function as a separate entity from the business in many ways, but at the end of the day they are still part of the same team. The primary goal of IT must always remain the success of the business, not profitability. It is this commonality of goal which allows IT to provide “cloud-like” service offerings which have been enriched with added vigilance on behalf of the business.
Four things need to happen in order for this change to work:
- A true catalog of service offerings must be developed
- IT must have the infrastructure to fulfill them
- Finance needs to buy into and support the new paradigm
- Most important is that the business needs to get excited
None of these changes are easy tasks and cannot happen in isolation. I have seen companies thrash for a year just trying to get a start on what this type of service catalog would look like. Developing the existing infrastructure is no less daunting. Worse, they often generate a “chicken and the egg” type situation where the two efforts deadlock. Development of the service offerings cannot move forward without knowing what IT can deliver on and IT cannot determine the infrastructure it needs to deliver without knowing the service offerings. Both of these efforts also tend to be limited by what finance will agree to make possible and nobody wants to get the business excited about something which may not happen.
So what can we do? Simple. Cheat. There are a number of companies out there who have already invested the time and money to figure out the service offerings that the business needs and how they should be priced. Even better, they have made this information publicly available. Who are these companies? The existing cloud providers! Go to Amazon, Google, and the others and look at what they offer. It may not fit exactly with what you have been thinking, but it is going to be close and close is good enough. In fact, given how much they have depending on it, the service offerings that a major cloud provider has come up with are probably better than good and by leveraging them, the deadlock is broken.
Unfortunately, the details on how cloud providers fulfill their offerings and determine pricing for them, etc… are not public. It is one of their deepest darkest secrets. But, we do not need to know how they do it. Given the service to be provided and the terms associated with it, we can work on figuring out the infrastructure that we will need to fulfill it. We also know the price the cloud provider has said they can provide it for profitably and we should be able to get close. If not, we can always assess the risks of using the cloud provider itself to provide infrastructure. There are risks to doing so, but as long as IT remains the central point of management, we can remain aware and conscious of those risks.
A crucial part of the finance team being willing to consider changing its rules and policies is having something believable and concrete to show them. By pointing them at an Amazon or Google and showing them the model you want to support, the finance department has something to work with. This is not going to be enough though. You need the excitement from the business to generate the business case and ROI from making the shift.
Create the Vision
So, here is where things get scary and just a bit crazy. In order for any of this to make sense, we need to get the business excited and become passionate about the idea ourselves. IT cannot transform into a “cloud provider-like” model alone. After all, the whole reason for doing this is to better serve the needs of the business. So, we start the campaign by painting the picture for our IT executives. When they buy into the idea, it is time to start selling the concept outside of IT, with all of the “mission statements” and “project charters” that IT folks tend to hate. Without “selling the vision,” all four of the essential efforts will not make progress. The business users need to rally and be in step to make this project a success.
Here are my thoughts on how to get the business excited about this transition. Pick something painful, like server provisioning, and build your effort around that change and make it a high level goal for the year. While the exact language for the goal is best left to the marketing types, the essence of my suggestion is to set a goal such as “On January 1st 2014, IT is providing internally the same type of service offerings available from a cloud provider.” While the effort to get there will be extensive throughout 2013, a broad-based understanding of the vision and the value to the organization makes achievement possible.