I think today is a good day to revisit PaaS is the future of the cloud services theme. Salesforce.com today announced their agreement to acquire Heroku, the ruby on rails PaaS provider. When people talk about Salesforce.com, the buzz is mostly around their CRM or Chatter and the discussions center around how it fits into the enterprise need for collaboration. Sitting here at Dreamforce, I see lot of excitement around Chatter (the new toy) and CRM (the reliable toy). Even though collaboration is hot, I see Salesforce’s future on a different layer, the platform layer. Today tech punditry is excited about the acquisition of Heroku by Salesforce but it is just a small step in their journey towards future enterprise domination. In this post, I will try to put forward my thoughts about where they are heading.
I see Salesforce more and more as a Platform player than anything else and I think it is the right strategy (Remember PaaS is the Future of Cloud Services?). Even though their proprietary Force.com platform gained some traction with developers, it never captured the imagination of the developers in the way they wanted. They realized that developers love open platforms and don’t want to get locked into a proprietary platform with a language unique to that platform. Salesforce.com realized this and they are trying to reverse course by embracing a more open approach. This was evident when they joined forces with VMware to offer the VMforce, their PaaS service for enterprise Java programmers.
When they announced Database.com yesterday, I clearly understood where they are going. By having a database service open to all developers irrespective of the platform or language they use, they signaled their willingness to engage with more developers with diverse backgrounds. We saw some criticism about the possible network latency affecting the success of database.com. I agree that it is a serious impediment to widespread adoption of Database.com. I also acknowledge that some companies attempted to do something similar and gave up because of network latency issues. But Database.com is just one small piece of their larger strategy.
Their strategy is not just to monetize Database.com somehow. Their strategy is to build a more open platform that could target enterprise developers. They are going to add support for more and more languages that could help developers build apps for the enterprise. Ruby through Heroku is another step in that direction. By keeping Heroku an independent platform, they are signaling their intention to be a more open platform provider. Heroku is not the end of their platform strategy. Believe me. They are going to embrace diverse range of platforms offering developers the choice they demand. By keeping Database.com open, they can easily lure enterprises who still see the world with hybrid lens. Enterprises can push their non-critical data to database.com while keeping their critical data inside. Then they could take advantage of Salesforce development platforms for their app needs seamlessly. That is why I thought Database.com is a smart move.
In my post about Database.com, I pointed out how Salesforce is also helping developers write modern applications with social and collaboration built into their apps. Today’s trend is about consumer apps having collaboration and social components built into the apps so that we can keep in touch with our friends and family seamlessly. Tomorrow’s trend is about building collaboration and social features tightly into enterprise applications. Marc Benioff, CEO of Salesforce, has been emphasizing this for a long time. Many among the punditry thought it is about pushing Chatter into the enterprise. While it is partly true, I always thought it is about pushing all of the enterprise software to be like Facebook natively (rather than as an add-on). For it to happen, social and mobile should be in the very DNA of the developer platform and Salesforce is trying to do just that.
I had this suspicion when they partnered with VMware to lure Java developers and, now, with Database.com and their plan to acquire/build PaaS offerings supporting many different languages, I am pretty convinced where they are going. if you try to put into context their strategy to build open platforms with their push to have social and mobile to be part of these development platforms, it is easy to understand where they are going. Clearly, they are pushing to get enterprises embrace modern day applications and they want to be at the center of this change. If they can convince the enterprise customers to walk along with them, Salesforce will be the company startups will be looking to dislodge in the next 10 years.
Related articles
- Salesforce.com Buys Heroku For $212 Million In Cash (techcrunch.com)
- Salesforce to buy Heroku’s Ruby Cloud for $212 Million (gigaom.com)
- Salesforce’s Database.com: Fat SaaS Here We Come! (enterpriseirregulars.com)
- PaaS bystanders squished as database giants spar (zdnet.com)
- Salesforce revs up the cloud data tier era with Database.com (zdnet.com)
- PaaS Strategy: Data not Code (enterpriseirregulars.com)
- Database.com: Salesforce’s New RDBMS as a Service Offering (cloudave.com)
- Dreamforce: Decade of Inspiration (enterpriseirregulars.com)
Krish,
From a business point of view, infrastructure business is a very low-margin business. Companies like Amazon is used to lower margins and they have optimized their strategy around lower margins. But it is safe to say that Salesforce is not used to lower margins. It’ll be interesting to see how this plays out in the long run.
On a related point…Database.com uses Oracle underneath to compete with Oracle? Heroku runs on Amazon EC2 and now will end up completing with Amazon? This doesn’t make much sense.
Yeah, I raised this question about Database.com running Oracle and Oracle’s strategy in my post yesterday. When I pushed them, they tried to spin it up saying they offer additional services (as if Oracle cannot offer them around their database).
Re: Heroku on EC2, I don’t expect them to stay on AWS infrastructure for long.
I totally agree with this article. PaaS is the wave of the future because most PaaS platforms have standardized their infrastructure components, infrastructure architecture and infrastructure operations to provide very low-cost services. These savings are passed on to their customers. This means that a startup product owner could start a PaaS-based software application and not have to worry about buying, installing, securing, and configuring servers, networks, scaling, caching, loadbalancing, database servers etc. This means that the time-to-market for any software idea can be greatly sped up. For example, an insightful and capable product manager and a great Ruby/Rails developer could conceive an application, validate the idea, develop and deploy it, and then validate the first release within a matter of weeks or even days. With heroku.com, the initial budget (without labour), only has to pay for a domain name and some cheap ads with AdSense/Facebook to get users. i.e. Not more than $100. That is a very inexpensive way to validate an idea.
Poorly conceived products are the greatest expense in software development, in my opinion, whether in the enterprise, or out in the wild. Anybody who has done software development at large companies is painfully aware of how much software development is thrown away on products nobody wants. Testing product ideas and weeding out the bad ones becomes much cheaper when a good PaaS (e.g. Heroku.com or Google App Engine) is used to host the software. This is where Salesforce/Heroku will really help lower the overall cost of software development.