Microsoft Azure started to come into its own as a viable alternative to AWS for enterprise IT shops in 2016.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
Hyperscale cloud providers experience a learning curve of outages and bugs as they operate at scale, and Microsoft was no different. But over the past year or so, those grumblings have largely subsided, as Microsoft stabilized the underlying architecture and added a range of new Azure features to compete directly with Amazon Web Services (AWS).
"People expect the second or third cloud provider out the gate would learn somehow vicariously through Amazon's mistakes, but there's only so much you can learn from the publicly released information," said Patrick McClory, senior vice president of platform engineering and delivery services at Datapipe, an AWS and Azure-managed service provider. "Even without that, Amazon is pretty upfront that there's a lot of organizational learning that needs to go into that."
Over the past two years, Azure has moved past the stumbles and concerns about the platform's clunkiness. Outages are few and far between thanks to improvements to the underlying networking, patching and automation, while latency has been addressed with improved throughput.
The shift to Azure Resource Manager was a major interface upgrade to help customers manage their environments. Other improvements around lifecycle management included the Azure Usage and Billing Portal, Azure Monitor and Application Insights.
"The vision was there, and it seems like the execution is finally getting there, too," said John Peluso, senior vice president of product management at AvePoint, a Microsoft independent software vendor that sells and uses Azure.
Like the other major public cloud providers, Azure has slowly moved toward microservices and away from virtual machines (VM), which are viewed as a hallmark of infrastructure as a service. Microsoft has continued to expand Azure features around containers, while also exploring more abstraction with services such as Azure Functions. Other services put the management of the underlying infrastructure in the hands of Microsoft, such as Azure Service Fabric for application lifecycle management, Power BI Embedded for live interactive reports on applications and Analysis Service for data modeling.
Machine learning is another area of focus for Microsoft, in response to increased customer interest. Milliman Inc., an actuarial service provider based in Seattle, runs its Integrate platform as a service on Azure, and wants to integrate this technology with Azure data services to create models, detect anomalies and improve the overall performance of their service.
"If we can get smarter about workloads we're running, it allows us to be a lot smarter around pricing and pricing opportunities for our end users," said Paul Maher, CTO of life technology solutions and principal at Milliman.
Openness and more variety on Azure
Microsoft as a whole has embraced openness under CEO Satya Nadella in a way that would have been unthinkable just a few years ago, and Azure is one of the best examples of that push. Red Hat Enterprise Linux became available on Azure, while Azure Container Service provides support for all the major container orchestration tools.
Two new VM groups were added: F-Series, with a focus on price performance, and the GPU-based N-Series. Other additions included Cool Blob Storage for cheap archival storage.
In addition, upgrades to Azure Security Center addressed advanced threat detection, web application firewalls and centralized security management.
Customers still wait for Azure Stack
It's hard to talk about all the advancements made to Azure in 2016 without mentioning what was supposed to be released, but wasn't. Azure Stack, the service that essentially brings the public cloud capabilities on-premises, was first mentioned in 2015, but was delayed as Microsoft tackled the challenges of scaling down its public cloud to fit an individual data center. (By comparison, the recent AWS-VMware partnership seeks to bring on-premises environments to the public cloud.) Customers have also raised concerns about the product's restrictions to specific hardware.
Now in technical preview, Azure Stack is expected to be fully available sometime in 2017.
Al Sadowskiresearch vice president, 451 Research
"That could be a game changer," said Al Sadowski, research vice president at 451 Research. "That's something that AWS is not doing. So truly having a full stack for hybrid scenarios on-premises and off-premises is an impressive story to tell."
Azure still has a ways to go before it truly challenges AWS, but Microsoft's relationships with enterprises could be a big advantage. In the meantime, users expect further integration of services and ease of use in 2017.
"The next phase for them is taking that enormous amount of services they offer and [building] an organization around it so it's a little bit more accessible from the point of view of a first-time user," McClory said. "I know they're working on it, but right now it's a big sea of services."
Trevor Jones is a news writer with TechTarget's Data Center and Virtualization Media Group. Contact him at firstname.lastname@example.org.
Top cloud computing predictions for 2017
How to invoke an Azure runbook with a webhook
Explore Azure cloud storage options