I said the same thing when we were in the extreme hype & FUD stage of SOA. Now we have simply changed the marketing slides from SOA to Cloud Computing and the same pattern of insanity commences.
Cloud Insanity – To repeat the same behaviors that caused SOA to fail and expect a different result.
For the creators of Cloud Computing FUD, here are some bullets for you to dwell on….
- Nothing is secure unless you architect for it. The cloud forces you to do this because the endpoints are outside of the enterprise
- Don’t hold cloud computing to a higher standard than your data center. Many data centers have bigger issues than the cloud. See previous bullet
- Don’t jump on every single cloud failure and declare the cloud too risky. Yes GMail has an issue on occasion but I can’t count all the times that the corporate Exchange server had an outage.
- The Cloud is just another implementation of an architecture. Use it when it makes sense.
For the creators of hype (primarily the vendors)
- If your stuff wasn’t designed specifically for the cloud it is just the same stuff rebranded as cloud computing…just like all of your “SOA” stuff
- Before you hype up the cloud, show me some real customer success stories. Don’t show me isolated cases where some guy with a credit card got a lot of work done for a little money. Show me how a customer built an end to end solution that connects to partners, suppliers, and combines both on-premise and off-premise services.
- If you can’t show me the case studies from the previous bullet, then don’t hype the cloud as the ultimate savior. Instead, describe it the way your customers use it, an efficient tool for processing adhoc and non-mission critical applications.
For the practitioners….
- If you have an opinion about the cloud that is not backed with your own research, don’t blog about it. Too many people are grabbing either the hype or the FUD and running with it based on assumptions.
- If you haven’t done your own evaluation and tried to solve real business problems with the cloud, then don’t call yourself an expert. If you have, then let’s have a discussion about the pros and cons.
The cloud offers great potential for those people who take a logical approach to solving real business problems. At the end of the day it comes down to business drivers and good old fashion architecture. To leverage the cloud to deliver enterprise solutions (not one offs, or isolated use cases) you really need an organization that is disciplined in architecture and capable of successfully implementing SOA. Since we failed so much delivering SOA, I cringe as the masses jump on the Cloud Train. That is not a knock on the cloud, it is a knock on the way many people are thinking about the cloud. The hypsters talk about the future state where everything runs in the cloud and all your problems are solved. What they don’t tell you is how do you get there. If you attend the conferences they will tell you how Joe Businesperson can run their credit card and have a system up and running in a day or two without needing IT. Yikes. It’s like all of your Access power users who created unvalidated decision support information that killed your PCI and SOX audits are now empowered with a credit card and unlimited computing resources! Let the nightmares begin. Companies who will have success creating enterprise solutions in the cloud will have these characteristics:
- Have an innovative culture
- Don’t buy into FUD or hype. Instead rely on EA or architecture best practices to validate the cloud.
- Focus on business drivers first and foremost and use the cloud where it makes sense for the business
- Are disciplined and believe in governance
- Have a successful SOA implementation
- Have talented personnel and a few really good architects
- Have C-level support
- Don’t have adversity between infrastructure and development silos
Just like SOA, you need to evaluate the impact of change to the organization. Below is a presentation I did about SOA and transformational change. Replace SOA with Cloud Computing, lather, rinse, repeat! It’s all the same.