AppFog's acquisition of Nodester: polyglot PaaS the future?
It’s fair to say that the platform-as-a-service market is
heating up. From huge open-source juggernauts like Red Hat to
younger startups on the up, it seems that all the hype might just
be amounting to something.
You also can’t fail to notice just how crowded the market has become so quickly. To truly stand out, companies have had fine-tune their pitches to developers and offer something unique. Not a day goes by without someone announcing new to their conveyor belt of products.
But one company truly striking out on their own (which we at JAXenter have eagerly been following over the past few months) is Portland’s infrastructure agnostic and polyglot-preaching AppFog, currently riding the crest of the PaaS wave.
Originally starting life off as PHP Fog, founder Lucas Carlson led the burgeoning PaaS down other avenues after initial success with a pure PHP PaaS. Seeing an opportunity to expand into other languages as part of the package, Carlson realised the potential and the pitfalls of the polyglot approach.
“We ran some funding ($9.8m in two rounds),” he says, “and with that funding we were able to expand our vision, to move from just PHP to deal with multiple languages. The problem with that up front was we didn’t want to give a bad experience on so many languages.
“We don’t want polyglot PaaS to be the bare minimum PaaS for every language. We want it to be the best Java, the best Node, the best PHP Paas there is.”
When asked why they were heading down this route, when some vendors opt for one language, Carlson said that whilst PaaS has been around for while, the biggest problem is that it hasn’t yet tackled a polyglot approach head on.
“The problem polyglot solves uniquely is one that enterprises have a crisis today with. These enterprises are scaling and managing sometimes 1,000 developers and you need something that supplies a standard interface across the development platforms and technologies,” he said.
The relationship with VMware’s in-beta platform Cloud Foundry provides the key here, since it follows an interesting trend in the PaaS market of co-opetition between competitors. The decision to hook up to Cloud Foundry a year ago, enabled them to utilise a universal API, with the hope of spanning across a plethora of infrastructure environments such as Rackspace and Microsoft Azure. AppFog effectively becomes the middle-man for enterprises, simplifying what can often be an arduous task.
It’s a very savvy move, allowing AppFog’s distinctive platform to permeate the community already invested with Cloud Foundry but also reiterates their desire to offer the user as much choice as possible, whilst reducing complexity at the same time. AppFog are also keen committers to Cloud Foundry, being an active contributor to the codebase.
Carlson told us that Cloud Foundry’s vision struck a chord with their company “The thing with Cloud Foundry that was very, very unique was that it approached it from the open source perspective so that there would be support for multiple languages. But instead of having to build up the support yourself every time and having expertise and the operational knowledge to run that, the people in the community would build it.
“For example, a day after they released Cloud Foundry, there was a pull request for Erlang, and now they had Erlang supported. Nobody had done Erlang-as-a-Service until Cloud Foundry and it helped me decide that this was the right tool worth moving forward with.”
Community building is certainly a big deal with AppFog as they try and create as many bridges as possible. Just this week, the company announced they had purchased Nodester, the most recognised PaaS for the Node.js community.
Carlson said that AppFog and Nodester had developed “an instant bond”:
“Nodester also spent two years forming a deep relationship with
the Node community and has been adding depth and breadth to our
support at AppFog,” he said. “On the depth side, it adds more
technology, getting some of the missing components like Websocket
support and on the breadth side, it adds a large community into the
AppFog ecosystem and helps us push forward.”
Node.js seems to be the darling of the wider industry at the moment. The hugely popular and highly scalable server side language hasn’t taken long to become one of the premier choices when writing a next-gen web application.
For AppFog, it ticks another box in their continuing quest for polyglot supremacy and strikes us as a no-brainer to purchase a company on the rise. For Nodester users, it opens the door to Ruby and Java within one platform. The acquisition also brings Websocket and Node 0.8 support to AppFog - an absolute must for any developer looking to create mobile and modern web apps. It’s just another string to AppFog’s bow, but one of the biggest additions yet.
But Carlson understands the challenges that lie ahead for AppFog and PaaS in general, namely vendor lock-in:
“Developers loves PaaS - that’s been the drive, the adoption factor and the only way it’s been maturing. CTOs and CIOs within enterprises don’t care or want to know about it and there’s a good reason for that. It’s not that they don’t want what AppFog has built, it’s because PaaS has come to mean more vendor lock-in. The problem for a CIO or CTO is that they’ve got 5 to 10 datacenters where code runs. To them, what PaaS has meant so far, oh great here’s another locked in datacenter.
“I think that’s the big open question - is PaaS is going to evolve or is there going to be a next-generation technology?”
Carlson isn’t the only one asking that, and we’re liable to agree with the latter. PaaS’s biggest problem is the definition often cast onto it by some who don’t believe it’s worthy. Perhaps a makeover is needed? Either way, AppFog seem to be leading the way towards the next-generation of PaaS.