I just returned from the Forrester EA Forum ’09 in Miami. I enjoyed several great sessions discussing the state of EA (which is in need of a bailout), agile EA, and Jeanne Ross’s talk about architecting for Agility. But one session had me scratching my head. It was the discussion on cloud computing. If I was new to cloud computing like many people in the audience, I would have left the session with the mindset that cloud computing is too insecure, too risky, and often not cost effective. Now to be fair to the folks at Forrester, they were presenting both the pros and cons of cloud computing. The problem is they spent a ton of time focusing on the cons. It got to the point where I had to raise my hand and defend cloud computing. Here were some of the statements I heard about the cons of cloud computing:
- “Is the cloud SOX compliant? No!”
- “Once you blow away your image, everything stays in memory and hackers can get it”
- “You have no control….”, “it is not secure”, etc.
I had to check the agenda to see if the title of the session was Cloud Computing FUD 101. So I raised my hand and made a case for cloud computing. I said something to the tune of …..
We are not being fair to cloud computing here. If you buy a rack of servers, they don’t come with SOX compliance, security, and are not hacker proof. You still need to architect for those things. This is just like the death of SOA, Everybody sees the next big thing and just starts doing it. You need architecture. Without it, this too will die”
When industry analysts don’t understand an already misunderstood buzzword and vendors are calling everything “cloud services” or building platforms that lock you in to their cloud, you have the early stages of the same disastrous hype cycle that led to SOA’s eventual funeral which I had the pleasure of attending last week. Don’t read this the wrong way. I am absolutely a huge fan of cloud computing and as a CTO for a startup, I am leveraging the cloud to clobber my competition. But let’s face it, how many new technologies, emerging trends, frameworks, or architectural styles does IT have to screw up before we learn our lesson? Cloud computing is the most disruptive technology since the adoption of the Internet and the birth of the personal computer. The platforms and the tools will mature over time and I predict 5 years from now, those not in the cloud will have a hard time competing. So as IT professionals, how can we save our industry from ourselves, and not let cloud computing die a slow agonizing death like its friend SOA?
Those of us who are trying the cloud, for better or worse, must share those experiences on our favorite social networks. Let’s hear the war stories from the trenches, not just from the generals and the media. Let’s educate the analysts, the vendors, the architects, management, our colleagues. When we see FUD, let’s call it out. When we see failures, let’s analyze them and learn from it. When we have successes, let’s share them. Let’s not blow this one. So let’s start with what we learned from SOA.
- Don’t talk to the business about the cloud, talk to the business about business drivers and benefits that the cloud will generate
- Use the cloud to solve business problems, not technology problems
- It still requires architecture and planning. You can outsource your infrastructure, but you still need to design for security, compliance, interoperability, etc.
- Governance is still critical.
- Same people + same process + new technology = same mess
- Don’t forget about organizational impacts
Let’s get this one right!