PaaS Is The Future Of Cloud Services: Battle Lines Are Clearly Drawn

Image via CrunchBase

Earlier this week, Heroku (previous CloudAve coverage), the PaaS player Salesforce acquired during last year’s Dreamforce, announced a major new release Celadon Cedar adding some powerful features targeting enterprise customers. They also announced full Node.js support and added Ruby 1.9.2 support. Some of the features include:

• New process model with support for background processes
• Procfiles for having a fine grained control over the processes
• Full HTTP 1.1 support
• Logplex, support for logs (not just the logs for apps but also for the infrastructure underneath)
• Few command line advancements
• Automatic language detection upon deployment
• Full Node.js support
• Ruby 1.9.2 support
• Ability to configure language using Procfile
• Full process isolation for better security and performance

etc.. In fact, the new features in this release clearly highlights the enterprise focus and Salesforce’s plans for Heroku.

Where are the battle lines then?

Before Salesforce.com acquired Heroku, the main attraction for Heroku platform was with individual developers and startups wanting to get platform power without the capital expenses on the underlying infrastructure. When Salesforce.com acquired them, it was evident that they want to offer enterprise level PaaS that is much more acceptable than Force.com platform. It became evident that Salesforce.com has bigger enterprise platform plans when they announced Database.com. Therefore, some of the features in this release is no surprise for me.

I see three trends in the enterprise PaaS space with each one having their own unique advantages to meet the diverse enterprise needs. I clearly see a battle for mindshare from these three camps. However, it is too early to predict who will get the lead in this market space. I will briefly discuss the three models below (Note: Don’t pick me on the names. I am just using some terms to differentiate different ways platform services are delivered).

• Heroku Model: I would also call this model traditional PaaS model because this is how the idea of PaaS was fed into the minds of users in the early days of cloud computing. This is the hosted PaaS offering where developers just push their apps like how they push them into a repository (like github) and the rest is magic. Google App Engine, Heroku, Engine Yard (though its model is a slight variant from Heroku), PHPfog, Orchestra, Redhat OpenShift Express, Azure, etc. belong to this model. This model is targeted at developers who don’t want to know or do anything beyond pushing their apps to a repository. This is one of the reasons why Heroku was hugely successful with startups and developers doing apps on social platforms.
• Amazon Model: When Amazon decided to enter the platform level of the stack, they took a “packaged” approach to PaaS. They package the necessary middleware to run the apps in such a way that the apps can be easily deployed and run on top of AWS cloud but those developers who wants to get their hands underneath can easily do so with Amazon Beanstalk. This model tries to remove the distinction between the IaaS and PaaS. Amazon is hoping that this approach will get them more customers from both startup side and enterprise side of the market. Windows Azure’s VMRole is targeted towards this crowd. This is clearly the model to suit the devops crowd.
• Federated PaaS Model: As enterprises started embracing a more federated cloud ecosystem at the infrastructure level, it was clear that there is a need for a PaaS layer for such a federated ecosystem. Cumulogic was the first one to offer a solution under this model. VMware’s Cloud Foundry is another player who completely changed the dynamics in this space. Redhat OpenShift’s Flex and Power offerings might fit in this model too. In my point of view, this model is going to take over enterprise market in the coming years.

In the next few years, we are going to see these three models competing against each other to grab the enterprise marketshare. Heroku might be attractive for SMEs because of Salesforce’s push but I don’t see them making a big impact on the large enterprise segment. As far as Amazon Beanstalk is concerned, I don’t see them gaining much on the enterprise side unless they do something to dramatically shift enterprise’s reluctance towards public cloud platforms. The players in the Federated PaaS Model are going to fight among themselves for the medium to large enterprise pie. Among them, Cloud Foundry has better advantage now but, depending on how they execute, Redhat can offer a tough competition. I am also waiting to see smaller players show their cards on their enterprise strategy.

The coming 2-3 years will show where the market is headed with respect to platform players. Even though Cloud Foundry and Heroku has garnered lots of buzz in the tech media, it is still anybody’s game. I think this space is already crowded and any new entrant has to go up in the platform layer to differentiate themselves from the existing ones. Keep a close watch on the players in this space and I assure you it is going to be an exciting ride.

Related articles

• VMWare Disrupts PaaS Space With Cloud Foundry – An Analysis (cloudave.com)
• Heroku Revamps with Logs, Processes and Node.js (gigaom.com)
• Software AG acquires Terracotta, makes open source developer play, gears up for in memory and PaaS (enterpriseirregulars.com)
• Right Scale And Zend Announce Portable PHP PaaS (cloudave.com)
• PaaS Consolidation: DotCloud Buys DuoStack (readwriteweb.com)
• Engine Yard Now Offers Free Option For New Users (cloudave.com)
• Heroku Gets Node.js and More in New Beta Version (readwriteweb.com)
• Cloud Foundry Shakes Up the PaaS Scene (datacenterknowledge.com)
• PaaS: why businesses should pass (econsultancy.com)