Intel, along with cloud giants Amazon and Google, is working on an open source hypervisor based on the rust-vmm project. The chipmaker discussed this and several other open source efforts at its Open Source Technology Summit, which kicked off yesterday.
The company “is and has been one of the largest contributors to open source,” said Imad Sousou, Intel corporate vice president and general manager of system software products. “Intel is the No. 1 contributor to the Linux kernel. We write 10% to 12% of the Linux kernel code.” For the record: Red Hat is No. 2, and it contributes about 6%, according to Sousou.
Intel hosts this summit every year. And if you’re wondering (as we were) why you’ve never heard of it, it’s because for the past 15 years it’s been a closed, internal event. This year Intel opened up the Open Source Technology Summit and invited customers, technology partners, and the press. It also added keynote speakers from Alibaba, Amazon, AT&T, Google, Huawei, Microsoft, and Red Hat, among other companies.
Of course, the discussions at the three-day event remain focused on open source software that is optimized for Intel hardware. But it also spanned the entire stack, targeting technologies including containers, artificial intelligence (AI), machine learning (ML), and other cloud-to-edge device workloads.
“This is where a lot of the innovation starts,” Sousou said. “The technology talked about here, you will probably see in production in about two years.”
One of these next-generation technologies is this new open source hypervisor built for cloud data centers.
“When you look at the hypervisors that power the entire cloud universe, all of it utilizes fairly old software,” Sousou said. “You have hundreds of thousands of lines of code to emulate five-inch floppy disks and all sorts of things that don’t exist anymore.”
So Intel and other industry giants got together to create a modern, configurable hypervisor.
The rust-vmm project got its start in December 2018, with Amazon, Google, Intel, and Red Hat engineers looking for the best way to share virtualization packages. The open source project provides a set of common hypervisor components developed by these and other contributors. The group said they are developing a commercial hypervisor based on rust-vmm.
Sousou said this technology will provide a more secure, higher performance container technology designed for cloud native environments. “This specific architecture that we are developing now is a base, and everybody can create their own customizable hypervisor” on top of the code base, he added.
Modernizing Open Source Code
Also in the vein of updating code and modernizing core infrastructure, Intel at the summit announced the ModernFW Initiative (FW for firmware). “That low-level software so far has been tailored to PCs and enterprise servers, and that really does not capture the need for cloud, for AI systems, and all the modern usages,” Sousou said.
This initiative has the goal to remove legacy code and modularize design, making the firmware more scalable. “What this will mean is that the BIOS firmware will be configurable, and it will be able to do start up really, really quickly and then get out of the way, which is really important when you are in a cloud environment or at the edge and you want to be able to boot in milliseconds, not several seconds,” Sousou said.
By delivering just enough code to boot the kernel, this approach also aims to reduce exposure to security risks and ease management.
Intel also committed to assigning developers to contribute code and test “critical system infrastructure” projects beyond the Linux kernel. Some of these open source projects include bash, chrony, the Fuzzing Project, GnuPG, libffi, the Linux Kernel Self Protection Project, OpenSSH, OpenSSL, and the R programming language. This is an effort to modernize these massive systems infrastructure projects so that they perform better — and, thus, the cloud and edge systems built on top of them will also run better — while also patching bugs and improving security, Sousou said.
Clear Linux OS
And finally, Intel announced new and updated open source software including:
- The Clear Linux Distribution is adding Clear Linux Developer Edition, which includes a new installer and store, and improved integration for Intel hardware features and workloads supporting the Deep Learning and Data Analytics software stacks.
- The Deep Learning Reference Stack, which is an integrated, highly-performant stack optimized for Intel Xeon Scalable processors. This stack includes Deep Learning Boost (Intel DL Boost) and is designed to accelerate AI use cases such as image recognition, object detection, speech recognition, and language translation.
- The Data Analytics Reference Stack developed to help enterprises analyze, classify, recognize, and process large amounts of data built on Xeon Scalable platforms using Apache Hadoop and Apache Spark.
- Fusion Edge Stacks support that consolidates workloads on connected devices using the ACRN device hypervisor, Clear Linux OS, Zephyr Project, and Android.
- And the Intel Robot software developer kit (SDK) that combines Intel hardware and software in one resource to simplify the process of creating AI-enabled robotics and automation systems, with an optimized computer vision stack.
“At the end of the day, we realize that software, and especially open source software, is a critical part of Intel’s growth,” Sousou said. “The $300 billion opportunity that Intel CEO Bob Swan talked about it going to be largely driven by software. This is why we are opening this [Open Source Technology Summit] up and have this heavy investment in all of these open source projects.”
Photo Credit: David B. Kinder/Intel Corporation