Andromeda: performance, isolation, and velocity at scale in cloud network virtualization Dalton et al., NSDI’18
Yesterday we took a look at the Microsoft Azure networking stack, today it’s the turn of the Google Cloud Platform. (It’s a very handy coincidence to have two such experience and system design report papers appearing side by side so that we can compare). Andromeda has similar design goals to AccelNet: performance close to hardware, serviceability, and the flexibility and velocity of a software-based architecture. The Google team solve those challenges in a very different way though, being prepared to make use of host cores (which you’ll recall the Azure team wanted to avoid).
We opted for a high-performance software-based architecture instead of a hardware-only solution like SR-IOV because software enables flexible, high-velocity feature deployment… Andromeda consumes a few percent of the CPU and memory on-host. One physical CPU core is reserved for the Andromeda dataplane… In the future, we plan to increase the dataplane CPU reservation to two physical cores on newer hosts with faster physical NICs and more CPU cores in order to improve VM network throughput.
Both the control plane and data plane use a hierarchical structure. The control Continue reading
The new NVMe-based storage array has built-in machine learning that uses predictive analytics and pattern recognition to improve storage performance.
Thanks to all who joined us for the Gluware DemoFriday: Software Enable your Brownfield Network – Automate Multi-Vendor QoS on Routers and NAC.
At Interop ITX, SDN pioneer talks about how software-defined networking has influenced the industry and other trends.
Enterprises are balking at moving Kubernetes into production environments because of too many choices. It "makes them basically an open source aggregator."
ONAP had started to run tests for various components it was developing. But it wanted to store test results and run tests more often. Meanwhile, OPNFV had already solved that problem.
The online payments company said it is paying half of what it paid to have six sites outfitted with MPLS to deploy Cato SD-WAN on 14 sites at a higher bandwidth.
The open source Operator Framework moves closer to automating the deployment and management of Kubernetes clusters.
Groups like the ORAN Forum, Cisco’s Open vRAN initiative, and TIP’s OpenRAN Group are all working to virtualize the RAN. The Linux Foundation has kept its distance, so far.
Company unveils its Virtual Cloud Network vision, based on its NSX network virtualization platform.
Never attribute to malice that which can be adequately explained by stupidity
Laughed heartily.
SSH is everywhere in the development or operations world now. For development it’s what allows you to push to GitHub. For operations it’s what allows you to reasonably securely log into Linux servers.
SSH