Microsoft Announces Azure ML, Cloud-based Machine Learning Platform That Can Predict Future Events


Microsoft has been on quite a cloud roll lately and today it announced a new cloud-based machine learning platform called Azure ML, which enables companies to use the power of the cloud to build applications and APIs based on big data and predict future events instead of looking backwards at what happened.

The product is built on the machine learning capabilities already available in several Microsoft products including Xbox and Bing and using predefined templates and workflows has been built to help companies launch predictive applications much more quickly than traditional development methods, even allowing customers to publish APIs and web services on top of the Azure ML platform.

Joseph Sirosh, corporate vice president at Microsoft, who was in charge of the Azure ML, and spent years at Amazon before joining Microsoft to lead this effort, said the platform enables customers and partners to build big data applications to predict, forecast and change future outcomes.

He says this ability to look forward instead of back is what really stands out in this product.

“Traditional data analysis let you predict the future. Machine learning lets you change the future,” Sirosh explained. He says by allowing you to detect patterns, you can forecast demand, predict disease outbreaks, anticipate when elevators need maintenance before they break and even predict and prevent crime, as just a few examples.

Sirosh says the cloud really changes the dynamic here because it provides the ability to scale, and the service takes care of much of the heavy lifting that would have taken weeks or months for companies trying to do it themselves in-house in a data center.

“The cloud solves the last mile problem, Sirosh explained. Before a service like this, you needed data scientists to identify the data set, and then have IT build an application to support that. This last part often took weeks or months to code and engineer at scale. He says Azure ML takes that process and provides a way to build that same application in hours.

What’s more is it supports more than 300 packages from the popular open source project R used by many data scientists.

Sirosh says the hope is that as more people use the platform and generate APIs and applications, and create what he called, “a virtuous cycle between data and APIs. ” People have data. They bring it to [Azure ML] to create APIs. People hook into applications then feed data back to the cloud and fuel more APIs, “he explained.

The product is currently in confidential preview, but Microsoft did mention a couple of examples including Max 451, a Microsoft partner working with large retailers to help predict which products customers are most likely to purchase, allowing them to stock their stores before the demand.

Carnegie Mellon University is working with Azure ML to help reduce energy costs in campus buildings by predicting and mitigating activities to reduce overall energy usage and cost.

Microsoft is not alone in this space, however. IBM launched Watson as a cloud service last winter for similar types of machine learning application building and just last week a startup called Ersatz Labs also launched a deep learning artificial intelligence cloud platform.

Azure ML goes into public preview next month. There is no word yet on the official launch date.

Microsoft Azure Gets New Tools For Hybrid Clouds And Simplified Cloud Storage Service For Businesses

Today is a big day for Microsoft. It’s making a large number of announcements around its developer tools at its enterprise-centric TechEd event in Houston today and in addition, it is also bringing a number of new features to its Azure cloud computing services. Among these are the general availability of Azure ExpressRoute for creating private connections between Azure and on-premise environments, a new simplified cloud storage service for Azure for businesses and a number of new security features.

Some of these features are available now, some are in preview and some will launch over the course of the summer. ExpressRoute is coming out of preview on Monday, for example. By working with AT&T, BT, Equinix, Level 3, SingTel, TelecityGroup, Verizon and Zadara Storage, ExpressRoute creates a direct and fast connection between Azure and the infrastructure that enterprises already have on their premises or in a co-location data center. Using these provider’s networks, ExpressRoute connections never touch the public internet, making them faster and more secure that typical connections. For enterprises that are working on adopting cloud computing through hybrid clouds, this is a huge deal. None of this comes cheap, though. During the preview, Microsoft offered a 50% discount and even then, a 10Gbps connection cost $5,000 per month with unlimited inbound data transfer and 250 terabytes of outgoing data transfer included. All of this comes with a 99.9 percent SLA.


Maybe the most interesting update, however, is the launch of Microsoft Azure Files in preview today. The idea here is that while cloud storage was designed with new applications in mind, it’s still very hard to move applications to the cloud when you have to rewrite your complete storage stack first. Microsoft argues that many of these legacy apps were written for on-premise applications where file shares were the de-facto storage standard. So to help businesses move these applications to the cloud, Azure Files essentially provides businesses with a cloud-based file sharing server that uses the standard SMB protocol that are part of the Windows API.


A few months ago, Microsoft acquired API management service Apiphany with the clear intent to integrate it into Azure. Today, some of that work is coming to fruition with the launch of Azure API Management in preview. This service helps businesses publish their APIs internally and to outside developers.


Among the new security solutions are the disaster recovery service Azure Site Recovery (previously known as Hyper-V Recovery Manager), which will launch in preview next month and a new anti-malware service for Azure in preview for both cloud service and virtual machines. In addition, Microsoft is partnering with Symantec and Trend Micro to integrate their antivirus technologies on Azure virtual machines.


Other updates the company announced today include the public previews of an internal load balancing service, an Azure Redis Cache service and the general availability of the A8 and A9 virtual machines for compute-intensive applications on Azure. Microsoft will also soon allow developers to permanently reserve an IP address from their own pool.

All of these new services clearly point at Microsoft’s interest in getting more enterprise customers onto Azure. While Microsoft may still be catching up to Amazon’s Web Services in many ways, Amazon has never quite focused on this market and Microsoft actually has a chance to leapfrog its competition here if it plays its cards right.