NSH_SFC Performance Report FD.io NSH_SFC and CSIT Team

Slides:



Advertisements
Similar presentations
Accelerating Network Intensive Workloads Using the DPDK netdev
Advertisements

SUPERVISOR 2 MODULE Quad Core – 2.12 GHz NX-OS software version 6.1 (or better) Required 12 GB of memory (NVRAM – 2 MB battery backup) CMP not supported.
Diagnostics. Module Objectives By the end of this module participants will be able to: Use diagnostic commands to troubleshoot and monitor performance.
Processor history / DX/SX SX/DX Pentium 1997 Pentium MMX
5.3 HS23 Blade Server. The HS23 blade server is a dual CPU socket blade running Intel´s new Xeon® processor, the E5-2600, and is the first IBM BladeCenter.
Cisco PIX 515E Firewall. Overview What a PIX Firewall can do Adaptive Security Algorithm Address Translation Cut-Through Proxy Access Control Network.
Jamel Callands Austin Chaet Carson Gallimore.  Downloading  Recommended Specifications  Features  Reporting and Monitoring  Questions.
Windows 7 Firewall.
MIDeA :A Multi-Parallel Instrusion Detection Architecture Author: Giorgos Vasiliadis, Michalis Polychronakis,Sotiris Ioannidis Publisher: CCS’11, October.
Hardware Trends. Contents Memory Hard Disks Processors Network Accessories Future.
DNS Dynamic Update Performance Study The Purpose Dynamic update and XFR is key approach to perform zone data replication and synchronization,
Srihari Makineni & Ravi Iyer Communications Technology Lab
Hyper Threading Technology. Introduction Hyper-threading is a technology developed by Intel Corporation for it’s Xeon processors with a 533 MHz system.
Module 1: Configuring Routing by Using Routing and Remote Access.
Hyper-V Performance, Scale & Architecture Changes Benjamin Armstrong Senior Program Manager Lead Microsoft Corporation VIR413.
Intel Research & Development ETA: Experience with an IA processor as a Packet Processing Engine HP Labs Computer Systems Colloquium August 2003 Greg Regnier.
Authors: Danhua Guo 、 Guangdeng Liao 、 Laxmi N. Bhuyan 、 Bin Liu 、 Jianxun Jason Ding Conf. : The 4th ACM/IEEE Symposium on Architectures for Networking.
Exploiting Task-level Concurrency in a Programmable Network Interface June 11, 2003 Hyong-youb Kim, Vijay S. Pai, and Scott Rixner Rice Computer Architecture.
Fd.io is the future Ed Warnicke fd.io Foundation1.
Dave Ward Faster Dave Ward fd.io Foundation.
Considerations for Benchmarking Virtual Networks Samuel Kommu, Jacob Rapp, Ben Basler,
August 2016 Tommy Long, Sergio Gonzalez Monroy VPP Cryptodev 1.
SDN Controller/ Orchestration/ FastDataStacks Joel Halpern (Ericsson) Frank Brockners (Cisco)
Status from Martin Josefsson Hardware level performance of Status from Martin Josefsson Hardware level performance of by Jesper Dangaard Brouer.
Open-source routing at 10Gb/s Olof Hagsand (KTH) Robert Olsson (Uppsala U) Bengt Görden (KTH) SNCNW May 2009 Project grants: Internetstiftelsen (IIS) Equipment:
TIZEN STUDIO INSTALLATION & ENVIRONMENT SETUP FOR DEVLAB
NFP: Enabling Network Function Parallelism in NFV
InterVLAN Routing 1. InterVLAN Routing 2. Multilayer Switching.
Only Use FD.io VPP to Achieve high performance service function chaining Yi Intel.
Network Load Balancing Addressing
Co-located Summit
This document contains information on a pre-launch desktop that is under NDA and is not yet available. Expected launch is: January 20, 2017.
/csit CSIT Readout to FD.io Board 08 February 2017
Ready-to-Deploy Service Function Chaining for Mobile Networks
New Approach to OVS Datapath Performance
TLDK overview Konstantin Ananyev 05/08/2016.
TYBIS IP-Matrix Virtualized Total Video Surveillance System Edge Technology, World Best Server Virtualization.
/csit CSIT Readout to FD.io Board 09 February 2017
Patrick Lu, Intel DCG/NPG/ASE
OpenStack’s networking-vpp
GPUNFV: a GPU-Accelerated NFV System
ODL SFC and VPP Integration
VPP overview Shwetha Bhandari
ODL SFC, Implementing IETF SFC November 14, 2016
6WIND MWC IPsec Demo Scalable Virtual IPsec Aggregation with DPDK for Road Warriors and Branch Offices Changed original subtitle. Original subtitle:
Need for high performance Data Plane
Chapter 6: Network Layer
Towards 10Gb/s open-source routing
Report from Netconf 2009 Jesper Dangaard Brouer
Multi-PCIe socket network device
Tetsuya Murakami IP Infusion, CTO
Networking overview Sujata
NFP: Enabling Network Function Parallelism in NFV
Patrick Lu, Intel DCG/NPG/ASE
Get the best out of VPP and inter-VM communications.
GEN: A GPU-Accelerated Elastic Framework for NFV
NFP: Enabling Network Function Parallelism in NFV
Open vSwitch HW offload over DPDK
Sangfor Cloud Security Pool, The First-ever NSH Use Case
Multiple vDPI Functions using DPDK and Hyperscan on OVS-DPDK Platform
Integrating OpenStack with DPDK for High Performance Applications
Enabling TSO in OvS-DPDK
Encrypting OVN tunnels with IPsec
Network Processors for a 1 MHz Trigger-DAQ System
A Scalable Approach to Virtual Switching
IP Control Gateway (IPCG)
NetCloud Hong Kong 2017/12/11 NetCloud Hong Kong 2017/12/11 PA-Flow:
Fabric metrics Qi Zhang(IBM Research – T.J. Watson) Hardware Software
A Closer Look at NFV Execution Models
Openstack Summit November 2017
Presentation transcript:

NSH_SFC 17.01 Performance Report FD.io NSH_SFC and CSIT Team INTEL CONFIDENTIAL

Background Measure NSH_SFC 17.01 performance with VPP 17.01 for different SFC ingredients Service Function Forwarder NSH Proxy NSH Classifier IXIA based PacketGen to be replaced by TREX in the automation test by CSIT team Baseline performance will be used to analyze potential negative performance impact introduced by software PacketGen or future patches As baseline for apple-to-apple performance comparison with DPDK OVS, once NSH patches to OVS are upstreamed Focus on single core performance currently, scale-up (assign more cores) testing will be covered in the future

Test Configuration IXIA Traffic Configuration IXIA 10G Niantic 2.2GHz Core 2 Core 3 Core 0 Core 1 Port B Port 1 DIMM 0 DIMM 1 Port A Port 0 DIMM 2 DIMM 3 IXIA Traffic Configuration NSH Classifier: Raw packet in different size SFF: VxLAN-GPE + NSH + Raw Packet NSH Proxy Inbound: VxLAN-GPE + NSH + Raw Packet NSH Proxy Outbound: VxLAN+ Raw Packet L4 packet inner IP header Inner MAC header UDP header Outer IP header Outer MAC header NSH VxLAN-GPE Header L4 packet inner Inner MAC header UDP header Outer IP header Outer MAC header NSH VxLAN-GPE Header L4 packet inner Inner MAC header UDP header Outer IP header Outer MAC header VxLAN L4 packet inner Inner MAC header

Device Under Test Hardware Configuration BIOS Configuration CPU Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz DIMM 2133 MHz, 64GB Total NIC 2x 82599ES 10-Gigabit SFI/SFP+ Network Connection PacketGen Ixia* 10 Gigabit Ethernet Traffic Generator (16 ports) Enhanced Intel Speedstep Disabled Turbo Boost Processor C3 Processor C6 Hyper-Threading Intel VT-d Enabled CPU Power and Performance Policy Performance Memory Freq. 2133 MHz Total Memory Size 64 GB Memory RAS and Performance Configuration -> NUMA Optimized ENABLED QPI B/W 9.6 GT/s MLC Streamer MLC Spatial Prefetcher DCU Data Prefetcher DCU Instruction Prefetcher Direct Cache Access (DCA) Software Configuration OS Ubuntu 14.04.1 Kernel version 4.2.0-27-generic VPP vpp v1.0.0-263~gd6f5b96 DPDK DPDK 16.11 NSH_SFC 17.01 RC2

NSH_SFC Performance in 17.01 Release NSH_SFC 17.01 performance drops from 5.8Mpps to 4.25Mpps for NSH Proxy 2 to 3 dedicated cores could achieve 10G line rate for small packet for all NSH_SFC use cases

Backup

NSH Proxy Performance in 16.09 Release - As Reference Data nsh-proxy-decap for 128B CPP Time (ns) Mpps 394.9 171.7 5.83