As a result, these rarefied skills have become extremely difficult to find and expensive to retain, driving up the cost of machine learning systems and making it difficult for enterprises and smaller web firms to apply the technology. In the data scientist talent shortage is opportunity, however, and a new breed of software platform is rising to meet this need. Building upon the low-level big data infrastructure now available, these new platforms seek to democratize machine learning and advanced analytics, making their benefits available to enterprises and firms who either can't afford or can't find enough PhDs and data scientists. The first of this coming wave of machine learning-powered platforms is launching at this week's O'Reilly Strata conference. Here are three companies leading the way.

Skytree

This is not to diminish the Skytree value proposition in the least. By analogy, if the skills required to use machine learning are akin to knowing how to drive, and the skills required to build a production machine learning system are akin to knowing how to build a car, the Amazons and Ebays of the world have built their machine learning 'vehicles' from the tires up, while what Skytree does is allow you to drive a Ferrari (a) without knowing how to build one and (b) on a Kia budget. Skytree Server is priced on a subscription basis, starting at $2,999 per year for up to 4 cores. It is also available as a Free Edition, which has the ability to process up to 100,000 rows of data.

BigML

Once the model has been generated, additional data can be fed into the system and the model will be used to make predictions about the prediction targets. BigML currently only supports decision tree models for machine learning. While this may be a limitation for true aficionados, the company argues that the decision tree technique is powerful because it can handle a wide variety of data types, is particularly intuitive and lends itself to visual representation, is a great place to start if you don't already know what kind of analysis to apply, and is easy to scale.

Continuing our driving analogy, BigML offers an easy-to-drive family sedan that appears at your driveway when you need it, takes you where you want to go, and presents helpful guidance on how you got there, ensuring that you're never lost. BigML is priced using a credit-based system at $0.05 per credit, with the number of credits required based on the size of your data, the size of your model, and the number of predictions you need to make.

Precog

At the heart of Precog are a scalable, custom-built analytics database coupled with a high-level analytics API that allows users to perform a variety of analyses by name (e.g. "optimize", "cluster", or "predict"), without getting bogged down by the details of which algorithm is best. In this way, Precog is probably most analogous to the kit cars you could order from the back of Popular Mechanics in the 80s, offering those willing to get their hands a bit dirty a way to create unique and customized high-performance vehicle without needing to engineer the engine and frame from scratch. Oh, and you can rent it, a la SaaS. Precog will become available this week to select alpha users, with a private beta expected shortly thereafter.The product is offered by ReportGrid, a year-old company initially focused on providing sexy embeddable analytics reports for SaaS companies.

The three companies profiled here represent very distinct approaches to making 'machine learning as a service' a reality, and I expect we'll see many more such offerings in the coming months.

Before you run off thinking I'm advocating the destruction of PaaS platforms, please realize that I am not. Rather, I'm referring to the shift away from monolithic, one-size-fits-all PaaS systems towards more open, loosely coupled platforms that makes it easy to consume code and services provided by third parties.

Early PaaS offerings, circa 2007-2009, were conceived of as all-in-one affairs. In fact, a big part of the value proposition that providers envisioned was its delivery via proprietary services and custom APIs that developers would use in their applications. Examples include App Engine's data store and Memcache services, the Force.com data store, the distributed cache and storage systems we built at Appistry and many more.

The fact is, the early players in the space had little choice but to roll their own. At the time, there were critical gaps in the market that needed to be filled in order for developers on PaaS platforms to be able to deliver rich, scalable applications. Fortunately for PaaS users, this is no longer the case for most application-level services.

REST Assured, We've Got Git

In order to build modern, scalable, connected web applications, developers must have access to a wide variety of third-party components and services upon which to build. With the proliferation of open source and SaaS services, these are now readily available on the open market. While both open source and SaaS predated the earliest PaaS offerings, in recent years the advent of GitHub and the popularity of REST-based Web services has played a significant role in expanding the selection of building blocks available to developers.

GitHub, by dramatically lowering the barriers to collaborating on and sharing open source projects, has become an "App Store" of sorts for developers, and is home to over one million projects. Likewise, the popularization of Roy Fielding's REST model for web service APIs has simplified developer access to the many application- and app-infrastructure-oriented SaaS services now available. It's now possible to store files, query and analyze data, send emails, create maps, subscribe to messages, encode videos, and much more, just by sending simple HTTP-based commands. (If you've never visited the ProgrammableWeb API Directory, the selection will blow your mind.)

This Cambrian explosion of high quality components and services has made web application development a much more productive affair for developers. And because the market has removed the burden of providing these low-level building blocks, those PaaS providers ready to embrace openness stand to gain great advantage.

Modular PaaS is Better for Providers, Too

While the end-user benefits of an open approach to PaaS, namely increased choice and reduced lock-in are apparent, the advantages of a modular approach to PaaS are two-sided, benefiting providers at least as much as users. This is because, as a PaaS provider, it's simply too hard to deliver both a solid application platform and the services that plug into it. For most businesses, such a thing would spread their development resources too thinly, even if they had the necessary domain expertise, which most don't. In addition, because the open source and SaaS genies have left their bottles, trying to do it all puts the provider at odds with their customers.

By building and offering an open platform able to easily consume third-party components and services, and by cultivating a thriving ecosystem of the tools' providers, second-generation PaaS vendors can improve their own chances of success while creating a better world for their users.

Past performance notwithstanding, many industry watchers have predicted 2012 to be the breakout year for PaaS in the enterprise. Gartner, for example, reportedly communicated at its November Application Architecture Development & Integration conference its belief that 2012 marks the beginning of a rise in PaaS adoption from almost zero (3% of enterprises) to nearly half of all enterprises (43%) in 2015.

While it remains to be seen whether 2012 goes down in the history books as the year PaaS makes good, much of the groundwork for PaaS' predicted success was laid in 2011. Here are some trends from the past year:

Heroku Beyond Ruby

Salesforce.com's acquisition of Heroku for $250M, an estimated 50-100x revenue, while announced in December 2010, set the stage for a year of brisk investment in PaaS. In January, on the heels of the acquisition, we saw a flurry of Heroku investments and product launches including PHPFog (PHP), Gondor.io (Python/Django), Nodejitsu (Node.js) and CloudBees (Java). With this activity, the reach of PaaS was significantly broadened.

The Rise of Polyglot Platforms

The early flood of single-language PaaS platforms gave way to a move towards multi-language platforms later in the year, perhaps precipitated by DotCloud's entrance in the market with a vision of "One Platform, Any Stack." Established providers like Red Hat OpenShift and Heroku broadly expanded platform support, while the aforementioned PHPFog relaunched as AppFog with a new multi-language platform.

The rise of so-called polyglot PaaS platforms, while derided by some as stifling innovation, is significant in that it marks a departure from early "one size fits few" approaches to PaaS, towards something more flexible, familiar, accommodating, and with a bit less lock-in... Just what the enterprise user is looking for.

Enter Cloud Foundry

VMware's Cloud Foundry, which launched in April of 2011 and covered extensively in ReadWriteCloud, was not the first multi-language PaaS. Nor was it the first open-source PaaS, or the first PaaS backed by a major player in enterprise IT. It wasn't the first PaaS to be readily deployable in both public and private cloud environments. Nor was it the first PaaS to embrace the power of an extended ecosystem of developers and partners.

What made Cloud Foundry a game-changer in 2011 is the fact that it was the first PaaS to offer all of those things--on your laptop, in your data center, or in the cloud.

PaaS Ecosystems Flourish

Last but not least, the developer services ecosystems that have formed around the major offerings were expanded greatly in 2011. Users of these platforms can now easily add a wide variety of services such as caching, messaging and databases (SQL and NoSQL) to their applications. These ecosystems are powerful in that they simultaneously have made PaaS platforms more productive for developers and more profitable for providers, while they have reduced the threat of lock-in for users.

The role of these ecosystems is key, and I'm planning to explore this topic further in a future article.

In the meantime, I'll continue rooting for the success of PaaS customers and providers, with full knowledge that they are building upon solid foundations laid throughout the last year.