While OpenStack is concerned with more than just Infrastructure-as-a-Service (IaaS) cloud these days, it’s still primarily an open-source cloud open-source consortium. In its latest release, OpenStack Stein, the cloud comes with significant network management, bare metal, and containers improvements.
For its users, which include many telecoms, the network management part is the most tasty part. OpenStack Neutron, its networking-as-a-service component, now boasts Network Segment Range Management. This enables cloud administrators to manage network segment type ranges dynamically. It uses a new application programming interface (API) extension to do this. Previously, you were stuck with manually editing configuration files.
Neutron also now treats treats bandwidth as a resource. Thus, it can work with the OpenStack Nova compute service to schedule instances only to hosts, which have enough bandwidth to do a given job effectively.
Finally, other Neutron API improvements support for Quality of Service (QoS) policy rule aliases. This enables sysadmins to to delete, show, and update QoS rules much more easily.
Ironic, OpenStack’s bare-metal provisioning service, offers better deployment templates. Standalone users can now request allocations of bare-metal nodes with submitted configuration data. Previously, you had to use pre-formed configuration drives.
As for Kubernetes — because who isn’t using Kubernetes for container orchestration? — OpenStack Magnum has greatly improved Kubernetes cluster launch time. With Stein, it will only take you five minutes per node instead of 10 to 12 minutes. With this release, too, you can now launch a fully integrated Kubernetes cluster with support for such core OpenStack services as Manila, Cinder, and Keystone on a pre-existing OpenStack cloud.
OpenStack is also adding some new services. These include:
- Blazar, the resource reservation service, introduced a new Resource Allocation API allowing operators to query the reserved state of their cloud resources.
- Placement enables you to target a candidate resource provider. This makes it easier to specify a workload migration host. In turn, this increases API performance by 50 percent for common scheduling operations. Nova’s internal Placement service will be removed by the the Train release scheduled for October 2019.
- Sahara, makes it easier to provision data processing frameworks, such as Apache Hadoop, Apache Spark, and Apache Storm, on OpenStack. It’s been been refactored into a easier-to-use architecture to make it easier to use this functionality.
Jonathan Bryce, the OpenStack Foundation executive director, summed this release up: “With Stein, operators gain new capabilities for bare metal management and networking, running high-performance workloads with GPUs, operating and Network functions virtualization (NFV) deployments. OpenStack has also become a powerful platform for managing Kubernetes clusters in private and multi-cloud deployments.”