The Tungsten Fabric (TF) community is excited and proud to announce our latest release, 5.1. The TF community has been hard at work on both community and technical challenges to ensure a rich and vibrant community to solve the toughest networking challenges regardless of public cloud, orchestrator, or workload. The 5.1 release reflects that effort. It is an excellent time to take a look at Tungsten Fabric as a developer or an operator for your networking needs in this multi-cloud world. Here is a quick summary of the TF 5.1 release highlights.
- Code quality first: all known P1 and P2 bugs are squashed prior to release!
- New features for containers and hybrid cloud to make life awesome for developers.
- Performance and scalability enhancements to delight network operators.
Additional information on what’s new in the 5.1 release
Here are some of the Tungsten Fabric 5.1 benefits and features in more depth!
- A strong commitment to code quality.
- The TF community has learned a lot over the years about how to create feature-rich networking for all orchestrators and public clouds. The latest release of TF squashes over 1600 P1 and P2 bugs prior to release, and there are currently zero known P1 and P2 issues! Special thanks to all TF community members for your dedication to optimizing development practices and the countless hours of effort on rigorous software testing.
- Networking feature enhancements for a variety of customer use cases.
- Tungsten Fabric 5.1 now supports container service chaining by supporting containers with multiple network interfaces. Service chaining is a powerful, differentiating feature of TF which is now available for all cloud-native / container workloads.
- The TF 5.1 control plane also supports extended communities allowing operators to build flexible routing policies to integrate with your existing network designs.
- Finally, are you trying to offer “cloud router” VRF services using BGP to your customers? The TF 5.1 implementation of BGPaaS now includes support for selectively peering with control nodes, so you can integrate VNFs running BGP while guaranteeing high availability when control nodes fail.
- High performance networking.
- TF 5.1 continues its dedication to high performance networking through supporting “fat flows” to optimize flow setup for high performance use cases. In scenarios where TF must support millions of flows for use cases such as subscriber internet access, fat flow support allows network operators to aggregate flows reducing latency in flow setup.
- TF 5.1 also features a number of flow table optimizations to reduce performance bottlenecks. Specifically, the TF vrouter now leverages Cuckoo hashing to remove performance bottlenecks where vrouter must perform numerous flow table lookups under heavy load.
- TF 5.1 introduces DPDK optimizations through batch processing to increase data plane performance. We want to meet your most demanding throughput needs!
Many clouds; one (Tungsten) Fabric
TF’s mission is to provide the network fabric for any cloud, any orchestrator, and any workload. Take a look at some of our latest enhancements to support this multi-cloud world.
- Tungsten Fabric as a networking overlay for any orchestrator and any workload.
- TF 5.1 supports the following orchestrators with its latest release.
- Kubernetes 1.12
- OpenShift 3.11
- OpenStack Ocata and newer
- VMware vCenter 6.7
- TF 5.1 supports the following orchestrators with its latest release.
- Tungsten Fabric in public cloud.
- Multi-cloud is table stakes for nearly every customer running workloads in public cloud. Tungsten Fabric 5.1 supports AWS and Azure with GCP support in progress to support developer’s and operator’s mult-cloud needs!
Want to learn more about the 5.1 release?
The Tungsten Fabric community documents all new features as blueprints. Check out all new 5.1 features here. (Note for those familiar with Tungsten Fabric and its history, we migrated blueprints from GitHub to Jira.)
Getting Started and Next Steps
Are you just getting started with TF or do you need a refresher? First, check out our architecture document. Then, try kicking the tires on Tungsten Fabric Carbide to launch a working TF cluster with Kubernetes in AWS in 15 minutes. We also have a tf-devstack project in beta with a number of community members trying it out before making it official. With our instance of tf-devstack, developers will be able to try out Tungsten Fabric on their laptops quickly and easily. Stay tuned for an announcement on our blog in the coming weeks.
Finally, we heard the community requests to clarify the release process and cycle. This is still a work in progress to meet the needs of all members of the community, but rest assured you will always be able to access the latest release in the Tungsten Fabric Docker Hub.
Do you have suggestions for us?
Join us in our community via regular meetings, Slack, or our mailing lists and bring your feedback. We want to hear from developers and operators alike, and we welcome your contributions!
