Another month, another monthly overview of the new and changed flings as published on https://flings.vmware.com/. There are three new flings and seven received an update.
Linux Driver for Precision Clock Virtual Device
The Linux Driver for Precision Clock Virtual Device fling is an alternative way for synchronising time using a newly introduced piece of virtual hardware in vSphere 7.
ptp_vmw is a Linux driver for VMware Precision Clock, a new type of virtual device available in ESXi 7.0 (hardware version 17 on-wards) that provides virtual machines with access to the underlying ESXi host’s system clock. Guests can use the device as a reference clock in Chrony time synchronization software to synchronize their system clocks with.
Precision Clock offers an alternative to existing methods of time synchronization in the guest, such as NTP. A potential benefit of using Precision Clock, when compared to a network time synchronization, is that it uses a VMware proprietary paravirtual interface between a virtual machine and the host to fetch time information. Achievable accuracy using network based time synchronization in a guest is limited by delay and variability in the virtual networking paths (including the guest’s own networking stack), especially under high loads. By avoiding virtual networking, time synchronization using Precision Clock can, potentially, achieve higher accuracy. See the vSphere 7.0 documentation at https://docs.vmware.com for more information about this virtual device.
This fling includes a Linux kernel module source RPM, which can be built and installed in a Linux system. Upon loading the driver, a PTP clock device is created, which can be consumed as a reference clock in Chrony. See included README file for more information.
Demo Appliance for Tanzu Kubernetes Grid
A Virtual Appliance that pre-bundles all required dependencies to help customers in learning and deploying standalone Tanzu Kubernetes Grid (TKG) clusters running on either VMware Cloud on AWS and/or vSphere 6.7 Update 3 environment for Proof of Concept, Demo and Dev/Test purposes.
This appliance will enable you to quickly go from zero to Kubernetes in less than 30 minutes with just an SSH client and a web browser!
- Quickly deploy TKG Clusters onto VMware Cloud on AWS or vSphere-based infrastructure
- Online vSphere Content Library to sync all TKG Demo Appliance dependencies
- Accompany step-by-step workshop-style guide
- Embedded Harbor registry pre-loaded with all required TKG and Demo Containers
- Support for Air-Gapped and Non-Internet accessible environments
- Sample demo applications including Persistent Volume, K8s 3-Tier Application with a LoadBalancer example
- Easily access and debug TKG Clusters using Octant
Supernova – Accelerating Machine Learning Inference
With machine learning is widely used in enterprises, big data are trained on the edge, inference services go to production either in the cloud or on the edge.
On the edge
- Edge devices have limited resources, space and power supply
- Edge servers cost much higher than devices
- Hardware accelerators are heterogeneous in architecture and various on interfaces and performance on the edge
In the cloud
- Accelerator market is dominated by Nvidia GPU
- Other options come as AMD GPU, Intel Habana Goya/Altera FPGA, AWS Inferentia, Xilinx FPGA etc
- Common inference interfaces from cloud to edge doesn’t appear generally
- Limitation on specific hardware accelerators or cloud leads to new vendor lock-in
Project Supernova is to build a common machine learning inference service framework by enabling machine learning inference accelerators across edge endpoint devices, edge systems and cloud, with or without hardware accelerators.
- Micro-service based architecture with Restful API
- Support heterogenous system architectures from leading vendors
- Support accelerator compilers to native code
- Neutral to ML training framework file formats
- Work on both edge devices and clouds
- Hardware CPU support:
- x86-64, ARM64
- Hardware accelerator support:
- Intel VPU, Google Edge TPU, Nvidia GPU
- Inference toolkit support: OpenVINO, TensorRT & Tenserflow Lite
- Training framework data format: Tensorflow, Caffe, ONNX, MxNet
VMware Appliance for Folding@Home
Do you have some cpu resources left to use for a good cause? The VMware appliance for folding@home makes life doing that just a bit easier.
May 6, 2020 – v1.0.4
- F@H software has been updated to latest 7.6.13
VMware Event Broker Appliance
The VMware Event Broker Appliance gives users makes live easier for creating business logic based on events.
Virtual Machine Compute Optimizer
The Virtual Machine Compute Optimizer (VMCO) is a Powershell script that uses the PowerCLI module to capture information about the hosts and VMS running in your vSphere environment, and reports back on whether the VMs are configured optimally based on the Host CPU and memory.
- Fixed errors in reporting for some VMs that are on hosts with 4 sockets
- Fixed “memory” missing from Details when VM memory spans pNUMA nodes
- Added ability to call function with “-simple” which only reports VM info (leaves out vCenter, Cluster, and Host)
VMware Machine Learning Platform
The VMware Machine Learning Platform was build to provide an end-to-end ML Platform.
- Added support for vSphere with Kubernetes and Tanzu Kubernetes GRID in addition to VMware
- Cloud Foundation/PKS
- Upgraded to Kubeflow 1.0 GA
- Added support for GPUs
- Introduced a new data registry component called Data Manager
- Upgraded minor components/libraries to the latest versions
- Added an easy-to-use installer
- Lots of bug fixes
vSphere Mobile Client
The vSphere Mobile client is the tool to have if you want to be able to an early check on your vCenter while running to your desk to do it on those nice and fancy big screens you have over there.
- Virtual keyboard for VM console, with all special keys available
- Details page for cluster objects
- iOS devices now have the VM console, still requires direct ESXi connection for both Android and iOS
- Library updates for better compatibility
- Host no longer shows as standalone when part of a cluster
- Issues for all objects are calculated similarly, by adding together fired alarms and configuration issues
- Virtual CPU count in VM summary page is now correct
- All details pages are showing information in a similar way
Horizon Session Recording
The Horizon Helpdesk Recording fling is an underestimated fling in my eyes, it gives you the opportunity to properly record whats’s happening in a users vdi session.
Note: Version 2.0.8 is a complete re-write of the whole fling, This fling does not support upgrading, this will require a new deployment, server and agent.
- The agent is now multi-threaded.
- The web service is now written in angular.
- The web service now supports high availability (see documentation).
- This release will only work with Horizon 7.9 or higher.
Horizon Helpdesk Utility
After Reach that sadly had to be pulled the Horizon Helpdesk Utility is on of the best flings to have ever been released for Horizon. FInally a fast tool that properly helps your helpdesk without having to go to the admin console.
- Fixed an intermittent issue with the agent crashed when viewing a pool / session.