Adoption of containers and serverless computing services is growing in the triple digits quarter to quarter. In addition, there is evidence that artificial intelligence and machine learning applications are rapidly growing among cloud users.
That’s the word from Cloudability, which just released data on its customer adoption patterns throughout 2017. Focusing on Amazon Web Services customers, the study’s authors analyzed data from over 1,500 organizations, covering more than $2.5 billion in spend in 2017. The dataset for this analysis exceeds 75 billion data points, three billion CPU hours, and 700 unique compute resources.
Among Amazon Web Services users, container adoption grew 246 percent during the fourth quarter of 2017, and 206 percent the quarter before that. Cloud customers have been gravitating to Kubernetes, with a 70 percent share of container orchestration solutions. This is up from a 45 percent share in the first quarter. Apache Mesos was the second-ranked engine at 25 percent; however, this share is down from 45 percent at the beginning of the year.
There has also been huge growth in adoption of serverless computing among cloud users. In the fourth quarter of 2017, serverless adoption grew by 667 percent among the sites tracked, the survey’s authors report. This is up from 321 percent just the quarter before. “Serverless continues to be attractive to organizations since it doesn’t require management of the infrastructure,” the report’s authors observe. “As companies migrate increasingly to the cloud and continue to build cloud-native architectures, we think the pace of serverless adoption will also continue to grow.”
The study’s authors also looked at cloud CPU consumption to draw conclusions about how people are deploying cloud power. The dominance of general-purpose workloads (employed 43 percent of the time) shows that most organizations start their cloud journey by moving development and test workloads, mostly provisioned as standard instances.
With increased application performance requirements, the hours of CPUs consumed grew by 85 percent over the last year, they state. “Given recent upticks in artificial intelligence and machine learning-related applications,we expect this trend to continue,” they state.
As evidence of potential adoption of resource-intensive services for AI and machine learning, the study shows. “Growth rates for high I/O instances show that customers are increasingly deploying production-scale databases in cloud. We are also starting to see the beginnings of high-GPU instance adoption, typically leveraged for AI/machine learning processing” — a growing trend in 2018.
Enterprise cloud consumers are also getting more bang for their buck, the study shows. “CPU Hours grew 84% while the spend grew only 35%. As more and more organizations increasingly continue to adopt cloud, cloud constructs such as Reserved Instances and Spot Instances are helping companies reduce spend.”