Networking and Virtualization Technologies
The Networking and Virtualization Technologies group is focused on two major areas— network virtualization and server virtualization.
The Network Virtualization team specializes in software define networks and data center networking. This team innovates, architects, and builds network virtualization solutions for large-scale highly virtualized multi-tenant environments. The Virtual Application Network (VAN), for example, contributed to the 2009 EU project RESERVOIR.
Lately, another network virtualization emerging solution known as DOVE (Distributed Overlay Virtual Ethernet) is expanding to become the major strategic SDN product offering by IBM Systems Networking. The team is currently engaged in number of activities: to advance the DOVE technology, doing further research in the area of software defined networking (SDN), and integrating into the cloud management stack (OpenStack). In addition, the team is actively engaged in partnerships and PoC initiatives, aiming to establish and complement the technology and to contribute to a complete orchestrated IT solution for data centers and clouds.
The Server Virtualization team specializes in operating system and hypervisor research and development, with a strong focus on I/O technologies. The team is currently engaged in a number of projects in areas such as I/O performance for virtual machines and nested virtualization in the Linux Kernel-based Virtual Machine (KVM) hypervisor. KVM hypervisor has gained a lot of attention in recent years and is an attractive open source hypervisor that is included in all Linux distributions. The team is also engaged in hypervisor research and development, with KVM as our platform of choice.
One of the focus areas of our Server Virtualization team is efficient I/O virtualization. During 2008, the team worked with the KVM community on direct assignment of PCI devices to virtual machines. We developed ELI (ExitLess interrupts) in collaboration with the Technion – Israel Institute of Technology. ELI is a novel software-only approach for handling interrupts within guest virtual machines directly and securely. Despite the proven performance advantage of device assignment and ELI, in many use cases paravirtual I/O is preferred or even required. Thus, we developed ELVIS (Exitless Virtual I/O System). ELVIS presents an efficient paravirtual I/O model, under which guests and the hypervisor exchange zero-overhead notifications while running on distinct cores.
The Server Virtualization team also focuses on nested virtualization. In classical machine virtualization, a hypervisor runs multiple operating systems simultaneously, each on its own virtual machine. In nested virtualization, a hypervisor can run multiple hypervisors with their associated virtual machines. The IBM Turtles Nested Virtualization project, which is part of the Linux/KVM hypervisor, runs multiple unmodified hypervisors, such as KVM and VMware, with various operating systems, including Linux and Windows.
Yet another focus area of the server virtualization team is the intersection of multi-core architectures and I/O stacks. Data storage technology today faces many challenges, including performance inefficiencies, inadequate dependability and integrity guarantees, limited scalability, loss of confidentiality, poor resource sharing, and increased ownership and management costs.
We participate in the EU project IOLanes, which aims to analyze and address these challenges throughout the I/O path. Our approach breaks down the I/O stack into four important layers: (a) application and middleware, (b) virtual machine, (c) host operating system, and (d) embedded storage controller. The proposed work analyzes and addresses the inefficiencies associated with these layers on multi-core CPUs, by designing an I/O stack that minimizes unnecessary overheads and scales with the number of cores. To solve part of the inefficiencies, we propose SplitX, an alternative virtualization model for multi-core systems. SplitX proposes architectural changes to the x86 architecture to support efficient simultaneous execution of guests and the hypervisor on dedicated CPU cores.
This is an ambitious EU FP7 project, in which we lead a consortium of 13 European companies and universities who collaborate on developing an advanced infrastructure for cloud computing, based on deep integration of virtualization and grid technologies.
This is an FP7 project that pursues advancements in the scalability and performance of I/O subsystems in multicore platforms. In this project, we work on hypervisor design and implementation, focusing on high-performance I/O in virtualized environments.