I/o paravirtualization at the device file boundary

Abstract

Paravirtualization is an important I/O virtualization technology since it uniquely provides all of the following benefits: the ability to share the device between multiple VMs, support for legacy devices without virtualization hardware, and high performance. However, existing paravirtualization solutions have one main limitation: they only support one I/O device class, and would require significant engineering effort to support new device classes and features. In this paper, we present Paradice, a solution that vastly simplifies I/O paravirtualization by using a common paravirtualization boundary for various I/O device classes: Unix device files. Using this boundary, the paravirtual drivers simply act as a classagnostic indirection layer between the application and the actual device driver.We address two fundamental challenges: supporting cross-VM driver memory operations without changes to applications or device drivers and providing fault and device data isolation between guest VMs despite device driver bugs. We implement Paradice for x86, the Xen hypervisor, and the Linux and FreeBSD OSes. Our implementation paravirtualizes various GPUs, input devices, cameras, an audio device, and an Ethernet card for the netmap framework with~7700 LoC, of which only~900 are device class-specific. Our measurements show that Paradice achieves performance close to native for different devices and applications including netmap, 3D HD games, and OpenCL applications.

A Hybrid I/O Virtualization Framework for RDMA-capable Network Interfaces

“…It uses a paravirtual approach to support complete bypass of the hypervisor on data operations. Our work on HyV contributes the following: (1) memory mapping API to allow using unmodified drivers on both guest and host, (2) generic, interconnect independent, RDMA verbs forwarding API to ease development of lightweight virtual providers, (3) demonstration of a real distributed application on a HyV-enabled cluster Paradice is a I/O paravirtualization framework to virtualize devices that use a (unix-style) device file interface to applications [2]. Like HyV, Paradice virtualizes at a high-level interface ("paravirtualization boundary") to target a larger number of devices in a generic way.…”

Partial migration technique for GPGPU tasks to Prevent GPU Memory Starvation in RPC‐based GPU Virtualization

“…Full GPU virtualization provides full GPU functionality to a VM, thereby enabling the VM to perform a variety of GPU tasks including 3D image rendering and GPGPU operations. Although full GPU virtualization provides the full functionality of the GPU to the VM, the host and guest operating system (OS), hypervisor, and GPU drivers must be modified because full GPU virtualization is system‐dependent and there are a limited number of virtualization platforms and OSs that can be applied. Finally, RPC‐based GPU virtualization does not directly virtualize a GPU; instead it virtualizes the GPU programming application programming interface (API).…”