Helder Almeida - Fotolia
As trends like serverless and containers continue to sweep the enterprise market, app development teams need to know what to expect.
We caught up with industry experts to get their biggest predictions for the future of cloud computing. Read on to make sure your team is prepared.
Functional programming comes to cloud, slow to reach private infrastructure
In 2018, mainstream application development will shift toward the programming of lambda functions to build stateless, event-driven cloud microservices. This trend, referred to as functional programming, will pick up pace during the year, as implementation of the enabling serverless platform permeates clouds. By the end of the year, developers will use functional methods to program over 50% of new microservices deployed in public cloud.
Functional programming and serverless technology will also make their way -- slowly -- to private clouds in 2018, thanks to a growing range of open source serverless software platforms, including Apache OpenWhisk, Fission, Gestalt, Nuclio, Fn and IronFunctions. Due to the slow adoption of serverless on-premises platforms, less than 10% of new code builds in private clouds will use functional code when 2018 comes to a close. Likewise, though functional programming has begun to come to cloud environments for edge microservices, the broad adoption of this technology for internet of things and mobility applications won't start until the end of the decade.
--James Kobielus, lead analyst at SiliconANGLE Wikibon
Serverless paves the way for event-driven architectures
Dave BartolettiVP and principal analyst, Forrester
Serverless programming patterns will begin to take over the aggregation tier of more enterprise apps. Enterprises continue to steadily add serverless patterns to new apps, as well as legacy apps being modernized. These apps will be more granular, utilize function as a service (FaaS) and be built on event-driven architectures, and they will be the canary apps that will guide modernization for the next five years. Watch for vendors to add serverless consumption models to cloud databases and other services. Soon, nearly every cloud service will have a serverless model, freeing developers from coding for infrastructure.
-Dave Bartoletti, VP and principal analyst at Forrester
Kubernetes wins container scheduling war
In 2018, we will see a race for how to most efficiently manage Kubernetes -- the current winner of the container scheduling and orchestration battle. Here it is absolutely key to understand that Kubernetes is just a standard and interface for effective, scalable container management. Containers are the real star, as their inherent advantages are independent from their scheduler. So, even if Kubernetes runs out of steam, containers are here to stay. This is especially true because containers constitute the foundation for some of the most exciting and fastest-growing technologies in 2018, namely FaaS and machine learning.
Long story short, containers will become mainstream in 2018. Enterprises will push hard to get them back under the control of IT operations, and vendors will evolve their monitoring and management tools to treat containers in a similar, but not the same, way as virtual machines.
-Torsten Volk, analyst at Enterprise Management Associates
Containers and serverless merge
We do expect to see some merging of containers and serverless as more synergies are realized, such as what we have begun to see with options through AWS and Cloud9. We also expect to see more open source options for companies that don't want to use the cloud, such as in cases of sensitive data security.
-Humberto Morales, chief architect at TeleSign, a communications API provider
Serverless containers will become popular in 2018. This concept goes beyond all of the new managed Kubernetes and containers-as-a-service offerings that were sprouting in 2017 in that developers now only have to upload the container and set a few scalability and cost parameters. AWS Fargate made the start, but all major cloud vendors will follow in 2018 to be able to absorb as many container workloads as possible.