Category Archives: User Story

A user explains how they employ Xen Project software

Q&A with GlobalLogic on the Xen Project and Automotive Virtualization

The Xen Project is commonly used in embedded scenarios due to its security features, light-weight architecture and open source community. These core attributes are now making it more pervasive in the automotive industry, which has similar demands to the embedded industry, especially when it comes to security requirements.

To better understand how the Xen Project is used in the automotive space, we sat down with the folks at GlobalLogic to discuss updates on its Nautilus platform, which uses the Xen Project hypervisor; why they originally chose Xen; how hypervisors generally work in the automotive space; and the company’s upcoming plans with automotive virtualization.

Last year when we talked to GlobalLogic, you mentioned that GPU Virtualization was the next phase of automotive innovation. Where are you at in terms of implementing GPU Virtualization?

We have successfully implemented our Nautilus platform’s GPU virtualization feature for several Tier 1 automotive vendors (located in Japan, the US, and Europe). This was a big win for us and we learned a lot along the way and experienced some major benefits. Mainly, GPU virtualization has eliminated almost all performance degradation during the rendering of heavy 3D graphics scenes, allowing us to create a new level of IVI systems.

Why is the hypervisor important for automotive virtualization and GPU Virtualization in general? Why is Xen Project the hypervisor of choice for you within this space?

The hypervisor allows a significant decrease to the cost of automotive production and reduces the cost of BOM because the functions that were previously executed on different CPUs can be run on separate VMs. At the same time, GPU virtualization is beneficial in the process of 2D/3D graphics rendering. Therefore, the use of hypervisor enables building systems that perform better than their more expensive completely-hardware analogues.

Moreover, there are less processors per board, which leads to higher fail-safety. Essentially, a virtual system divided into a number of small subsystems is cheaper to maintain.

At the dawn of our project, GlobalLogic engineers considered various hypervisors, and finally decided that Xen Project was the most suitable solution because it is open source and has a rich history of application in various fields. Using the Xen Project, lets us concentrate on specific vehicle-related challenges instead of reinventing a virtualization solution.

What are the top three benefits you get from using the Xen hypervisor?

The first benefit that we have experienced is the decreased time to market for the manufacturers. Secondly, our customers get demos for free – if we used a proprietary product, we couldn’t afford this. Finally, it is great to experience the constant support of the global community and the community-driven approach to vulnerability detecting and fixing that we get with the Xen Project.

Were there any challenges with implementing Xen? How did you overcome these challenges?

The main challenges that we had with Xen and GPU virtualization was related to the different based ARM platforms. To overcome this, we developed a bench of drivers and extended the environment around them.

What are the next stages of growth for with automotive virtualization? Any trends that we should watch out for?

GlobalLogic is actively working on the commercialization of the Nautilus platform. We are expanding the GPU feature to a network of customers and vehicle models. At the same time, we are expanding the functionality of virtualization in areas like self-driving, advanced driver assistance systems (ADAS), connected services, safety, etc.

Tips and Tricks for Making VM Migration More Secure

A challenge for any cloud installation is the constant tradeoff of availability versus security. In general, the more fluid your cloud system (i.e., making virtualized resources available on demand more quickly and easily), the more your system becomes open to certain cyberattacks. This tradeoff is perhaps most acute during active virtual machine (VM) migration, when a VM is moved from one physical host to another transparently, without disruption of the VM’s operations. Live virtual machine migration is a crucial operation in the day-to-day management of modern cloud environment.

Linux.com recently published an article from John Shackleton of Adventium Labs that focuses on how to recognize and avoid common attacks with VM migration. Read the full article here.

 

The Power of Hypervisor-Based Containers

The modern trend towards cloud-native apps seems to be set to kill hypervisors with a long slow death. Paradoxically, it is the massive success of hypervisors and infrastructure-as-a-service during the last 15 years that enabled this trend.

Stefano Stabellini provides an overview of the rise of containers and how hypervisors are co-existing and thriving in the era of containers. Read more here.

Xen Project 2017 Predictions: The Growth and Changes of the Hypervisor in 2017 and More

Embedded systems become virtualized, IoT security concerns continue and the container community diversifies… What else will happen to the hypervisor and beyond in 2017? Two members of the Xen Project, Stefano Stabellini and James Bulpin, provide insight on where the hypervisor is going in 2017 and other virtualization and infrastructure trends to watch out for in this VMblog post.

CES 2015: Smart Cars are the New Smart Phone

This is a reprint of the following Linux.com article by Alex Agizim, VP, CTO Embedded Systems at GlobalLogic

“Smart car” technology had a huge presence at CES 2015, from BMW’s 360-degree collision avoidance and parking assist features to Audi’s Human Machine Interface (HMI) that connects to an iPhone or Android device. And with both Apple and Google jumping into the market with their CarPlay and Android Auto IVI systems, the automotive industry is on the brink of some significant changes.

For example, thanks to new developments in open source virtualization, OEMs and car manufacturers are closer than ever to achieving a secure, flexible, robust, and customizable integrated cockpit — one that keeps drivers safe while meeting consumers’ connected car expectations. Already well-known for providing security, stability, and isolation in the datacenter, automotive virtualization is gaining wider attention due to additional hardening and new support for ARM.

While this is certainly exciting, virtualization remains a roadblock to some in the smart car industry. I personally had the opportunity to demonstrate GlobalLogic’s Nautilus platform for automotive virtualization at GENIVI’s CES demo and networking event. Leveraging a TI J6 SoC, I demo’d a dual-screen virtual cockpit with one screen emulating a Linux-powered driver information display, and the other screen emulating an Android-powered IVI system. The entire configuration ran on Xen Project Hypervisor 4.5 with three domains: Dom0 (thin control), DomU (Linux), and DomU (Android).

During the demo, I showcased how Nautilus achieves an overall system boot time of 8 seconds, an early RVC of 1.5 seconds, and secure and reliable peripheral sharing (including GPUs). Most importantly, I demonstrated how even if the Android virtual machine crashes, it has absolutely no influence on the mission-critical Linux virtual machine. With Nautilus automotive software, developers can host a number of VMs that are completely sandboxed from each other, thereby ensuring that all vehicle services will continue to operate even if one specific component fails.


The demo was well-received by GENIVI’s attendees, and I got the impression that many Tier 1 OEMs were thinking about using virtualization in their next-gen platforms. This is a huge milestone because, up until very recently, virtualization had a bad rep in the automotive industry. Previous attempts at virtualization using ARM A9 architecture ultimately failed because there was no hardware support for it. Many were also highly reluctant to use open source technology because it lacked proper compliance to strict auto industry regulations. But with platforms like Nautilus, developers can leverage cutting-edge open source technology that is ISO 26262 certification ready to create secure and reliable automotive virtualization experiences.

In fact, GlobalLogic’s goal is to make Nautilus part of the reference Automotive Grade Linux (AGL) software, an open source project that is developing a common, Linux-based software stack for the connected car. We are also a founding leader for Xen Project’s Embedded and Automotive initiative. GlobalLogic is working to add the Xen-based technology to the AGL spec and is further developing the platform’s real-time scheduling and peripheral sharing features to improve the use of a single physical CPU for multiple guest OSes and peripheral devices. We’ll soon be extending QNX and Tizen IVI 3.0 support to improve the functionality of other features. Finally, we are also expanding Nautilus to support even more SoCs in the next six months, such as Renesas R-Car H2/M2, which offers hardware support for virtualization.

Based on my work with the Nautilus platform and my observations of the general automotive industry, I wouldn’t be surprised to see the first PoCs for automotive virtualization coming out of China and Japan later this year. The momentum behind smart car technology development is very strong right now, and I’m excited to see what happens when automotive OEMs finally start taking advantage of virtualization’s many possibilities.

Using Xen Project on OpenStack “Juno” via Libvirt

By Xing Lin

This document describes steps I took to setup a compute node based on Ubuntu 14.04 for OpenStack “juno”, using the Xen Project via libvirt approach. Openstack does not support this approach well as it is in Group C of the hypervisor support matrix for Openstack. You can hardly find any tutorial online describing this approach and this might be the first. Let’s get started!

Prerequisites

Follow “OpenStack Installation Guide for Ubuntu 14.04″ to setup the control node and network node, following the three-node architecture with OpenStack Networking (neutron). This involves lots of configuration and could take a day or two. Check that the control node and network node is working.

Continue reading