Monday, May 09, 2011

I Was Wrong: The Next Big Thing is NOT Cloud Computing

Yes, after 5 years of being a big fan of cloud computing. I admit defeat. I am, as of today, changing my worldview. Cloud Computing is no longer the future. And that's the bad news.

But the good news is that Cloud Computing is now an accepted reality - its the present - the market has tipped. The big server hugging companies of the past are "all in" with cloud, or are building next generation "in memory" cloud systems, or rolling out "cloud boxes" that you can take to go. While the false clouds proliferate, the customer is getting savvier.

More importantly, just 2 years ago, most CIO conversations I had started out with trying to convince them the value of cloud computing - especially the 'real cloud' which has been summarized perfectly by my friends at Heroku as 'no software, no private cloud', 'multi-tenancy' and 'abstraction=value'. Today, CIOs are no longer asking what is cloud computing? They are no longer asking do I need to do cloud computing? Is cloud computing a fad (based on generous free education by their existing leading vendors)?

The Conversation for the CIO has Shifted: How can I best leverage the Cloud?

Here are the top 5 conversations I am seeing in the industry.

1. What applications can or should I run in the Cloud? The answer here depends on the individual customers needs but the shift is quite radical. They want to find the best cloud computing app to meet their needs and then look for reasons why there needs cannot be met by that cloud app. This is a bias - and its based on good reason. When you buy a good SaaS app, you are not just getting functionality and performance that you see today - but most SaaS apps are continuously improving with every new release. You don't have to wait 5-15 years to have a massive upgrade/re-write project just because you want the new improvements in your app - features or architecture. For example, when a SaaS app makes the UI better, you can just turn it on. When a SaaS app adds new APIs or new protocols (REST), you simply start using it. You don't have to buy half a million dollars worth of middleware so you can send and receive XML messages. This aspect of cloud apps while subtle is a game changer - and once a CIO understands this by experiencing it first hand, he's much more willing to buy a cloud app that will likely keep pace with innovations than an on-premise app that will require major re-work to just make simple changes.

2. Which Cloud Platform should I standardize on? This is not a question I was hearing a few years ago. Thanks in part to the proliferation of cloud platform choices ranging from and Heroku to Amazon Web Services to Google App Engine to Azure, customers are looking to make a few strategic bets. CIOs want to validate, verify and then approve the use of a select one or few platforms. The answer often depends on what kind of applications you are looking to build and what level of security and open-ness you want. is offering its customers (and please read my disclaimer) an open, social and mobile platform which is proven and built on technology running the world's leading enterprise cloud database for over 10 years. But no single platform will meet the needs of all your applications - hence an open platform that interoperates is a key consideration.

3. How do I get started? The most important question is the getting started question. Unlike a few years ago, there is a wide array of choices in how to get started. If you are looking for a packaged app, application stores like AppExchange or Google App Marketplace are good places to look. If you are looking to build an application, if its a business application you want to pick a platform. The key to getting the value out of a platform is abstracting the underlying infrastructure and servers - a true Platform as a Service. Infrastructure as a Service is a reasonable choice if what you are building is so unique that you want to take the DIY approach.

4. Who can help me with Cloud adoption? There is a vast array of resources and partners now with dedicated cloud practices (much more so than just a couple of years ago) ranging from Accenture & Deloitte to Appirio & ModelMetrics and many, many more. These companies have not only taken the time to learn the new cloud technologies but can help with implementation, change management and be a guide in understanding what apps to move or build in the cloud first.

5. What about Security? This still comes up but less so as an objection but much more as a requirement for the business applications. They want to run their apps on a cloud they can trust. And they want to get the information that has made 10s of thousands of other customers rely on the cloud. Security and trust have to be earned daily by the cloud providers. But the FUD from the on-premise vendors that was working a few years ago is no longer working - customers want information on how the cloud providers ensure security. A very valid request that all of us should make from anyone we do business with.

This shift had taken place in silicon valley a few years ago where VCs would almost require you to build on the cloud and not 'burn' your money on building out your own. And now I see a similar mindset taking hold with many of the forward looking CIOs.

And yes, there are always a few that don't want to do anything with the cloud. But then there is a significant minority of people who have never and probably will never deposit a check or cash via an ATM.

1 comment:

Being Guided said...

I think that the biggest problem with cloud computing is letting CIOs drive the agenda into business-as-usual. That's why has zero innovation. Siteforce is just another CMS. Look at Azure..., AWS..., et al. Deadly dull stuff from left-brain dominated producers and consumers.

What the cloud needs is an infusion of Web 2.0-style enabling - the promise of business users as active contributors to creating simple cloud apps. So, in this context, Siteforce CMS should have the UI of Squarespace - not another IT-centric CMS. Clicks-and-Code should be replicated as a business logic PaaS on Heroku.

Sadly, I fear none of this Coghead-like stuff will happen. We'll go on perpetuating the grey science of left-brain IT - from Cobol to Scala - it's all about preserving the elite and discouraging simplification of IT.