I recently came across an interesting post on GigaOM: Which is less expensive: Amazon or self-hosted? by Charlie Oppenheimer. The point of the article is to walk through an example that supports his claim that “if you have a significant amount of load that is well known and predictable then you may be able to save some money by bringing a portion or all of that inside,” or in other words, self-host. Amazon is MORE expensive. While the majority of his analysis was pretty good, there were a few gaps that - in my opinion - need to be addressed.
- How often is this type of decision ever ONLY about cost? Rarely. There is a lot of strategy work and planning that should go into the Cloud vs. self-host decision. If you were an ISV start-up seeking funding, how many VCs would fund your business plan if it contained a large budget for buying hosting infrastructure hardware? My guess is zero. Deployment strategies can and should vary. Fortunately, there are ways to manage your deployment within a Cloud environment that can give you more flexibility. Leveraging reserve instances instead of running entirely on-demand could dramatically reduce your run-time capacity requirements without sacrificing the ability to keep up with fluctuating demand. Plus, it will ultimately lower the overall cost of deployment, too. A true win-win.
- If you ARE focusing on cost, then you need to consider whether or not it is more effective for your business to work with Capital Expense (CAPEX) or Operating Expense (OPEX). More often, I hear from our partners and customers that OPEX is the way to go because it is much more predictable and stable from a budgeting and forecasting perspective. The finance department will thank you for it.
- Labor costs CANNOT be excluded from the analysis or the conversation. There is just too much variation between the resource demands of a self-hosted environment vs. a Cloud-based environment in terms of time, money, and personnel needed to scope, implement, manage, and maintain the system. A big part of why we built Progress Arcade was to address the need to simplify application deployment in the Cloud for our partners and customers. Excluding labor costs from your analysis is like refusing to set aside money to pay your taxes each year. In either case, your revenue model will be distorted and incomplete. You can’t see the complete picture.
Bottom line: do your homework and make an informed decision based upon YOUR unique set of requirements. Don’t let someone else “calculate” your success.
Thanks and, as always, please feel free to drop me a line and let me know what you think.