Editor’s note: This is the second post in a series about cloud computing by Greg Arnette, founder and CTO of cloud archiving company Sonian (an OpenView portfolio company). Read his first post for 3 Questions to Decide Whether Cloud Computing Technology is Truly Right for Your Company.
As a veteran of the cloud computing industry, I’ve been thinking a lot lately about the roller coaster ride that some companies go through when they adopt cloud technology for the first time. Recently, I came up with a pretty fitting (if not exact) analogy.
In the beginning, the feeling of jumping aboard the cloud bandwagon is euphoric and exciting. You feel like it’s opening new doors and it’s going to make your life easier. You feel like you’re getting somewhere, moving forward. Then, reality sets in and you realize that cloud adoption is a lot like learning to ride a bike. It isn’t as easy it seems. There are myriad challenges and pitfalls that can derail you, and if you don’t handle them the right way the whole experience can spiral out of control.
The good news? Similar to bicycling, the cloud gets easier and significantly more enjoyable with practice and experience.
For companies that are either considering integrating cloud technology or are in the midst of that process right now, it’s important to understand the key stages of cloud adoption and what to expect from each of them:
1) Initial Euphoria
This stage marks the beginning of cloud adoption for most companies. The cloud is new, different, and expected to solve a key problem. What’s not to love? This stage is often personified by an “eyes wide open,” headlong dive into cloud technology.
Unfortunately, that euphoria typically dissipates when companies realize that the cloud technology they’ve invested in isn’t as easy or reliable as they thought it would be, and that it takes more time (and money) to implement than they initially planned for. A few examples of potential hardship include:
- Estimating total costs in the cloud: Estimating costs and understanding cloud “purchasing” techniques is a common first-timer cloud hardship. The buying process is a different paradigm compared to buying IT services in a non-cloud environment. The positive benefits of “pay as you go” can desensitize folks from paying attention to the incremental consumption throughout a billing cycle (typically monthly), sometimes to the point of sticker shock. In the old world, IT had purchasing patterns that included pre-approvals, purchase orders, and net-30 day invoices that lowered the risk of budget blowouts.
- Investing in initial architecture patters to achieve reliable cloud operations: Software architectures need to be designed for the cloud. This means that you need to assemble the mandatory requirement to assume failure at all levels, build application reliability into software, and not take it for granted that reliability will come from the underlying infrastructure. The cloud itself is very reliable, but only if you assume the individual components are not and instead compensate at the software layer.
- Robust logging and monitoring that indicate when a service is failing: The simple lesson here is that, as a cloud adopter, you don’t want to fail silently. Each major cloud software component needs to “shout loudly” via logging and monitoring to alert on failure. Far too often, I have seen companies implement the monitoring framework last instead of first. That’s a big mistake.
If a business can get past the hardship stage, it often enters enlightenment. At this stage, the company begins to grasp the nuances of cloud technology and work out most of its early kinks.
4) Steady State Accomplishment
At this point, the company has developed a sound cloud cadence, brought its costs under control, and determined the right cloud architecture. From there, the business can begin taking advantage of the cloud’s true potential and build out its cloud services as it grows and matures.
For some companies, the cloud is a fun new toy. It’s trendy and popular, which can cause those businesses to become enamored with its potential, even if the cloud isn’t really something their customers care about. Unfortunately, those teams very rarely get past the hardship stage, largely because of the three pitfalls I discussed above.
As a startup founder, your goal should be to release a minimum viable product as quickly as possible. If the cloud can help you achieve that goal more efficiently, then by all means begin exploring your options. If, on the other hand, the cloud simply distracts you from more important things, then it might be best to wait until the cloud is a better fit for your business model.