Topic: A Deep Dive into Memory Access. Company: Intel Title: Software Engineer Name: Wang, Zhihong

Size: px
Start display at page:

Download "Topic: A Deep Dive into Memory Access. Company: Intel Title: Software Engineer Name: Wang, Zhihong"

Transcription

1 Topic: A Deep Dive into Memory Access Company: Intel Title: Software Engineer Name: Wang, Zhihong

2 A Typical NFV Scenario: PVP Guest Forwarding Engine virtio vhost Forwarding Engine NIC Ring ops What s actually going on? memcpy DDIO Guest virtio virtio RX TX Shared Memory vhost TX NIC RX vhost RX NIC TX Ring ops memcpy DDIO

3 Overview of Memory System 0 1 N-1 N N+1 CPU 0 CPU 1 2N-1 MESIF protocol Memory 0 Memory 1

4 Overview of Memory System (cont d) AVX 1 to maximize bandwidth 0 Load 2 1 N-1 N N+1 CPU 0 Haswell cache parameters 3 CPU 1 Line Fastest Cache level size latency 5 (Bytes) (Cycle) 4 6 L1D L ~34 Varies Capacity Peak bandwidth (KB) (Bytes/cycle) Memory L2 and L1D 0 in Memory 1 other cores 64 2N-1 64 (Load) + 32 (Store)

5 Let s Do It! Guest Forwarding Engine virtio vhost Forwarding Engine NIC Target for our analysis Ring ops Where the data flows memcpy DDIO Guest virtio virtio RX TX Shared Memory vhost TX NIC RX vhost RX NIC TX Ring ops memcpy DDIO

6 First Impression Guest -> Guest -> Guest N-1 N N+1 CPU 0 CPU 1 2N-1 FWD RX from NIC TX to vhost RX from vhost TX to NIC Memory 0 Memory 1 VM FWD RX from virtio TX to virtio

7 Unexpectedly Guest -> -> L1 Cross-core copies?? Guest CPU 0 -> Guest N-1 First try Notice CPU cycle measurement disturbs overall performance Memory 0

8 Under The Hood Guest -> Guest updates ring only, doesn t touch the data Guest N-1 N N+1 CPU 0 -> Guest CPU 1 2N-1 Data locality in cache: Who operates the data Memory 0 Memory 1

9 Guest Read The Packet Guest -> Guest R N-1 N N+1 CPU 0 CPU 1 -> Guest FWD RX from NIC TX to vhost RX from vhost TX to NIC Memory 0 Memory 1 2N-1 VM FWD RX from virtio Read packet TX to virtio

10 Still Doesn t Feel Right Guest -> No change?? Guest R CPU 0 N-1 -> Guest Memory 0 Guest read packet

11 Under The Hood Guest -> Cache line can be shared when no modification Guest R N-1 N N+1 CPU 0 CPU 1 -> Guest FWD RX from NIC TX to vhost RX from vhost TX to NIC Memory 0 Memory 1 2N-1 VM FWD RX from virtio Read packet TX to virtio

12 Guest Edit The Packet Guest -> Guest M N-1 N N+1 CPU 0 CPU 1 -> Guest FWD RX from NIC TX to vhost Memory 0 RX from vhost Memory 1 TX to NIC 2N-1 VM FWD RX from virtio Edit packet TX to virtio

13 Write-back Guest -> No change? Cross-core copies Guest M CPU 0 N-1 -> Guest Memory 0 Guest edit packet

14 Go See Some C Code desc_addr = gpa_to_vva(dev, desc->addr); rte_prefetch0((void *)(uintptr_t)desc_addr); Oh I see S/W Prefetching to reduce latency rte_memcpy(rte_pktmbuf_mtod_offset(cur, void *, mbuf_offset), (void *)((uintptr_t)(desc_addr + desc_offset)), cpy_len);

15 Without S/W Prefetching Guest -> Now I understand Bring it back right away! Guest M CPU 0 N-1 -> Guest Memory 0 Guest edit packet; No prefetching

16 How About Guest In Another Guest -> Node? Guest 1 N-1 M N+1 CPU 0 CPU 1 -> Guest FWD RX from NIC TX to vhost Memory 0 RX from vhost Memory 1 TX to NIC 2N-1 VM FWD RX from virtio Edit packet TX to virtio

17 Better NOT Keep related processes on the same node Guest edit packet; No prefetching

18 rte_memcpy()? Why Even Bother? Warm copy DPDK s scenario AVX load/store Alignment handling Guest edit packet; Guest on the same node

19 AVX For Bandwidth xmm0 = _mm_loadu_si128(src); _mm_storeu_si128(dst, xmm0); ymm0 = _mm256_loadu_si256(src); _mm256_storeu_si256(dst, ymm0); 2x peak bandwidth + 40% + 53% AVX512 is coming Guest read packet; Guest on the same node

20 Alignment Matters rte_memcpy((void *)((uint8_t *)dst + 1), src, len - 1); rte_memcpy(dst, src, len); Just like coupons FREE gifts if you use them + 17% + 15% Guest read packet; Guest on the same node

21 Takeaways See actual memory behaviors under the hood Intel 64 and IA-32 Architectures Optimization Reference Manual Benefit from new IA technologies AVX, DDIO

22

23 Cache Allocation Technology 0 1 N-1 N N+1 CPU 0 CPU 1 2N-1 CMT + CAT Noisy neighbor: One Memory core is 0 requesting huge amount of data Memory 1 What if another HIGH priority core is very latency sensitive?

RDMA-like VirtIO Network Device for Palacios Virtual Machines

RDMA-like VirtIO Network Device for Palacios Virtual Machines RDMA-like VirtIO Network Device for Palacios Virtual Machines Kevin Pedretti UNM ID: 101511969 CS-591 Special Topics in Virtualization May 10, 2012 Abstract This project developed an RDMA-like VirtIO network

More information

Achieve Low Latency NFV with Openstack*

Achieve Low Latency NFV with Openstack* Achieve Low Latency NFV with Openstack* Yunhong Jiang Yunhong.Jiang@intel.com *Other names and brands may be claimed as the property of others. Agenda NFV and network latency Why network latency on NFV

More information

KVM as The NFV Hypervisor

KVM as The NFV Hypervisor KVM as The NFV Hypervisor Jun Nakajima Contributors: Mesut Ergin, Yunhong Jiang, Krishna Murthy, James Tsai, Wei Wang, Huawei Xie, Yang Zhang 1 Legal Disclaimer INFORMATION IN THIS DOCUMENT IS PROVIDED

More information

Understanding The Performance of DPDK as a Computer Architect

Understanding The Performance of DPDK as a Computer Architect Understanding The Performance of DPDK as a Computer Architect XIAOBAN WU *, PEILONG LI *, YAN LUO *, LIANG- MIN (LARRY) WANG +, MARC PEPIN +, AND JOHN MORGAN + * UNIVERSITY OF MASSACHUSETTS LOWELL + INTEL

More information

Virtio/vhost status update

Virtio/vhost status update Virtio/vhost status update Yuanhan Liu Aug 2016 outline Performance Multiple Queue Vhost TSO Functionality/Stability Live migration Reconnect Vhost PMD Todo Vhost-pci Vhost Tx

More information

vswitch Acceleration with Hardware Offloading CHEN ZHIHUI JUNE 2018

vswitch Acceleration with Hardware Offloading CHEN ZHIHUI JUNE 2018 x vswitch Acceleration with Hardware Offloading CHEN ZHIHUI JUNE 2018 Current Network Solution for Virtualization Control Plane Control Plane virtio virtio user space PF VF2 user space TAP1 SW Datapath

More information

The dark powers on Intel processor boards

The dark powers on Intel processor boards The dark powers on Intel processor boards Processing Resources (3U VPX) Boards with Multicore CPUs: Up to 16 cores using Intel Xeon D-1577 on TR C4x/msd Boards with 4-Core CPUs and Multiple Graphical Execution

More information

Next Gen Virtual Switch. CloudNetEngine Founder & CTO Jun Xiao

Next Gen Virtual Switch. CloudNetEngine Founder & CTO Jun Xiao Next Gen Virtual Switch CloudNetEngine Founder & CTO Jun Xiao Agenda Thoughts on next generation virtual switch Technical deep dive on CloudNetEngine virtual switch Q & A 2 Major vswitches categorized

More information

New Approach to OVS Datapath Performance. Founder of CloudNetEngine Jun Xiao

New Approach to OVS Datapath Performance. Founder of CloudNetEngine Jun Xiao New Approach to OVS Datapath Performance Founder of CloudNetEngine Jun Xiao Agenda VM virtual network datapath evolvement Technical deep dive on a new OVS datapath Performance comparisons Q & A 2 VM virtual

More information

The.pdf version of this slide deck will have missing info, due to use of animations. The original.pptx deck is available here:

The.pdf version of this slide deck will have missing info, due to use of animations. The original.pptx deck is available here: The.pdf version of this slide deck will have missing info, due to use of animations. The original.pptx deck is available here: https://wiki.opnfv.org/download/attachments/10293193/vsperf-dataplane-perf-cap-bench.pptx?api=v2

More information

A Comparison of Performance and Accuracy of Measurement Algorithms in Software

A Comparison of Performance and Accuracy of Measurement Algorithms in Software A Comparison of Performance and Accuracy of Measurement Algorithms in Software Omid Alipourfard, Masoud Moshref 1, Yang Zhou 2, Tong Yang 2, Minlan Yu 3 Yale University, Barefoot Networks 1, Peking University

More information

Fast packet processing in the cloud. Dániel Géhberger Ericsson Research

Fast packet processing in the cloud. Dániel Géhberger Ericsson Research Fast packet processing in the cloud Dániel Géhberger Ericsson Research Outline Motivation Service chains Hardware related topics, acceleration Virtualization basics Software performance and acceleration

More information

DPDK Vhost/Virtio Performance Report Release 17.08

DPDK Vhost/Virtio Performance Report Release 17.08 DPDK Vhost/Virtio Performance Report Test Date: August 15 th 2017 Author: Intel DPDK Validation team Revision History Date Revision Comment August 15 th, 2017 1.0 Initial document for release 2 Contents

More information

OpenFlow Software Switch & Intel DPDK. performance analysis

OpenFlow Software Switch & Intel DPDK. performance analysis OpenFlow Software Switch & Intel DPDK performance analysis Agenda Background Intel DPDK OpenFlow 1.3 implementation sketch Prototype design and setup Results Future work, optimization ideas OF 1.3 prototype

More information

DPDK Vhost/Virtio Performance Report Release 18.05

DPDK Vhost/Virtio Performance Report Release 18.05 DPDK Vhost/Virtio Performance Report Test Date: Jun 1 2018 Author: Intel DPDK Validation Team Revision History Date Revision Comment Jun 1st, 2018 1.0 Initial document for release 2 Release 18.02 Contents

More information

Accelerating VM networking through XDP. Jason Wang Red Hat

Accelerating VM networking through XDP. Jason Wang Red Hat Accelerating VM networking through XDP Jason Wang Red Hat Agenda Kernel VS userspace Introduction to XDP XDP for VM Use cases Benchmark and TODO Q&A Kernel Networking datapath TAP A driver to transmit

More information

Submission instructions (read carefully): SS17 / Assignment 4 Instructor: Markus Püschel. ETH Zurich

Submission instructions (read carefully): SS17 / Assignment 4 Instructor: Markus Püschel. ETH Zurich 263-2300-00: How To Write Fast Numerical Code Assignment 4: 120 points Due Date: Th, April 13th, 17:00 http://www.inf.ethz.ch/personal/markusp/teaching/263-2300-eth-spring17/course.html Questions: fastcode@lists.inf.ethz.ch

More information

Xen Network I/O Performance Analysis and Opportunities for Improvement

Xen Network I/O Performance Analysis and Opportunities for Improvement Xen Network I/O Performance Analysis and Opportunities for Improvement J. Renato Santos G. (John) Janakiraman Yoshio Turner HP Labs Xen Summit April 17-18, 27 23 Hewlett-Packard Development Company, L.P.

More information

A Userspace Packet Switch for Virtual Machines

A Userspace Packet Switch for Virtual Machines SHRINKING THE HYPERVISOR ONE SUBSYSTEM AT A TIME A Userspace Packet Switch for Virtual Machines Julian Stecklina OS Group, TU Dresden jsteckli@os.inf.tu-dresden.de VEE 2014, Salt Lake City 1 Motivation

More information

Quo Vadis Virtio? Michael S. Tsirkin Red Hat

Quo Vadis Virtio? Michael S. Tsirkin Red Hat Quo Vadis Virtio? 26 Michael S. Tsirkin Red Hat Uses material from https://lwn.net/kernel/ldd3/ Gcompris, tuxpaint, childplay Distributed under the Creative commons license, except logos which are C/TM

More information

DPDK Vhost/Virtio Performance Report Release 18.11

DPDK Vhost/Virtio Performance Report Release 18.11 DPDK Vhost/Virtio Performance Report Test Date: December 3st 2018 Author: Intel DPDK Validation Team Revision History Date Revision Comment December 3st, 2018 1.0 Initial document for release 2 Contents

More information

Design of Vhost-pci - designing a new virtio device for inter-vm communication

Design of Vhost-pci - designing a new virtio device for inter-vm communication Design of Vhost-pci - designing a new virtio device for inter-vm communication Wei Wang wei.w.wang@intel.com Contributors: Jun Nakajima, Mesut Ergin, James Tsai, Guangrong Xiao, Mallesh Koujalagi, Huawei

More information

Enhancing SSD Control of NVMe Devices for Hyperscale Applications. Luca Bert - Seagate Chris Petersen - Facebook

Enhancing SSD Control of NVMe Devices for Hyperscale Applications. Luca Bert - Seagate Chris Petersen - Facebook Enhancing SSD Control of NVMe Devices for Hyperscale Applications Luca Bert - Seagate Chris Petersen - Facebook Agenda Introduction & overview (Luca) Problem statement & proposed solution (Chris) SSD implication

More information

Great Reality #2: You ve Got to Know Assembly Does not generate random values Arithmetic operations have important mathematical properties

Great Reality #2: You ve Got to Know Assembly Does not generate random values Arithmetic operations have important mathematical properties Overview Course Overview Course theme Five realities Computer Systems 1 2 Course Theme: Abstraction Is Good But Don t Forget Reality Most CS courses emphasize abstraction Abstract data types Asymptotic

More information

Lecture: Large Caches, Virtual Memory. Topics: cache innovations (Sections 2.4, B.4, B.5)

Lecture: Large Caches, Virtual Memory. Topics: cache innovations (Sections 2.4, B.4, B.5) Lecture: Large Caches, Virtual Memory Topics: cache innovations (Sections 2.4, B.4, B.5) 1 Techniques to Reduce Cache Misses Victim caches Better replacement policies pseudo-lru, NRU Prefetching, cache

More information

KVM PERFORMANCE OPTIMIZATIONS INTERNALS. Rik van Riel Sr Software Engineer, Red Hat Inc. Thu May

KVM PERFORMANCE OPTIMIZATIONS INTERNALS. Rik van Riel Sr Software Engineer, Red Hat Inc. Thu May KVM PERFORMANCE OPTIMIZATIONS INTERNALS Rik van Riel Sr Software Engineer, Red Hat Inc. Thu May 5 2011 KVM performance optimizations What is virtualization performance? Optimizations in RHEL 6.0 Selected

More information

Cache Memories. From Bryant and O Hallaron, Computer Systems. A Programmer s Perspective. Chapter 6.

Cache Memories. From Bryant and O Hallaron, Computer Systems. A Programmer s Perspective. Chapter 6. Cache Memories From Bryant and O Hallaron, Computer Systems. A Programmer s Perspective. Chapter 6. Today Cache memory organization and operation Performance impact of caches The memory mountain Rearranging

More information

Tungsten Fabric Optimization by DPDK ZHAOYAN CHEN YIPENG WANG

Tungsten Fabric Optimization by DPDK ZHAOYAN CHEN YIPENG WANG x Tungsten Fabric Optimization by DPDK ZHAOYAN CHEN YIPENG WANG Agenda Introduce Tungsten Fabric Support More CPU cores MPLS over GRE Optimization Hash Table Optimization Batch RX for VM and Fabric What

More information

Changpeng Liu. Cloud Storage Software Engineer. Intel Data Center Group

Changpeng Liu. Cloud Storage Software Engineer. Intel Data Center Group Changpeng Liu Cloud Storage Software Engineer Intel Data Center Group Notices & Disclaimers Intel technologies features and benefits depend on system configuration and may require enabled hardware, software

More information

G-NET: Effective GPU Sharing In NFV Systems

G-NET: Effective GPU Sharing In NFV Systems G-NET: Effective Sharing In NFV Systems Kai Zhang*, Bingsheng He^, Jiayu Hu #, Zeke Wang^, Bei Hua #, Jiayi Meng #, Lishan Yang # *Fudan University ^National University of Singapore #University of Science

More information

Lecture: Large Caches, Virtual Memory. Topics: cache innovations (Sections 2.4, B.4, B.5)

Lecture: Large Caches, Virtual Memory. Topics: cache innovations (Sections 2.4, B.4, B.5) Lecture: Large Caches, Virtual Memory Topics: cache innovations (Sections 2.4, B.4, B.5) 1 More Cache Basics caches are split as instruction and data; L2 and L3 are unified The /L2 hierarchy can be inclusive,

More information

These slides do not give detailed coverage of the material. See class notes and solved problems (last page) for more information.

These slides do not give detailed coverage of the material. See class notes and solved problems (last page) for more information. 11 1 This Set 11 1 These slides do not give detailed coverage of the material. See class notes and solved problems (last page) for more information. Text covers multiple-issue machines in Chapter 4, but

More information

Introducing Cache Pseudo-Locking to reduce memory access latency. Reinette Chatre

Introducing Cache Pseudo-Locking to reduce memory access latency. Reinette Chatre Introducing Cache Pseudo-Locking to reduce memory access latency Reinette Chatre About me Software Engineer at Intel (~12 years) Open Source Technology Center (OTC) Currently Enabling Cache Pseudo-Locking

More information

DPDK Roadmap. Tim O Driscoll & Chris Wright Open Networking Summit 2017

DPDK Roadmap. Tim O Driscoll & Chris Wright Open Networking Summit 2017 DPDK Roadmap Tim O Driscoll & Chris Wright Open Networking Summit 2017 Agenda Overview: What is DPDK? What problems does it solve? Open source community and transition to Linux Foundation: Why is this

More information

DPDK Summit 2016 OpenContrail vrouter / DPDK Architecture. Raja Sivaramakrishnan, Distinguished Engineer Aniket Daptari, Sr.

DPDK Summit 2016 OpenContrail vrouter / DPDK Architecture. Raja Sivaramakrishnan, Distinguished Engineer Aniket Daptari, Sr. DPDK Summit 2016 OpenContrail vrouter / DPDK Architecture Raja Sivaramakrishnan, Distinguished Engineer Aniket Daptari, Sr. Product Manager CONTRAIL (MULTI-VENDOR) ARCHITECTURE ORCHESTRATOR Interoperates

More information

Intel New RDT Features and Implementation Introduction

Intel New RDT Features and Implementation Introduction Intel New RDT Features and Implementation Introduction Yi Sun Jun. 10 th, 2017 1 Agenda Shared Resource Contention Solution: Intel Resource Director Technology (RDT) Performance Improvement Proofs New

More information

Nova Scheduler: Optimizing, Configuring and Deploying NFV VNF's on OpenStack

Nova Scheduler: Optimizing, Configuring and Deploying NFV VNF's on OpenStack Nova Scheduler: Optimizing, Configuring and Deploying NFV VNF's on OpenStack Ian Jolliffe, Chris Friesen WHEN IT MATTERS, IT RUNS ON WIND RIVER. 2017 WIND RIVER. ALL RIGHTS RESERVED. Ian Jolliffe 2 2017

More information

Bridging OPNFV and ETSI Yardstick and the methodology for pre-deployment validation of NFV Infrastructure

Bridging OPNFV and ETSI Yardstick and the methodology for pre-deployment validation of NFV Infrastructure Bridging OPNFV and ETSI Yardstick and the methodology for pre-deployment validation of NFV Infrastructure Ana Cunha (Ericsson) ana.cunha@ericsson.com Agenda The facts The questions The ETSI-NFV methodology

More information

Data Path acceleration techniques in a NFV world

Data Path acceleration techniques in a NFV world Data Path acceleration techniques in a NFV world Mohanraj Venkatachalam, Purnendu Ghosh Abstract NFV is a revolutionary approach offering greater flexibility and scalability in the deployment of virtual

More information

DPDK Summit China 2017

DPDK Summit China 2017 Summit China 2017 Embedded Network Architecture Optimization Based on Lin Hao T1 Networks Agenda Our History What is an embedded network device Challenge to us Requirements for device today Our solution

More information

Most of the slides in this lecture are either from or adapted from slides provided by the authors of the textbook Computer Systems: A Programmer s

Most of the slides in this lecture are either from or adapted from slides provided by the authors of the textbook Computer Systems: A Programmer s Most of the slides in this lecture are either from or adapted from slides provided by the authors of the textbook Computer Systems: A Programmer s Perspective, 2 nd Edition and are provided from the website

More information

Improve VNF safety with Vhost-User/DPDK IOMMU support

Improve VNF safety with Vhost-User/DPDK IOMMU support Improve VNF safety with Vhost-User/DPDK IOMMU support No UIO anymore! Maxime Coquelin Software Engineer KVM Forum 2017 AGENDA Background Vhost-user device IOTLB implementation Benchmarks Future improvements

More information

Reliably Scalable Name Prefix Lookup! Haowei Yuan and Patrick Crowley! Washington University in St. Louis!! ANCS 2015! 5/8/2015!

Reliably Scalable Name Prefix Lookup! Haowei Yuan and Patrick Crowley! Washington University in St. Louis!! ANCS 2015! 5/8/2015! Reliably Scalable Name Prefix Lookup! Haowei Yuan and Patrick Crowley! Washington University in St. Louis!! ANCS 2015! 5/8/2015! ! My Topic for Today! Goal: a reliable longest name prefix lookup performance

More information

Optimizing Memory Bandwidth

Optimizing Memory Bandwidth Optimizing Memory Bandwidth Don t settle for just a byte or two. Grab a whole fistful of cache. Mike Wall Member of Technical Staff Developer Performance Team Advanced Micro Devices, Inc. make PC performance

More information

Netchannel 2: Optimizing Network Performance

Netchannel 2: Optimizing Network Performance Netchannel 2: Optimizing Network Performance J. Renato Santos +, G. (John) Janakiraman + Yoshio Turner +, Ian Pratt * + HP Labs - * XenSource/Citrix Xen Summit Nov 14-16, 2007 2003 Hewlett-Packard Development

More information

CS5460: Operating Systems Lecture 14: Memory Management (Chapter 8)

CS5460: Operating Systems Lecture 14: Memory Management (Chapter 8) CS5460: Operating Systems Lecture 14: Memory Management (Chapter 8) Important from last time We re trying to build efficient virtual address spaces Why?? Virtual / physical translation is done by HW and

More information

Bringing the Power of ebpf to Open vswitch. Linux Plumber 2018 William Tu, Joe Stringer, Yifeng Sun, Yi-Hung Wei VMware Inc. and Cilium.

Bringing the Power of ebpf to Open vswitch. Linux Plumber 2018 William Tu, Joe Stringer, Yifeng Sun, Yi-Hung Wei VMware Inc. and Cilium. Bringing the Power of ebpf to Open vswitch Linux Plumber 2018 William Tu, Joe Stringer, Yifeng Sun, Yi-Hung Wei VMware Inc. and Cilium.io 1 Outline Introduction and Motivation OVS-eBPF Project OVS-AF_XDP

More information

Afternoon Session 20

Afternoon Session 20 Afternoon Session 20 DPDK Sample Apps L2fwd Crypto Ethtool L2fwd IVSHME M Exception Path KNI Perf Thread VMDq Timer RxTx Callbacks PTP Client L3fwd VF L2fwd CAT TEP Term Quota & W mark Multi Process L2fwd

More information

Agilio CX 2x40GbE with OVS-TC

Agilio CX 2x40GbE with OVS-TC PERFORMANCE REPORT Agilio CX 2x4GbE with OVS-TC OVS-TC WITH AN AGILIO CX SMARTNIC CAN IMPROVE A SIMPLE L2 FORWARDING USE CASE AT LEAST 2X. WHEN SCALED TO REAL LIFE USE CASES WITH COMPLEX RULES TUNNELING

More information

Hash Table Design and Optimization for Software Virtual Switches

Hash Table Design and Optimization for Software Virtual Switches Hash Table Design and Optimization for Software Virtual Switches P R E S E N T E R : R E N WA N G Y I P E N G WA N G, S A M E H G O B R I E L, R E N WA N G, C H A R L I E TA I, C R I S T I A N D U M I

More information

VIRTIO: VHOST DATA PATH ACCELERATION TORWARDS NFV CLOUD. CUNMING LIANG, Intel

VIRTIO: VHOST DATA PATH ACCELERATION TORWARDS NFV CLOUD. CUNMING LIANG, Intel VIRTIO: VHOST DATA PATH ACCELERATION TORWARDS NFV CLOUD CUNMING LIANG, Intel Agenda Towards NFV Cloud vhost Data Path Acceleration vdpa Intro vdpa Design vdpa Implementation Summary & Future Work Towards

More information

Distributed caching for cloud computing

Distributed caching for cloud computing Distributed caching for cloud computing Maxime Lorrillere, Julien Sopena, Sébastien Monnet et Pierre Sens February 11, 2013 Maxime Lorrillere (LIP6/UPMC/CNRS) February 11, 2013 1 / 16 Introduction Context

More information

VIRTIO-NET: VHOST DATA PATH ACCELERATION TORWARDS NFV CLOUD. CUNMING LIANG, Intel

VIRTIO-NET: VHOST DATA PATH ACCELERATION TORWARDS NFV CLOUD. CUNMING LIANG, Intel VIRTIO-NET: VHOST DATA PATH ACCELERATION TORWARDS NFV CLOUD CUNMING LIANG, Intel Agenda Towards NFV Cloud Background & Motivation vhost Data Path Acceleration Intro Design Impl Summary & Future Work Towards

More information

Are You Insured Against Your Noisy Neighbor Sunku Ranganath, Intel Corporation Sridhar Rao, Spirent Communications

Are You Insured Against Your Noisy Neighbor Sunku Ranganath, Intel Corporation Sridhar Rao, Spirent Communications Are You Insured Against Your Noisy Neighbor Sunku Ranganath, Intel Corporation Sridhar Rao, Spirent Communications @SunkuRanganath, @ngignir Legal Disclaimer 2018 Intel Corporation. Intel, the Intel logo,

More information

Service Edge Virtualization - Hardware Considerations for Optimum Performance

Service Edge Virtualization - Hardware Considerations for Optimum Performance Service Edge Virtualization - Hardware Considerations for Optimum Performance Executive Summary This whitepaper provides a high level overview of Intel based server hardware components and their impact

More information

BESS: A Virtual Switch Tailored for NFV

BESS: A Virtual Switch Tailored for NFV BESS: A Virtual Switch Tailored for NFV Sangjin Han, Aurojit Panda, Brian Kim, Keon Jang, Joshua Reich, Saikrishna Edupuganti, Christian Maciocco, Sylvia Ratnasamy, Scott Shenker https://github.com/netsys/bess

More information

VALE: a switched ethernet for virtual machines

VALE: a switched ethernet for virtual machines L < > T H local VALE VALE -- Page 1/23 VALE: a switched ethernet for virtual machines Luigi Rizzo, Giuseppe Lettieri Università di Pisa http://info.iet.unipi.it/~luigi/vale/ Motivation Make sw packet processing

More information

VDPA: VHOST-MDEV AS NEW VHOST PROTOCOL TRANSPORT

VDPA: VHOST-MDEV AS NEW VHOST PROTOCOL TRANSPORT VDPA: VHOST-MDEV AS NEW VHOST PROTOCOL TRANSPORT CUNMING(Steve) LIANG, Intel cunming.liang AT intel.com KVM Forum 2018, Edinburgh October, 2018 Background KVM Forum 2018 / Edinburg / 2018 Intel Corporation

More information

Intel s Architecture for NFV

Intel s Architecture for NFV Intel s Architecture for NFV Evolution from specialized technology to mainstream programming Net Futures 2015 Network applications Legal Disclaimer INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION

More information

Computer Labs: Profiling and Optimization

Computer Labs: Profiling and Optimization Computer Labs: Profiling and Optimization 2 o MIEIC Pedro F. Souto (pfs@fe.up.pt) December 15, 2010 Optimization Speed matters, and it depends mostly on the right choice of Data structures Algorithms If

More information

Dataplane Networking journey in Containers

Dataplane Networking journey in Containers Dataplane Networking journey in Gary Loughnane gary.loughnane@intel.com Kuralamudhan Ramakrishnan kuralamudhan.ramakrishnan@intel.com DPDK Summit Userspace - Dublin- 2017 Discussion topics Container Deployment

More information

Advanced Computer Networks. End Host Optimization

Advanced Computer Networks. End Host Optimization Oriana Riva, Department of Computer Science ETH Zürich 263 3501 00 End Host Optimization Patrick Stuedi Spring Semester 2017 1 Today End-host optimizations: NUMA-aware networking Kernel-bypass Remote Direct

More information

Got Loss? Get zovn! Daniel Crisan, Robert Birke, Gilles Cressier, Cyriel Minkenberg, and Mitch Gusat. ACM SIGCOMM 2013, August, Hong Kong, China

Got Loss? Get zovn! Daniel Crisan, Robert Birke, Gilles Cressier, Cyriel Minkenberg, and Mitch Gusat. ACM SIGCOMM 2013, August, Hong Kong, China Got Loss? Get zovn! Daniel Crisan, Robert Birke, Gilles Cressier, Cyriel Minkenberg, and Mitch Gusat ACM SIGCOMM 2013, 12-16 August, Hong Kong, China Virtualized Server 1 Application Performance in Virtualized

More information

Datacenter Network Solutions Group

Datacenter Network Solutions Group 1 Enabling NFV features in kubernetes IVAN COUGHLAN IVAN.COUGHLAN@INTEL.COM Software Architect Kuralamudhan Ramakrishnan kuralamudhan.ramakrishnan@intel.com Senior Software Engineer Data Center Network

More information

Memory Hierarchy Computing Systems & Performance MSc Informatics Eng. Memory Hierarchy (most slides are borrowed)

Memory Hierarchy Computing Systems & Performance MSc Informatics Eng. Memory Hierarchy (most slides are borrowed) Computing Systems & Performance Memory Hierarchy MSc Informatics Eng. 2011/12 A.J.Proença Memory Hierarchy (most slides are borrowed) AJProença, Computer Systems & Performance, MEI, UMinho, 2011/12 1 2

More information

Lecture: Cache Hierarchies. Topics: cache innovations (Sections B.1-B.3, 2.1)

Lecture: Cache Hierarchies. Topics: cache innovations (Sections B.1-B.3, 2.1) Lecture: Cache Hierarchies Topics: cache innovations (Sections B.1-B.3, 2.1) 1 Types of Cache Misses Compulsory misses: happens the first time a memory word is accessed the misses for an infinite cache

More information

Memory Hierarchy Computing Systems & Performance MSc Informatics Eng. Memory Hierarchy (most slides are borrowed)

Memory Hierarchy Computing Systems & Performance MSc Informatics Eng. Memory Hierarchy (most slides are borrowed) Computing Systems & Performance Memory Hierarchy MSc Informatics Eng. 2012/13 A.J.Proença Memory Hierarchy (most slides are borrowed) AJProença, Computer Systems & Performance, MEI, UMinho, 2012/13 1 2

More information

ResQ: Enabling SLOs in Network Function Virtualization

ResQ: Enabling SLOs in Network Function Virtualization ResQ: Enabling SLOs in Network Function Virtualization Amin Tootoonchian* Aurojit Panda Chang Lan Melvin Walls Katerina Argyraki Sylvia Ratnasamy Scott Shenker *Intel Labs UC Berkeley ICSI NYU Nefeli EPFL

More information

Problem 1. (15 points):

Problem 1. (15 points): CMU 15-418/618: Parallel Computer Architecture and Programming Practice Exercise 1 A Task Queue on a Multi-Core, Multi-Threaded CPU Problem 1. (15 points): The figure below shows a simple single-core CPU

More information

Agenda How DPDK can be used for your Application DPDK Ecosystem boosting your Development Meet the Community Challenges

Agenda How DPDK can be used for your Application DPDK Ecosystem boosting your Development Meet the Community Challenges SPEED MATTERS. All rights reserved. All brand names, trademarks and copyright information cited in this presentation shall remain the property of its registered owners. Agenda How DPDK can be used for

More information

Addressing the Memory Wall

Addressing the Memory Wall Lecture 26: Addressing the Memory Wall Parallel Computer Architecture and Programming CMU 15-418/15-618, Spring 2015 Tunes Cage the Elephant Back Against the Wall (Cage the Elephant) This song is for the

More information

Why memory hierarchy? Memory hierarchy. Memory hierarchy goals. CS2410: Computer Architecture. L1 cache design. Sangyeun Cho

Why memory hierarchy? Memory hierarchy. Memory hierarchy goals. CS2410: Computer Architecture. L1 cache design. Sangyeun Cho Why memory hierarchy? L1 cache design Sangyeun Cho Computer Science Department Memory hierarchy Memory hierarchy goals Smaller Faster More expensive per byte CPU Regs L1 cache L2 cache SRAM SRAM To provide

More information

W H I T E P A P E R. Comparison of Storage Protocol Performance in VMware vsphere 4

W H I T E P A P E R. Comparison of Storage Protocol Performance in VMware vsphere 4 W H I T E P A P E R Comparison of Storage Protocol Performance in VMware vsphere 4 Table of Contents Introduction................................................................... 3 Executive Summary............................................................

More information

Support for Smart NICs. Ian Pratt

Support for Smart NICs. Ian Pratt Support for Smart NICs Ian Pratt Outline Xen I/O Overview Why network I/O is harder than block Smart NIC taxonomy How Xen can exploit them Enhancing Network device channel NetChannel2 proposal I/O Architecture

More information

PASTE: A Networking API for Non-Volatile Main Memory

PASTE: A Networking API for Non-Volatile Main Memory PASTE: A Networking API for Non-Volatile Main Memory Michio Honda (NEC Laboratories Europe) Lars Eggert (NetApp) Douglas Santry (NetApp) TSVAREA@IETF 99, Prague May 22th 2017 More details at our HotNets

More information

Userspace NVMe Driver in QEMU

Userspace NVMe Driver in QEMU Userspace NVMe Driver in QEMU Fam Zheng Senior Software Engineer KVM Form 2017, Prague About NVMe Non-Volatile Memory Express A scalable host interface specification like SCSI and virtio Up to 64k I/O

More information

DPDK Summit China 2017

DPDK Summit China 2017 DPDK Summit China 2017 2 DPDK in container Status Quo and Future Directions Jianfeng Tan, June 2017 3 LEGAL DISCLAIMER No license (express or implied, by estoppel or otherwise) to any intellectual property

More information

COSC 6385 Computer Architecture - Memory Hierarchy Design (III)

COSC 6385 Computer Architecture - Memory Hierarchy Design (III) COSC 6385 Computer Architecture - Memory Hierarchy Design (III) Fall 2006 Reducing cache miss penalty Five techniques Multilevel caches Critical word first and early restart Giving priority to read misses

More information

Next Generation Technology from Intel Intel Pentium 4 Processor

Next Generation Technology from Intel Intel Pentium 4 Processor Next Generation Technology from Intel Intel Pentium 4 Processor 1 The Intel Pentium 4 Processor Platform Intel s highest performance processor for desktop PCs Targeted at consumer enthusiasts and business

More information

Crosstalk between VMs. Alexander Komarov, Application Engineer Software and Services Group Developer Relations Division EMEA

Crosstalk between VMs. Alexander Komarov, Application Engineer Software and Services Group Developer Relations Division EMEA Crosstalk between VMs Alexander Komarov, Application Engineer Software and Services Group Developer Relations Division EMEA 2 September 2015 Legal Disclaimer & Optimization Notice INFORMATION IN THIS DOCUMENT

More information

Building high performance network functions in VPP. Ole Trøan, VPP contributor FOSDEM 2018

Building high performance network functions in VPP. Ole Trøan, VPP contributor FOSDEM 2018 Building high performance network functions in VPP Ole Trøan, ot@cisco.com, VPP contributor FOSDEM 2018 1 2 This talk? Goal: Make you into VPP developers Agenda: VPP architecture An example decomposed

More information

ELEC 5200/6200 Computer Architecture and Design Spring 2017 Lecture 7: Memory Organization Part II

ELEC 5200/6200 Computer Architecture and Design Spring 2017 Lecture 7: Memory Organization Part II ELEC 5200/6200 Computer Architecture and Design Spring 2017 Lecture 7: Organization Part II Ujjwal Guin, Assistant Professor Department of Electrical and Computer Engineering Auburn University, Auburn,

More information

CACHE MEMORIES ADVANCED COMPUTER ARCHITECTURES. Slides by: Pedro Tomás

CACHE MEMORIES ADVANCED COMPUTER ARCHITECTURES. Slides by: Pedro Tomás CACHE MEMORIES Slides by: Pedro Tomás Additional reading: Computer Architecture: A Quantitative Approach, 5th edition, Chapter 2 and Appendix B, John L. Hennessy and David A. Patterson, Morgan Kaufmann,

More information

DPDK Tunneling Offload RONY EFRAIM & YONGSEOK KOH MELLANOX

DPDK Tunneling Offload RONY EFRAIM & YONGSEOK KOH MELLANOX x DPDK Tunneling Offload RONY EFRAIM & YONGSEOK KOH MELLANOX Rony Efraim Introduction to DC w/ overlay network Modern data center (DC) use overly network like Virtual Extensible LAN (VXLAN) and GENEVE

More information

How to Write Fast Numerical Code

How to Write Fast Numerical Code How to Write Fast Numerical Code Lecture: Memory hierarchy, locality, caches Instructor: Markus Püschel TA: Alen Stojanov, Georg Ofenbeck, Gagandeep Singh Organization Temporal and spatial locality Memory

More information

Zero-copy Receive for Virtualized Network Devices

Zero-copy Receive for Virtualized Network Devices Zero-copy Receive for Virtualized Network Devices Kalman Meth, Joel Nider and Mike Rapoport IBM Research Labs - Haifa {meth,joeln,rapoport}@il.ibm.com Abstract. When receiving network traffic on guest

More information

An FPGA-Based Optical IOH Architecture for Embedded System

An FPGA-Based Optical IOH Architecture for Embedded System An FPGA-Based Optical IOH Architecture for Embedded System Saravana.S Assistant Professor, Bharath University, Chennai 600073, India Abstract Data traffic has tremendously increased and is still increasing

More information

Virtual Memory. Virtual Memory

Virtual Memory. Virtual Memory Virtual Memory Virtual Memory Main memory is cache for secondary storage Secondary storage (disk) holds the complete virtual address space Only a portion of the virtual address space lives in the physical

More information

Caches and Memory. Anne Bracy CS 3410 Computer Science Cornell University. See P&H Chapter: , 5.8, 5.10, 5.13, 5.15, 5.17

Caches and Memory. Anne Bracy CS 3410 Computer Science Cornell University. See P&H Chapter: , 5.8, 5.10, 5.13, 5.15, 5.17 Caches and emory Anne Bracy CS 34 Computer Science Cornell University Slides by Anne Bracy with 34 slides by Professors Weatherspoon, Bala, ckee, and Sirer. See P&H Chapter: 5.-5.4, 5.8, 5., 5.3, 5.5,

More information

Supporting Fine-Grained Network Functions through Intel DPDK

Supporting Fine-Grained Network Functions through Intel DPDK Supporting Fine-Grained Network Functions through Intel DPDK Ivano Cerrato, Mauro Annarumma, Fulvio Risso - Politecnico di Torino, Italy EWSDN 2014, September 1st 2014 This project is co-funded by the

More information

6.9. Communicating to the Outside World: Cluster Networking

6.9. Communicating to the Outside World: Cluster Networking 6.9 Communicating to the Outside World: Cluster Networking This online section describes the networking hardware and software used to connect the nodes of cluster together. As there are whole books and

More information

Programming Parallel Computers

Programming Parallel Computers ICS-E4020 Programming Parallel Computers Jukka Suomela Jaakko Lehtinen Samuli Laine Aalto University Spring 2016 users.ics.aalto.fi/suomela/ppc-2016/ New code must be parallel! otherwise a computer from

More information

Improving Driver Performance A Worked Example

Improving Driver Performance A Worked Example USERSPACE, October 2016 Improving Driver Performance A Worked Example Bruce Richardson Legal Disclaimers No license (express or implied, by estoppel or otherwise) to any intellectual property rights is

More information

Learning with Purpose

Learning with Purpose Network Measurement for 100Gbps Links Using Multicore Processors Xiaoban Wu, Dr. Peilong Li, Dr. Yongyi Ran, Prof. Yan Luo Department of Electrical and Computer Engineering University of Massachusetts

More information

DPDK Integration within F5 BIG-IP BRENT BLOOD, SR MANAGER SOFTWARE ENGINEERING VIJAY MANICKAM, SR SOFTWARE ENGINEER

DPDK Integration within F5 BIG-IP BRENT BLOOD, SR MANAGER SOFTWARE ENGINEERING VIJAY MANICKAM, SR SOFTWARE ENGINEER x DPDK Integration within F5 BIG-IP BRENT BLOOD, SR MANAGER SOFTWARE ENGINEERING VIJAY MANICKAM, SR SOFTWARE ENGINEER F5 Company Snapshot Founded: 1996 IPO: June 1999 Employees: 4,395 Headquarters: Seattle,

More information

PacketShader: A GPU-Accelerated Software Router

PacketShader: A GPU-Accelerated Software Router PacketShader: A GPU-Accelerated Software Router Sangjin Han In collaboration with: Keon Jang, KyoungSoo Park, Sue Moon Advanced Networking Lab, CS, KAIST Networked and Distributed Computing Systems Lab,

More information

SoftRDMA: Rekindling High Performance Software RDMA over Commodity Ethernet

SoftRDMA: Rekindling High Performance Software RDMA over Commodity Ethernet SoftRDMA: Rekindling High Performance Software RDMA over Commodity Ethernet Mao Miao, Fengyuan Ren, Xiaohui Luo, Jing Xie, Qingkai Meng, Wenxue Cheng Dept. of Computer Science and Technology, Tsinghua

More information

Improve Performance of Kube-proxy and GTP-U using VPP

Improve Performance of Kube-proxy and GTP-U using VPP Improve Performance of Kube-proxy and GTP-U using VPP Hongjun Ni (hongjun.ni@intel.com) Danny Zhou (danny.zhou@intel.com) Johnson Li (johnson.li@intel.com) Network Platform Group, DCG, Intel Acknowledgement:

More information

I/O Handling. ECE 650 Systems Programming & Engineering Duke University, Spring Based on Operating Systems Concepts, Silberschatz Chapter 13

I/O Handling. ECE 650 Systems Programming & Engineering Duke University, Spring Based on Operating Systems Concepts, Silberschatz Chapter 13 I/O Handling ECE 650 Systems Programming & Engineering Duke University, Spring 2018 Based on Operating Systems Concepts, Silberschatz Chapter 13 Input/Output (I/O) Typical application flow consists of

More information

Freescale, the Freescale logo, AltiVec, C-5, CodeTEST, CodeWarrior, ColdFire, C-Ware, t he Energy Efficient Solutions logo, mobilegt, PowerQUICC,

Freescale, the Freescale logo, AltiVec, C-5, CodeTEST, CodeWarrior, ColdFire, C-Ware, t he Energy Efficient Solutions logo, mobilegt, PowerQUICC, Freescale, the Freescale logo, AltiVec, C-5, CodeTEST, CodeWarrior, ColdFire, C-Ware, t he Energy Efficient Solutions logo, mobilegt, PowerQUICC, QorIQ, StarCore and Symphony are trademarks of Freescale

More information