Who Pays for Parallel Development Pizza Parties in the Cloud?
As we move development and test activities into cloud-based labs, we have noticed an age-old challenge arise to some companies’ adoption efforts. While it is tempting to focus on the value or ROI we can achieve from cloud dev/test environments, we should never forget to ask, “Exactly who pays for all this?”
Is accounting getting in the way of agile development in the cloud?
A lot of companies treat their first forays into the cloud like a pizza party. Since the monthly costs are usually far cheaper than the up-front cost of building conventional labs, the budget owner of the cloud initiative might take that cost savings as a cue to offer cloud labs up as a “flat rate pizza party” to all dev and test units in IT.
If we tell everyone to “come and get it,” and start using our pool of cloud lab resources, they’ll be more likely to adopt them faster, and start realizing the associated benefits of on-demand access… right?
In reality, some teams will show up and come to expect the pizza party every week. Others will only show up once a month, but eat everything. And some guys will start staking out a claim on several boxes ahead of time. But enough about pizza. What I’m trying to say is this:
A free-for-all funding model, even if well intentioned, can be the worst enemy of long-term cloud adoption.
Managing usage quotas and chargebacks at a group or departmental level may be the deciding factor in the success of your company’s adoption of cloud-based dev/test environments.
Any critical resource that is ungoverned has the potential to create a “Lord of the Flies” type of struggle for control, even among teams that should be aligned for a common purpose. Software development is a highly variable exercise, and the consumption of environments for dev and test is equally volatile. You can’t blame some people for hoarding.
Peak usage happens for different groups at different times. Development teams running regressions during the preparation of a major release. Performance teams running a major load test. An IT tiger team responding to a failure with an all-hands break fix exercise. And in between these uses, there’s huge valleys of low utilization.
In volatile demand scenarios, cloud labs are ideally suited to meet these needs with elasticity. But that doesn’t mean everyone is paying their fair share for this capacity. If we gauge the performance of different groups on a profit-and-loss basis, the groups who use fewer cloud resources may start to resent accepting costs for teams that use more cloud capacity, if their usage is not metered or tied to any cost chargebacks. In addition, without controls or quotas on usage in place, teams may extend the system sprawl of hoarded but underutilized systems into the cloud rather than returning that capacity to the pool.
Quotas and chargebacks do not mean we must automatically charge teams who need more cloud lab resources a higher rate than teams who don’t. A disaster recovery team may have a large need for cloud lab images, without a corresponding profit, but their costs are offset by risk. Quotas do mean companies should allocate the usage of cloud to teams or departments, meter their usage, and use that information for accountability and planning for future capacity.
I walked a few steps over to the office of our CFO, Frank Colich. He had a brief post-lunch moment of solitude, so I ran the concept by him to see if he had a different model.
“I would look at the capacity in that pool of cloud environments like cash on hand for software development,” Frank says.
So, you have budgeted allocations on the left of the cash on hand (environments on demand), and expected value for outcomes on the right? Well, Frank’s way of thinking made a lot of sense.
Let’s look at funding controls in another way and talk about compliance, or ensuring usage. If I am a CIO or CTO, I am betting “all in” on moving dev/test activities to cloud whenever possible not just for cost savings, but to increase agility and get more features to market. Therefore I want to monitor quotas to ensure that teams are using their cloud resources with increasing regularity.
I also want to make sure all my teams are fully utilizing cloud instead of reverting to conventional methods of building environments. So when the next year’s budget cycle arrives, I can continue to justify and grow my cloud initiative to that CFO. ROI may be king, but “use it or lose it” is also not going away any time soon.
All technologists could stand to learn a new trick or two about slicing up the pie from the guys with the balance sheets.
Feel like taking Skytap out for a test drive before diving into costs and quotas? Sign up for a free trial today and check out all that Skytap has to offer for yourself!