On December 18, I had an interesting Twitter conversation with Mark Thiele (@mthiele10) about moving to the cloud based on cost. There is a cost perspective to consider as cloud services can be very expensive. When does it make sense to go to the cloud? There are two scenarios to consider when talking about going to the cloud. While we were hampered by the 140-character limit, I think the message is clear.
Moving to the Cloud: A discussion
My initial question is about the costs of moving to the cloud. There is some point in time when moving to the cloud makes sense, but when is it? I see lots of discussions about migrating to the cloud, but they seem to be either for greenfield (brand new organizations) or wholesale movement to the cloud. That type of ‘rip-and-replace’ is very expensive.
Mark’s response was quite thoughtful. It is a tough question, as it differs for almost every organization. A small enterprise/business may be more energetic about adopting cloud, but they may still opt to have their own hardware from a control perspective, or they may find that the cloud application they need just does not exist yet.
Yet even as we look at moving to the cloud, cost is a factor. We have all heard horror stories about bills from Amazon. Utility computing often has high utility rates, so working the best deal for your organization is important. If cost were not an issue, we would all be ‘in the cloud’ today. Cost includes not only utility computing costs but also those related to compliance, security, etc. We must include everything in the analysis.
Then there is the natural migration approach as Mark defines it. A good time to consider the cloud is when making a hardware or software upgrade. Is the cost of the hardware and software upgrade significantly higher than the cost of moving to the cloud? Business growth would dictate new hardware purchases and software licenses, and this growth could open up the cloud as a cost effective option.
I think growth may be the best motivator moving forward, as the cloud is ripe for hybrid approaches where the cloud is used as a utility for more compute power as needed. As you can see from Mark’s response, he agrees that growth may be the best motivator. As we outgrow our current virtualized data centers, the costs associated with renting or buying more space, cooling, and power go up quite a bit. This may be when the cloud becomes feasible. The cloud may also become cost effective when there is a major app change, one that is cloud-capable. However, we still have to deal with maintenance of a business-critical application. That trumps all other concerns. Migration of business-critical apps to the cloud is difficult at best.
All of this implies that the number 1 reason to move to the cloud would be that moving the application is cost-effective. Without the ability to move a business-critical application, cloud usage will be limited to the lowest hanging fruit or perhaps some small amount of development and test, which may not be the way to start in the cloud (a discussion for another time).
This raises the question: Does an enterprise have enough developers to develop cloud-based business critical application so that migration can be made to the cloud? Can they handle the deployment model necessary to support the application in the cloud? Or are the developers in a break/fix development mode instead of looking into future applications? Or perhaps agile development is really in the way? (Another discussion for another time).
One may counter: Why not use the existing applications that are cloud ready and move towards Software-as-a-Service (SaaS) deployment model? This may also be cost effective; however, the application must already exist. If it does not, is it up and coming? Many applications have been moved to the cloud, including Exchange, Oracle, SAP, etc. The list grows daily.
We are on the cusp of massive cloud adoption, but to get there, we need to have business-critical applications in the cloud.
I would like to thank @mthiele10 (EVP Data Center, Cloud enthusiast, Pres & Founder Data Center Pulse. SF Bay Area · http://datacenterpulse.org) for a great conversation.
Conclusion
Many of the early adopters are there, many cloud providers are ready, but when you’re a small enterprise with your own hardware, ditching that investment is a tough nut to swallow from a cost perspective. I would think it is even harder for larger organizations. However, if you are just starting out, the cloud may be the cost-effective way to go, allowing for a large presence with low costs, with the warning that the cloud as it currently exists can get very expensive very fast, depending on the deal you make. The bigger you are, the better the deal.
As businesses grow, moving to the cloud is a real possibility; cost is a factor for everyone, small, medium, or large. Exactly when it becomes cost-effective depends entirely on the organization; hopefully this conversation will help with those decisions.