2015 was a banner year for the cloud

All the leading clouds added major new capabilities as adoption ramped up -- and one major player bowed out

2015 was a banner year for the cloud
Wonderlane via Flickr

The cloud turned out to be the platform on which the most interesting enterprise tech was built this year. Why not? After all, cloud is the new hardware -- it makes sense that the most advanced tech would capitalize on the most advanced infrastructure.

We also saw some dramatic shifts among the major cloud players. Without question, 2015 was a year to remember. Let's have a look at the most significant developments: 

Amazon doubled its lead. In May, Gartner declared that AWS had 10 times the cloud server capacity of its nearest 14 rivals, up from its 2013 estimate of five times the capacity. Those are some awfully round numbers, but hey, they're the best we have. As InfoWorld's Andrew Oliver says, "'Cloud' generally means Amazon Web Services … real dominance is when the people assume by default."

Cloud became the machine learning platform of choice. Many want the benefits of machine learning for their applications, but few want to go to the trouble of setting up, say, a GPU cluster for deep learning. Amazon Machine Learning and Microsoft Azure Machine Learning both debuted in 2015, enabling anyone writing applications on those clouds to tap machine learning APIs. IBM acquired AlchemyAPI and added it to its Watson Developer Cloud, while Google open sourced its TensorFlow machine learning project.

Microsoft set a benchmark for the hybrid cloud. Microsoft made its intentions clear: Windows Server and System Center 2016, together with Azure Stack, will deliver a local environment very similar to that of the Azure public cloud. The goal is to create a true hybrid architecture, where workloads are fully portable between private and public clouds. The Azure Service Fabric PaaS, introduced in April and currently in preview, will run in both of those environments as well.

Google got serious. This summer, Google announced the formation of the CNCF (Cloud Native Computing Foundation), which will take Kubernetes as a starting point to build out an ecosystem for container scheduling, management, and orchestration. Eventually, those capabilities could be deployed in both public and private clouds, enabling true application portability at the container layer. It's a big deal -- if Google demonstrates sustained commitment. The hire of legendary VMware co-founder Diane Greene to head up Google Cloud is a good sign.

HP headed for the exit. Finally, after struggling to find customers for its Helion Public Cloud, HP opted to concentrate on private and managed cloud capabilities instead. This was the second major exit: Rackspace, once the No. 2 cloud behind Amazon, made a similar move in 2014. Along with the big four -- Amazon, Microsoft, Google, and IBM -- a few independents remain, including Joyent, which brings some nice secure container technology to the table.

I should add that all the major IaaS clouds -- plus Salesforce -- added IoT event processing capabilities. There's widespread agreement that clouds will provide the hub for IoT.

Analytics capabilities, such as support for Spark, were added, too. But data migration is such a big issue, it's going to take a while before the public cloud becomes the center of the universe for big data and business intelligence. AWS is clearly looking ahead to that eventuality with the October announcement of its first full-blown SaaS application, Amazon QuickSight, which will be generally available in 2016.

It may be that the great enterprise migration to the public cloud will happen sooner than we think. Perhaps the most intriguing cloud trend of 2015 was the sustained background buzz I heard from various sources that more enterprises than ever were considering moving major workloads to the public cloud. That's something we'll be tracking closely in 2016.

Copyright © 2015 IDG Communications, Inc.

How to choose a low-code development platform