Why the cloud? In 2016, it was the lure of the new

This year the cloud became not only the new normal, but also an irresistible ecosystem of the most exciting new enterprise tech

Why the cloud? In 2016, it was the lure of the new
Credit: Thinkstock

Enterprises have all sorts of justifications for moving to the cloud: avoiding capital expense, adding scalability to applications, even cloud lust on the part of CEOs who want to “get out of the IT business” (um, sorry, administration still required).

But 2016 saw one reason rise to the top: Incredible new features all pre-provisioned and waiting for you in the cloud. Sure, you could stand up a GPU cluster and run your own deep learning algorithms, or jump into IoT by assembling an event-driven platform in your own data center. But … would you?

Not every potential cloud customer wants to leap into machine learning or IoT right away. But the major public clouds offer so much new functionality and the potential is so great,  particularly with machine learning, that lack of access to that stuff amounts to a competitive disadvantage.

For a simple example, say you want real-time language translation with near-human levels of accuracy. You could try and set up the software and infrastructure to do that yourself, but in a year or two when the accuracy beats that of humans, how quickly can you upgrade? A cloud service will deliver those improvements as they arrive.

Besides, developers play with new cloud APIs whether they tell management about it or not, so you might as well harness that and at least experiment with developing new cloud applications. Your other choice is to prohibit developers from experimenting with that stuff on company time – and chase away the best and brightest.

Here are the four main areas where the cloud offers not just functionality, but continuous improvement:

Machine learning: Welcome to the hottest area in tech. Judging by InfoWorld’s own traffic patterns, Google’s TensorFlow deep learning service seems to be the main reason potential customers consider Google Cloud Platform. Microsoft offers its Azure Machine Learning; IBM Bluemix provides Watson in the cloud. Amazon played aggressive catch-up at its re:Invent conference, introducing its Rekognition, Polly, and Lex machine learning services and announcing that MXNet would be its deep learning framework.

IoT platforms: The top five public clouds – AWS, Salesforce, Microsoft Azure, Google Cloud Platform, and IBM Bluemix – all have IoT platforms for securely connecting devices and developing event-driven applications. Amazon stirred the pot at re:Invent when it announced AWS Greengrass, a software core (and SDK) designed to run on IoT devices, enabling those devices to run AWS Lambda functions and connect securely to the AWS IoT platform.

Serverless computing: The industry has a long history of piling abstraction on top of abstraction. With serverless computing, worrying about infrastructure, even the virtual kind, becomes a thing of the past for developers. Serverless computing also encourages developers to grab functions from a library and string them together, minimizing the amount of original code that needs to be written. AWS Lambda is the best-known example of serverless computing, but other clouds have followed suit. Microsoft has Azure Functions and Google offers Cloud Functions.

Container management: Containers promise all sorts of agility benefits, but they need to be managed and orchestrated. The industry appears to have settled on Kubernetes as the solution of choice, one supported by all the major public clouds. Kubernetes is open source so it can be set up on premises, but rest assured most customers will opt for it as a cloud service instead. Plus, the recent introduction of the Amazon EC2 container scheduler Blox proves that you can expect all sorts of related services to emerge over time.

These are just the highest profile advanced technology areas. For example, the public cloud is also a natural place for compute-intensive analytics, because you can spin up and spin down servers as needed as well as take advantage of machine learning to make sense of results. The ever-shifting, open source Hadoop/Spark ecosystem keeps adding new projects, which the public clouds are quick to absorb and make available as services to customers.

Tapping compute, storage, and networking resources without having to procure, provision, and maintain them on premises is one thing. That was the first-order value proposition of the cloud. Today, we’re seeing vast cloud ecosystems emerge, which are becoming the go-to platforms for the most exciting new technology. Can any enterprise afford to ignore that?