Oracle at the KubeCon + CloudNativeCon Europe 2018 conference announced that Oracle Container Engine has been tightly integrated with Kubernetes and the Fn serverless framework it developed will support both the CloudEvents specification being developed by the Cloud Native Computing Foundation and a serverless framework being advanced by the Massachusetts Institute of Technology (MIT).
Jon Reeve, senior director of product management at Oracle, says Oracle Container Engine for Kubernetes will reduce the barriers to adoption for the de facto container orchestration standard by integrating it with the storage, networking and management service that Oracle has baked into Oracle Cloud Infrastructure (OCI), its infrastructure-as-a-service (IaaS) platform.
Specifically, Oracle Container Engine for Kubernetes enables user nodes to be deployed inside a virtual cloud network (VCN) that manages all IP addresses, subnets, route tables and gateways using the VCN in addition to providing load balancing capabilities.
In terms of persistent storage, Oracle Container Engine for Kubernetes provides access to OCI Volume to dynamically provision block storage resource and a OCI Flexvolume Driver to mount OCI block storage volumes to Kubernetes Pods using the flexvolume plugin interface.
The technology also provides support for two different levels of role-based access controls and support for the Payment Card Industry Data Security Standard (PCI DSS).
Reeve says the lack of integration with enterprise-class storage, networking and compliance management tools has been holding back adoption of Kubernetes in enterprise IT environments. Most IT organizations do not want to have to go to the trouble and expense of setting up dedicated storage and networking capabilities when they have already deployed those services to support other classes of applications.
Additionally, Reeve says Oracle plans to extend that same level of support to Oracle Cloud Machine, an instance of a managed service provided by Oracle that deploys an instance of a private cloud in an on-premises environment. Oracle is also committed to providing support for containers and Kubernetes running on either virtual machines, bare-metal x86 servers or graphical processor units (GPUs) both in the cloud and on-premises, says Reeve. Oracle claims two bare-metal NVIDIA Tesla P100 GPUs running parallelized workloads on Kubernetes hosted on the OCI can achieve more than 21 TFLOPS of single-precision performance per instance.
Reeve says Oracle is also committed to ensuring serverless computing frameworks can be deployed on any cloud or on-premises IT environment. The goal is to make them a seamless extension of any containerized application, says Reeve. As part of that commitment, Oracle also revealed it contributed a function-as-a-service (FaaS) provider to the serverless framework developed at MIT to enable container-native functions to run on any Fn Cluster.
While Oracle may not have the heft of rivals such as Amazon Web Services (AWS) in the cloud, the rise of containerized applications presents Oracle with an opportunity to close that gap by aggressively bundling Kubernetes-based services with the IaaS platform it already provides. Oracle was not in a position to compete for the first wave of hypervisor-based applications that make up the bulk of current cloud applications. But Reeve says the support Oracle today provides for both virtual machines and bare-metal servers in Oracle Cloud makes the cloud services a more attractive option for many developers.