Review: Microsoft Azure reaches beyond Windows

A long time ago in a century getting further and further away, Bill Gates compared MSN with the exploding World Wide Web, saw the future, and pivoted nicely to embrace the Internet. A few decades later, someone at Microsoft looked at the cloud and recognized that the old days of selling Windows Server OS licenses were fading. Today we have Microsoft Azure, Microsoft's pivot to cloud computing.

Azure is a cloud filled with racks and racks of machines like other clouds, but it also offers a wider collection of the building blocks that enterprise managers need to assemble modern, flexible websites.

There are common offerings, such as virtual machines, databases, and storage blocks. Then there are some not-so-common additions such as machine learning tools, parallel processing engines, service buses, networks, and connections to data farms. There are also some tools for debugging your code, sending emails, and installing databases like MongoDB or Cassandra or ClearDB's version of MySQL. You can draw on many tools made by companies that aren't Microsoft and many that are open source.

The wealth of options shows that Microsoft is actively trying to build a system that makes it easy for developers to produce a working website using the tools of their choice. Azure is not just delivering commodity hardware running Microsoft and leaving the rest up to you. It's starting to make it easier to bolt together all of the parts. In many cases, Microsoft bolted together many of the parts -- or at least threaded the nuts -- and your only job is to pour in some of your data. The process is rarely as simple as, say, sending an email, but it's dramatically easier than the old paradigm.

Microsoft's melting pot


The Azure service is a godsend for those who are heavily invested in Microsoft's operating systems. Many of the other clouds offer only Linux or BSD machines, or charge more to build out a Microsoft Windows server. Azure rents some basic Windows machines at the same bargain rate as Linux. The fatter machines with more power, though, will cost more if you want Windows running on them instead of Linux.

Did I say Linux? Yes, because Microsoft's embrace of open source is on full display with Azure. You can boot up a virtual machine and install popular Linux distros like Ubuntu Server 14.04 or CentOS 7.1. There is a surprisingly large number of open source distros available -- enough so that Azure feels more like an open marketplace than a Windows shop. You can spin up machines from basic distributions built by companies like Canonical or OpenLogic, or draw on hundreds of images that are more finished machines than foundations for your own creation. There are so many options, it's almost impossible to actually make a decision.

There's an Apache Solr installation from Bitnami that's ready to index your documents, or a Chef server complete with a license for maintaining 100 nodes (with more available). Or if you do want to start with a foundation layer, there are some virtual boxes that start with IBM's WebSphere application server. You'll find hundreds, if not thousands, of machine images in the selection box when you're building up a virtual machine.

If you want to build an application on Windows Server and the .Net stack, those options are still there, along with nice integrations with Visual Studio. But so are nice services for Node.js, PHP, Python, Java, and cross-platform mobile app developers.

Beyond virtual machines

It's probably a mistake to spend too much time focusing on the raw machines. Most of the new innovations at Azure are being sold as services, and some of the fanciest are referred to as engagements, machine learning, or insights. Many are quite clever and might be said to deliver answers instead of raw compute cycles.

The machine learning option, for instance, takes the data from your database and analyzes it. There's no code, just a drag-and-drop interface that links together various algorithms into a pipeline. The data comes in and then gets fed into various machine learning algorithms until you find the signal that you want to find.

Some people might imagine that the tool's interface is so simple that you don't need to be a programmer. That might be stretching it -- a non-data-scientist will probably be lost, no matter how easy it can be to drag the tiles into place. Still, the simplicity shows that Microsoft doesn't imagine Azure to be a cloud filled with commodity machines.

Another option is the collection of data lake services that store and analyze your data. Microsoft is bundling together Hadoop, Spark, and Storm while storing the data in HBase or U-SQL. It offers a thin layer of visualization that makes it a bit easier to watch what's happening and track your job. All of this runs on basic Microsoft cloud machines accessing data stored in Microsoft's disk farms, but you don't need to worry about that; you can just focus on Hadoop and your higher level questions.

There are many more similar options. One service called Stream Analytics is meant to "perform real-time analytics for your Internet of things solutions." Another offers Event Hubs to track what all of your things are doing when they phone home. Yet another provides some identity management with a directory service. The Web interfaces for these are long on slick graphics and short on the traditional command lines. And I'm only scratching the surface.

Media platform services

One of the biggest new areas is devoted to delivering "media," which largely means video and perhaps some audio. Microsoft has built out a content delivery network to make it simpler to stream whatever you want to the customers, no matter what their screen size or location happens to be. The tools will try to match the content to the screen to save bandwidth using the best set of encoders.  

If you want to add some encryption and rights management, Microsoft wants to help. Azure offers top-grade AES encryption and multiple options for managing keys and rights with systems like PlayReady or Widevine.

All of this leads Microsoft to say that Azure provides all of the features you need to build your own cloud-based DVR for content, both produced and live. It's an ambitious goal that will make it easier for video producers big and small to deliver their own content without relying on legacies like Netflix or Hulu. It's not hard for anyone with a credit card to open up a video shop online and then pay for the bandwidth if and when customers show up.

The scope of Azure and the ambition of Microsoft are both so great that it may be time to look for a new metaphor to replace the "cloud." Microsoft wants to do more than just rent out commodity machines by the boatload. Sure, there are still plenty of powerful machines at good prices. But the interesting stuff are the services that do the real work. You can roll out sophisticated analysis and juggle multiple streams of data, all from a Web interface.

In a sense, Microsoft is doing what it did decades ago when it rolled together a word processor and a spreadsheet into a single package called Office. The company has always found its success when it produced an interface that was simple enough for the average user but sophisticated enough to offer more of the power of the computer.

Microsoft is doing the same now with big data analytics and the Internet of things. It is building a nice, manageable set of tools so the average customer can conjure up amazing mathematical and analytical results that were once achievable only by the most highly skilled with the largest budgets. It's a compelling solution that's bound to find many takers.

Related articles

• Make the most of Microsoft Azure Service FabricRelated articles
• First look: Microsoft's Azure App Services sweeten cloud development
• Review: Microsoft Azure beats Amazon and Google for mobile development