Let the PaaS wars begin
Picking a PaaS path is harder than ever for newer enterprises
By now, you probably know what the latest trend in cloud computing, Platform-as-a-Service is and what it can do for your business. But equally, with an embarrassment of riches at your disposal, it's difficult to know which is the right choice or whether you should embrace PaaS right now or hold off.
Recently, a tranche of companies have appeared in the space vying for enterprise attention by each offering something different - whether that's focusing on high concurrency, latency or so on. They've never really tried to go for all at the same time, when the revolution has yet to be televised, well at least fully.
That was until Red Hat jumped first last week - laying down the foundations for their PaaS strategy and how OpenShift would utilise a large number of Red Hat's technologies (Red Hat Storage and JBoss Enterprise Middleware forming the backbone) to entice enterprises who had previously steered well clear. Their ace in the hole appears to be in offering a multitude of management options, by giving the enterprise a choice - either control it the old fashioned way in ITops or go for the new age DIY Devops world, which allows for more open source options with the developers in control.
The biggest stumbling block for enterprises currently is operations; many are just too fearful of giving developers too much power in an age where risks no longer hold the rewards they once did.
Lead Developer of Cloud Solutions at New Relic, Adron Hall knows what it's like making these tough decisions, having helped enterprises great and small make informed choices. He believes there's two main reasons for enteprise scepticism:
Firstly, enterprise shops often fear the drastic nature of the change that PaaS brings with it. It breaks the traditional bureaucratic flow of most enterprises and speeds up the business cycle in huge ways. Secondly, many people in enterprise shops still don't even know about PaaS and when they do learn about the technology it is hard to believe.
At this point I think the majority of Enterprises are just becoming familiar with cloud technology. Many still haven't even learned about or know that PaaS technologies exist. Those that are just now finding about them and stepping into cloud technologies like PaaS will have an immediate advantage over competitors.
Education is a huge problem when it comes to cloud computing, sometimes it feels like the blind leading the blind. Only through deeper understanding can any enterprise learn of the benefits, and Red Hat knows this. Their attempt to target the untapped market is a bold move putting it mildly, but perhaps it's at exactly the right time.
451 Research reckon that the enterprise PaaS market could be worth $3 billion by 2015, comfortably easing past the ailing SaaS market and leaving it in its wake. Back to the present day, there's an array of vendors currently trying to play catchup, working out where they fit in. Hall believes that 'around 30% of all new startups use PaaS Solutions to get started' so there's clearly room for maneouver here.
Red Hat's OpenShift wants to be an everyman's cloud, which means treading on the toes of VMware's Cloud Foundry - both have extremely active open source communities behind them, both have interesting deals with smaller vendors like JasperSoft and both believe that they way forward is through a particularly hybrid model. VMware of course has the edge when it comes to virtualisation.
But is this biting off more than you can chew? OpenShift has built its empire from the ground up and steadily, and clearly they've put a lot thought into it, but is it better to focus one section first and then challenge others, rather than tackling it in one fell swoop?
The key challenge for any given PaaS is changing CIO mentality. Developers know exactly what OpenShift et al are offering but do the suits? Hall, who works with AppFog and Tier3 on a regular basis, adds:
CTOs and CIOs need to realize that in almost every scenario these days they need to be providing a way to access and implement these services vs implementing the environments to build out these services. The later option is becoming more and more risky for a company and for the office of CTO/CIO to encourage. Enabling a business through software and services implemented via cloud provisions such as SaaS, PaaS, and IaaS in some scenarios is often the best option for most companies.
There are a number [of risks to the enteprise], a rather large book could be written and it still wouldn't include them all. One of the key things to think about is that traditional architectures don't hold up well in a PaaS environment. New thinking and new architectures need to be utilized to gain full benefits of PaaS usage. These new architectures take into account much more scalable, geographically dispersed, and more computer- and data-intensive capabilities over the limited traditional-style architectures. Above all, the architecture decisions are the biggest change for using PaaS effectively.
What does the future hold for PaaS? It's a bright one no doubt, but some PaaS specialists will fall by the wayside through making bad decisions in which markets they target. In time, enterprises will get up to speed in understanding which PaaS works best for them - through better documentation and informers.
At the moment though it's anyone's guess who will come out on top in this game - it won't be one vendor for sure. Red Hat has laid down the gauntlet by jumping in headfirst with a wide variety in their cloud palette. You wouldn't bet against them.
Stay tuned tomorrow for our special Java Tech Journal devoted entirely to the world of PaaS.