The OpenStack community today released Train, the 20th version of the open source cloud infrastructure software – widely deployed in the cloud and hosting industry. For the Train release, OpenStack received 25,500 code changes by 1,125 developers from 150 different companies. With the Train release, the OpenStack community has delivered features targeting emerging use cases like AI and ML while improving data security and infrastructure resource management for enterprises.
OpenStack is uniquely suited to deployments of diverse architectures: bare metal, virtual machines (VMs), graphics processing units (GPUs) and containers. The software now powers more than 75 public cloud data centers and thousands of private clouds “at a scale of more than 10 million compute cores.”
Among the dozens of enhancements provided in Train, the 3 highlights include:
- Enhanced security and data protection
- Advancements for artificial intelligence (AI) and machine learning (ML) use cases
- Improved resource management and tracking
Enhanced Security and Data Protection
- Support for building software RAID – Ironic bare metal service protects services from disk failures.
- Hardware-based encryption – Nova features a new framework supporting hardware-based encryption of guest memory to protect users against attackers or rogue administrators snooping on their workloads when using the libvirt compute driver. This feature is useful for multi-tenant environments and environments with publicly accessible hardware.
- Data protection orchestration – Karbor adds events notifications for plan, checkpoint, restore, scheduled and trigger operations. This feature allows users to backup image boot servers with the new added data which is located on the root disk. Upstream development was led by China Mobile.
Increased Accelerator Support for AI/Machine Learning
- Accelerator Lifecycle Management – In the Cyborg project, the Cyborg-Nova interaction spec creates a blueprint for launching and managing VMs with accelerators.
Improved Resource Management & Tracking
- The Train release completes the transition of the Placement capability into a standalone service, available independent of Nova. In Train, service response time is 0.7 seconds down from 16.9 seconds prior to the Stein release.
- Nova features improvements to the scheduler to more intelligently filter results from the Placement service.
- Watcher receives data from Placement, improving the Watcher compute data model.
Additional Release Highlights
- Enhanced high availability – Several Cinder drivers add support for new features like multi-attach and consistency groups.
- Nova – Nova offers live migration support for servers with a NUMA topology, pinned CPUs and/or huge pages, when using the libvirt compute driver. Nova also delivers live migration support for servers with SR-IOV ports attached when using the libvirt compute driver. This enhancement, driven by the StarlingX project, enables a hypervisor to run both pinned and unpinned workloads in the same hypervisor.
- Documentation – Improved documentation was a community-wide goal for the Train cycle in an effort to be a more accessible open source community. OpenStack operators can now access PDF-generated documentation by project.
“OpenStack is the market’s leading choice of open-source infrastructure for containers, VMs and bare metal in private cloud,” said Mark Collier, COO of the OpenStack Foundation. “Looking forward to 2022, market watchers like 451 Research see an emerging $7.7 billion market for OpenStack products and services and $4.3 billion for application containers. As the overall open source cloud market continues its march toward eight figures in revenue and beyond, it’s clear that the OpenStack and application container markets are advancing hand in hand.”
OpenStack Train Already in Production on Launch Day
VEXXHOST, a Canadian cloud service provider, uses OpenStack to power its public cloud and hosted private cloud deployments. The company employs a rapid-deployment cycle that tracks and tests against the latest stable branch of each software release. As a result, today VEXXHOST is running the Train release in production via its private cloud deployments, targeting its public cloud datacenters next.
“With this new release out, we are excited to be able to offer the latest OpenStack has to offer on the first day,” said Mohammed Naser, CEO of VEXXHOST. “Pushing out new capabilities and features this quickly is something that our users have come to expect and very much enjoy. As we’ve done on other recent releases, we’re thrilled to deliver that once again with Train.”