User Tools

Site Tools


open_vswitch_for_nfv

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
open_vswitch_for_nfv [2015/07/02 15:36]
Mark D. Gray [Committers and Contributors:]
open_vswitch_for_nfv [2016/03/04 11:41] (current)
Simona Coppola [Proposed Release Schedule:]
Line 7: Line 7:
  
 This project proposes:\\ This project proposes:\\
-  * Modifying the OPNFV build to include a deployment option for the software-accelerated userspace Open vSwitch build. This change will significantly ​increase ​the performance of the NFVI for network I/O. +  * Modifying the OPNFV build to include a deployment option for the software-accelerated userspace Open vSwitch build. This change will significantly ​improve ​the performance of the NFVI for network I/O. 
-  * Future work that will encompass collaborative development within the Open vSwitch project to increase ​the performance of the software-accelerated userspace Open vSwitch and increasing its suitability for Telco NFV deployments.+  * Future work that will encompass collaborative development within the Open vSwitch project to improve ​the performance of the software-accelerated userspace Open vSwitch and increasing its suitability for Telco NFV deployments.
  
 The current release of OPNFV (Arno) enables continuous integration,​ automated deployment and testing of components from upstream projects such as Ceph, KVM, OpenDaylight,​ OpenStack and Open vSwitch. The combination of these provides an initial build of the NFV Infrastructure (NFVI) and Virtual Infrastructure Manager (VIM) components of the ETSI NFV architecture.\\ \\ The current release of OPNFV (Arno) enables continuous integration,​ automated deployment and testing of components from upstream projects such as Ceph, KVM, OpenDaylight,​ OpenStack and Open vSwitch. The combination of these provides an initial build of the NFV Infrastructure (NFVI) and Virtual Infrastructure Manager (VIM) components of the ETSI NFV architecture.\\ \\
-Open vSwitch is an open source virtual switch providing Virtual Networking services in the NFVI. It provides network connectivity from Virtual Machines (VMs) to the physical network and between Virtual Machines. The default deployment of Open vSwitch in OPNFV implements the software fast path as a Linux Kernel Module. This kernel module caches recently used flows in order to quickly forward subsequent packets in the same flow space. The kernel datapath, however, is not necessarily designed or targeted to meet the requirements for the NFVI, and we need to make collaborative efforts toward ​enabling ​NFV features. Although Open vSwitch gives adequate performance in many cloud and enterprise use cases, when used in Telco NFV use cases, the performance may not be sufficient in terms of throughput and latency, particularly for small packet sizes.\\ \\+Open vSwitch is an open source virtual switch providing Virtual Networking services in the NFVI. It provides network connectivity from Virtual Machines (VMs) to the physical network and between Virtual Machines. The default deployment of Open vSwitch in OPNFV implements the software fast path as a Linux Kernel Module. This kernel module caches recently used flows in order to quickly forward subsequent packets in the same flow space. The kernel datapath, however, is not necessarily designed or targeted to meet the requirements for the NFVI, and we need to make collaborative efforts toward ​realizing ​NFV requirements. Although Open vSwitch gives adequate performance in many cloud and enterprise use cases, when used in Telco NFV use cases, the performance may not be sufficient in terms of throughput and latency, particularly for small packet sizes.\\ \\
 Technologies such as DPDK and ODP have shown that running packet processing workloads in userspace software can improve the performance in terms of throughput and latency. The advantage of this is apparent in recent work done on the Open vSwitch userspace forwarding plane which has shown 10x improvement in packet throughput with respect to the Open vSwitch kernel space dataplane. This is demonstrated via the OPNFV vsperf project. \\ \\ Technologies such as DPDK and ODP have shown that running packet processing workloads in userspace software can improve the performance in terms of throughput and latency. The advantage of this is apparent in recent work done on the Open vSwitch userspace forwarding plane which has shown 10x improvement in packet throughput with respect to the Open vSwitch kernel space dataplane. This is demonstrated via the OPNFV vsperf project. \\ \\
  
Line 67: Line 67:
   * Joseph Gasparakis <​joseph.gasparakis@intel.com>​   * Joseph Gasparakis <​joseph.gasparakis@intel.com>​
   * Billy O Mahony <​billy.o.mahony@intel.com>​   * Billy O Mahony <​billy.o.mahony@intel.com>​
 +  * Hongbo Tianhongbo <​hongbo.tianhongbo@huawei.com>​
 +  * Tom Herbert <​therbert@redhat.com>​
 __Names and affiliations of any other contributors__ __Names and affiliations of any other contributors__
   * Mike Lynch <​Michael.a.lynch@intel.com>​   * Mike Lynch <​Michael.a.lynch@intel.com>​
   * Robin Giller <​robin.giller@intel.com>​   * Robin Giller <​robin.giller@intel.com>​
- 
 ==== Planned deliverables ==== ==== Planned deliverables ====
  
Line 81: Line 82:
   * Target intersection of integration of the software-accelerated userspace datapath with OPNFV B Release. ​   * Target intersection of integration of the software-accelerated userspace datapath with OPNFV B Release. ​
   * Future development work to improve the performance of the software-accelerated userspace datapath will be targeted for later releases.   * Future development work to improve the performance of the software-accelerated userspace datapath will be targeted for later releases.
 +
 +==== Release C Planning: ====
 +
 +  * [[Release C_Planning:​Priority list|Release C Project proposal List]]
  
 ---- ----
open_vswitch_for_nfv.1435851395.txt.gz · Last modified: 2015/07/02 15:36 by Mark D. Gray