SoNIC: Precise Real1me So3ware Access and Control of Wired Networks. Ki Suh Lee, Han Wang, Hakim Weatherspoon Cornell University

Size: px
Start display at page:

Download "SoNIC: Precise Real1me So3ware Access and Control of Wired Networks. Ki Suh Lee, Han Wang, Hakim Weatherspoon Cornell University"

Transcription

1 SoNIC: Precise Real1me So3ware Access and Control of Wired s Ki Suh Lee, Han Wang, Hakim Weatherspoon Cornell University 4/11/13 SoNIC NSDI

2 Interpacket Delay and Research Link Interpacket gap, spacing, arrival 1me, Important metric for network research Packet Genera1on Can be improved with access to the PHY Increasing Throughput IPG Packet i Packet i+1 IPD Detec1ng 1ming channel Packet Capture Characteriza1on Es1ma1ng bandwidth 4/11/13 SoNIC NSDI

3 Research enlightened via the PHY Link Valuable informa1on: Idle characters IPG Packet i Packet i+1 IPD Can provide precise 1ming base for control Each bit is ~97 ps wide 4/11/13 SoNIC NSDI

4 Research enlightened via the PHY Link Valuable informa1on: 12 /I/s Idle = 100bits characters = 9.7ns IPG Packet i Packet i+1 One Idle character (/I/) = 7~8 bits Each bit is ~97 ps wide Can provide precise 1ming base for control Packet Genera1on Detec1ng 1ming channel Packet Capture 4/11/13 SoNIC NSDI

5 Principle #1: Precision Precise network measurements is enabled via access to the physical layer (and the idle characters and bits within interpacket gap) 4/11/13 SoNIC NSDI

6 How to control the idle characters (bits)? Link Access to the en1re stream is required IPG Packet i Packet i+1 Issue1: The PHY is simply a black box No interface from NIC or OS Valuable informa1on is invisible (discarded) Packet i Packet i+1 Packet i+2 Packet i Packet i+1 Issue2: Limited access to hardware We are network systems researchers a.k.a. we like so3ware Packet i+2 4/11/13 SoNIC NSDI

7 Principle #2: So3ware Systems researchers need so3ware access to the physical layer 4/11/13 SoNIC NSDI

8 Precision + So3ware = Physics equipment??? BiFocals [IMC 10 Freedman, Marian, Lee, Birman, Weatherspoon, Xu] Enabled novel network research Precision + So3ware = Laser + Oscilloscope + Offline analysis Allowed precise control in so3ware Limita1ons Offline (not real'me) Limited Buffering Expensive 4/11/13 SoNIC NSDI

9 Principle #3: Real1me systems researchers need access and control of the physical layer (interpacket gap) con1nuously in real1me 4/11/13 SoNIC NSDI

10 Challenge Link Goal: Control every bit in so-ware in real'me IPG Packet i Packet i+1 IPD Enable novel network research Challenge Requires unprecedented so3ware access to the PHY 4/11/13 SoNIC NSDI

11 Outline Introduc1on SoNIC: So3ware- defined Interface Card Background: 10GbE Stack Design Research s Conclusion 4/11/13 SoNIC NSDI

12 SoNIC: So3ware- defined Interface Card Implements the PHY in so3ware IPG Link Enabling control and access to every bit in real1me With commodity components Thus, enabling novel network research How? Packet i Packet i+1 IPD Backgrounds: 10 GbE stack Design and implementa1on Hardware & So3ware Op1miza1ons 4/11/13 SoNIC NSDI

13 10GbE Stack L3 Hdr L2 Hdr L3 Hdr Link Preamble Eth Hdr L2 Hdr L3 Hdr CRC Gap 64/66b PCS Encode Decode 64 bit Idle 2 bit characters syncheader (/I/) Gigabits /S/ /D/ /D/ /D/ /D/ /T/ /E/ Scrambler Descrambler 16 bit Gearbox Blocksync PMA PMD /11/13 SoNIC NSDI

14 10GbE Stack SW L2 Hdr L3 Hdr Packet i Packet i+1 L3 Hdr Link Preamble Eth Hdr L2 Hdr L3 Hdr CRC Gap 64/66b PCS Encode Decode HW /S/ /D/ /D/ /D/ /D/ /T/ /E/ Scrambler Descrambler Packet i Packet i+1 Gearbox Blocksync PMA PMD Commodity NIC 4/11/13 SoNIC NSDI

15 10GbE Stack L3 Hdr SW Link Preamble Eth Hdr L2 Hdr L2 Hdr L3 Hdr L3 Hdr HW CRC Link Gap 64/66b PCS Encode Decode Scrambler Descrambler SW 64/66b PCS /S/ Packet i /D/ /D/ Packet /D/ i+1 /D/ Encode /T/ Decode /E/ Scrambler Descrambler Gearbox Blocksync PMA PMD HW SoNIC NetFPGA Gearbox Blocksync PMA PMD 4/11/13 SoNIC NSDI

16 SoNIC Design L3 Hdr L2 Hdr L3 Hdr Link Preamble Eth Hdr L2 Hdr L3 Hdr CRC Gap 64/66b PCS Encode Decode Scrambler Descrambler SW /S/ /D/ /D/ /D/ /D/ /T/ /E/ Gearbox Blocksync PMA PMD HW SoNIC 4/11/13 SoNIC NSDI

17 SoNIC Design and Architecture L3 Hdr APP Userspace L2 Hdr L3 Hdr APP Kernel Link Preamble Eth Hdr L2 Hdr TX MAC L3 Hdr RX MAC CRC Gap 64/66b PCS Encode Decode Scrambler Descrambler SW /S/ /D/ /D/ /D/ /D/ /T/ /E/ TX PCS RX PCS Gearbox Blocksync PMA PMD HW Transceiver Transceiver SoNIC Gearbox Blocksync Hardware SFP+ 4/11/13 SoNIC NSDI

18 SoNIC Design: Hardware 64/66b PCS Encode Scrambler Gearbox Link PMA PMD Decode Descrambler Blocksync SW HW To deliver every bit from/to so3ware High- speed transceivers PCIe Gen2 (=32Gbps) Op1mized DMA engine SFP+ SFP+ FPGA PCIeGen2 4/11/13 SoNIC NSDI

19 SoNIC Design: So3ware Port 0 APP Port 1 APP Link TX MAC RX MAC TX MAC RX MAC 64/66b PCS Encode Decode Scrambler Descrambler Gearbox Blocksync PMA PMD SW HW TX PCS RX PCS TX PCS RX PCS Dedicated Kernel Threads TX / RX PCS, TX / RX MAC threads APP thread: Interface to userspace Packet i Packet i+1 4/11/13 SoNIC NSDI

20 SoNIC Design: Synchroniza1on Port 0 APP Low- latency FIFOs Port 1 APP Link TX MAC RX MAC TX MAC RX MAC 64/66b PCS Encode Decode Scrambler Descrambler Gearbox Blocksync PMA SW HW TX PCS SFP+ SFP+ RX PCS FPGA TX PCS RX PCS Pointer- polling No Interrupts PMD PCIeGen2 4/11/13 SoNIC NSDI

21 SoNIC Design: Op1miza1ons 64/66b PCS Encode Scrambler Gearbox Link PMA PMD Decode Descrambler Blocksync Scrambler Naïve Implementa1on s ç state d ç data for i = 0 è 63 do in ç (d >> i) & 1 out ç (in (s >> 38) (s >> 57))&1 s ç (s << 1) out r ç r (out << i) state ç s end for CRC computa1on DMA engine G( x) = x58+ x Gbps 21 Gbps Op1mized Implementa1on s ç state d ç data r ç (s >> 6) (s >> 25) d r ç r (r << 39) (r << 58) state ç r 4/11/13 SoNIC NSDI

22 SoNIC Design: Interface and Control Hardware control: ioctl syscall I/O : character device interface Sample C code for packet genera1on and capture 1: #include "sonic.h" 2: 3: struct sonic_pkt_gen_info info = { 4:.mode = 0, 5:.pkt_num = UL, 6:.pkt_len = 1518, 7:.mac_src = "00:11:22:33:44:55", 8:.mac_dst = "aa:bb:cc:dd:ee:ff", 9:.ip_src = " ", 10:.ip_dst = " ", 11:.port_src = 5000, 12:.port_dst = 5000, 13:.idle = 12, 14: }; 15: 16: /* OPEN DEVICE*/ 17: fd1 = open(sonic_control_path, O_RDWR); 18: fd2 = open(sonic_port1_path, O_RDONLY); 19: /* CONFIG SONIC CARD FOR PACKET GEN*/ 20: ioctl(fd1, SONIC_IOC_RESET) 21: ioctl(fd1, SONIC_IOC_SET_MODE, PKT_GEN_CAP) 22: ioctl(fd1, SONIC_IOC_PORT0_INFO_SET, &info) 23 24: /* START EXPERIMENT*/ 25: ioctl(fd1, SONIC_IOC_START) 26: // wait 1ll experiment finishes 27: ioctl(fd1, SONIC_IOC_STOP) 28: 29: /* CAPTURE PACKET */ 30: while ((ret = read(fd2, buf, 65536)) > 0) { 31: // process data 32: } 33: 34: close(fd1); 35: close(fd2); 4/11/13 SoNIC 22

23 Outline Introduc1on SoNIC: So3ware- defined Interface Card Research s Packet Genera1on Packet Capture Covert 1ming channel Conclusion 4/11/13 SoNIC NSDI

24 Research s Link Interpacket delays and gaps IPG Packet i Packet i+1 IPD Packet Genera1on Detec1ng 1ming channel Packet Capture 4/11/13 SoNIC NSDI

25 Packet Genera1on and Capture Basic func1ons for network research Genera1on: SoNIC allows control of IPGs in # of /I/s Capture: SoNIC captures what was sent with IPGs in bits APP APP TX MAC RX MAC TX MAC RX MAC TX PCS RX PCS TX PCS RX PCS 1518B 1518B 1518B 1518B 1518B 9Gbps, IPD =13992 bits (1357ns) 4/11/13 SoNIC NSDI

26 Packet Genera1on SoNIC allows precise CDF of control generated of IPGs IPDs 1" TX MAC TX PCS APP RX MAC RX PCS CDF 0.8" 0.6" 0.4" 0.2" Specialized NIC Higher variance SoNIC Zero variance!!! SoNIC" Sniffer"10G" TX MAC TX PCS APP RX MAC RX PCS 0" 1000" 1500" 2000" 2500" 3000" Interpacket delays (ns) 1518B 1518B 1518B 1518B 1518B 9Gbps, IPD =13992 bits (1357ns) 4/11/13 SoNIC NSDI

27 Packet Capture SoNIC captures what CDF of is sent captured IPDs 1" TX MAC APP RX MAC CDF 0.8" 0.6" 0.4" SoNIC" Kernel" Userspace" Sniffer"10G" TX MAC APP RX MAC TX PCS RX PCS 0.2" TX PCS RX PCS 0" 0" 1000" 2000" 3000" 4000" 5000" Interpacket delays (ns) 1518B 1518B 1518B 1518B 1518B 9Gbps, IPD =13992 bits (1357ns) 4/11/13 SoNIC NSDI

28 Covert Timing Channel Embedding signals into interpacket gaps. Large gap: 1 Small gap: 0 Covert 1ming channel by modula1ng IPGs at 100ns TX MAC TX PCS APP RX MAC RX PCS Packet i Packet i+1 Packet i Packet i+1 Overt channel at 3 Gbps Covert channel at 250 kbps Over 4- hops with < 1% BER TX MAC TX PCS APP RX MAC RX PCS 4/11/13 SoNIC NSDI

29 Covert Timing Channel Modula'ng IPGS at 100ns scale (=128 /I/s) CDF 1" SoNIC" 3562 /I/s 0.8" Kernel" /I/s /I/s 0.6" BER = 0.37% TX MAC APP RX MAC 0.4" TX MAC APP RX MAC TX PCS RX PCS 0.2" 0 1 TX PCS RX PCS 0" 500" 1500" 2500" 3500" Interpacket delays (ns) 4500" 1 : /I/s 0 : /I/s 1 : a /I/s 0 : 3562 a /I/s 4/11/13 SoNIC NSDI

30 Contribu1ons Research Unprecedented access to the PHY with commodity hardware A pla orm for cross- network- layer research Can improve network research applica1ons Engineering Precise control of interpacket gaps (delays) Design and implementa1on of the PHY in so3ware Novel scalable hardware design Op1miza1ons / Parallelism Status Measurements in large scale: DCN, GENI, 40 GbE 4/11/13 SoNIC NSDI

31 Conclusion Precise Real1me So3ware Access to the PHY Commodity components An FPGA development board, Intel architecture applica1ons measurements characteriza1on steganography Webpage: h p://sonic.cs.cornell.edu SoNIC is available Open Source. 4/11/13 SoNIC NSDI

32 Thank you Demo tonight! h p://sonic.cs.cornell.edu

Data Center Traffic and Measurements: SoNIC

Data Center Traffic and Measurements: SoNIC Center Traffic and Measurements: SoNIC Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems and ing November 12, 2014 Slides from USENIX symposium on ed Systems

More information

Globally Synchronized time via Datacenter Networks

Globally Synchronized time via Datacenter Networks 1 Globally Synchronized time via Datacenter Networks Ki Suh Lee Cornell University Joint work with Han Wang, Vishal Shrivastav and Hakim Weatherspoon 2 Synchronized Clocks Fundamental for network and distributed

More information

Data Center Networks and Switching and Queueing and Covert Timing Channels

Data Center Networks and Switching and Queueing and Covert Timing Channels Data Center Networks and Switching and Queueing and Covert Timing Channels Hakim Weatherspoon Associate Professor, Dept of Computer Science CS 5413: High Performance Computing and Networking March 10,

More information

PHY Covert Channels: Can you see the Idles?

PHY Covert Channels: Can you see the Idles? PHY Covert Channels: Can you see the Idles? Ki Suh Lee, Han Wang, and Hakim Weatherspoon, Cornell University https://www.usenix.org/conference/nsdi4/technical-sessions/presentation/lee This paper is included

More information

TCP Over SoNIC. Xuke Fang Cornell University. XingXiang Lao Cornell University

TCP Over SoNIC. Xuke Fang Cornell University. XingXiang Lao Cornell University TCP Over SoNIC Xuke Fang Cornell University XingXiang Lao Cornell University ABSTRACT SoNIC [1], a Software-defined Network Interface Card, which provides the access to the physical layer and data link

More information

NetSlices: Scalable Mul/- Core Packet Processing in User- Space

NetSlices: Scalable Mul/- Core Packet Processing in User- Space NetSlices: Scalable Mul/- Core Packet Processing in - Space Tudor Marian, Ki Suh Lee, Hakim Weatherspoon Cornell University Presented by Ki Suh Lee Packet Processors Essen/al for evolving networks Sophis/cated

More information

Data Center Traffic and Measurements: Available Bandwidth Estimation Hakim Weatherspoon

Data Center Traffic and Measurements: Available Bandwidth Estimation Hakim Weatherspoon Data Center Traffic and Measurements: Available Bandwidth Estimation Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems and Networking November 14, 2014

More information

Software Routers: NetMap

Software Routers: NetMap Software Routers: NetMap Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems and Networking October 8, 2014 Slides from the NetMap: A Novel Framework for

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

40-Gbps and 100-Gbps Ethernet in Altera Devices

40-Gbps and 100-Gbps Ethernet in Altera Devices 40-Gbps and 100-Gbps Ethernet in Altera Devices Transceiver Portfolio Workshops 2009 Agenda 40/100 GbE standards 40/100 GbE IP in Altera devices Stratix IV GT FPGA features and advantages Altera standards

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

Towards Fully Synchronized (and Programmable) Datacenter Networks

Towards Fully Synchronized (and Programmable) Datacenter Networks Towards Fully Synchronized (and Programmable) Datacenter Networks Vishal Shrivastav, Cornell University 28 Jul 2016 University of Cambridge Outline DTP: Datacenter Time Protocol [SIGCOMM 16] SHOAL: Synchronized

More information

Hector Codes - SoNIC Steganography

Hector Codes - SoNIC Steganography Hector Codes - SoNIC Steganography Hector A. Tosado Jimenez August 7, 2011 1 Introduction asdin our world we tend to rely too much on a variety of networks in order to have a functional society. May it

More information

GASPP: A GPU- Accelerated Stateful Packet Processing Framework

GASPP: A GPU- Accelerated Stateful Packet Processing Framework GASPP: A GPU- Accelerated Stateful Packet Processing Framework Giorgos Vasiliadis, FORTH- ICS, Greece Lazaros Koromilas, FORTH- ICS, Greece Michalis Polychronakis, Columbia University, USA So5ris Ioannidis,

More information

Packet Clustering Introduced by Routers: Modeling, Analysis and Experiments

Packet Clustering Introduced by Routers: Modeling, Analysis and Experiments Packet Clustering Introduced by Routers: Modeling, Analysis and Experiments Chiun Lin Lim 1, Ki Suh Lee 2, Han Wang 2, Hakim Weatherspoon 2, Ao Tang 1 1 School of Electrical and Computer Engineering, Cornell

More information

PowerPC on NetFPGA CSE 237B. Erik Rubow

PowerPC on NetFPGA CSE 237B. Erik Rubow PowerPC on NetFPGA CSE 237B Erik Rubow NetFPGA PCI card + FPGA + 4 GbE ports FPGA (Virtex II Pro) has 2 PowerPC hard cores Untapped resource within NetFPGA community Goals Evaluate performance of on chip

More information

OSNT A Community-owned platform for high-performance and low-cost network testing

OSNT A Community-owned platform for high-performance and low-cost network testing May, 14 2015 OSNT A Community-owned platform for high-performance and low-cost network testing Gianni Antichi gianni.antichi@cl.cam.ac.uk 1 Let s start with the context. We all know why we need network

More information

High Performance Packet Processing with FlexNIC

High Performance Packet Processing with FlexNIC High Performance Packet Processing with FlexNIC Antoine Kaufmann, Naveen Kr. Sharma Thomas Anderson, Arvind Krishnamurthy University of Washington Simon Peter The University of Texas at Austin Ethernet

More information

Impact of Cache Coherence Protocols on the Processing of Network Traffic

Impact of Cache Coherence Protocols on the Processing of Network Traffic Impact of Cache Coherence Protocols on the Processing of Network Traffic Amit Kumar and Ram Huggahalli Communication Technology Lab Corporate Technology Group Intel Corporation 12/3/2007 Outline Background

More information

Near- Data Computa.on: It s Not (Just) About Performance

Near- Data Computa.on: It s Not (Just) About Performance Near- Data Computa.on: It s Not (Just) About Performance Steven Swanson Non- Vola0le Systems Laboratory Computer Science and Engineering University of California, San Diego 1 Solid State Memories NAND

More information

PE310G4TSF4I71 Quad Port SFP+ 10 Gigabit Ethernet PCI Express Time Stamp Server Adapter Intel Based

PE310G4TSF4I71 Quad Port SFP+ 10 Gigabit Ethernet PCI Express Time Stamp Server Adapter Intel Based PE310G4TSF4I71 Quad Port SFP+ 10 Gigabit Ethernet PCI Express Time Stamp Server Adapter Intel Based Product Description Silicom s 40 Gigabit Ethernet PCI Express Time Stamping server adapter is designed

More information

NetFPGA Hardware Architecture

NetFPGA Hardware Architecture NetFPGA Hardware Architecture Jeffrey Shafer Some slides adapted from Stanford NetFPGA tutorials NetFPGA http://netfpga.org 2 NetFPGA Components Virtex-II Pro 5 FPGA 53,136 logic cells 4,176 Kbit block

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

WORKLOAD CHARACTERIZATION OF INTERACTIVE CLOUD SERVICES BIG AND SMALL SERVER PLATFORMS

WORKLOAD CHARACTERIZATION OF INTERACTIVE CLOUD SERVICES BIG AND SMALL SERVER PLATFORMS WORKLOAD CHARACTERIZATION OF INTERACTIVE CLOUD SERVICES ON BIG AND SMALL SERVER PLATFORMS Shuang Chen*, Shay Galon**, Christina Delimitrou*, Srilatha Manne**, and José Martínez* *Cornell University **Cavium

More information

INT G bit TCP Offload Engine SOC

INT G bit TCP Offload Engine SOC INT 10011 10 G bit TCP Offload Engine SOC Product brief, features and benefits summary: Highly customizable hardware IP block. Easily portable to ASIC flow, Xilinx/Altera FPGAs or Structured ASIC flow.

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

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

CS 5413: High Performance Systems and Networking

CS 5413: High Performance Systems and Networking CS 5413: High Performance Systems and Networking Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems and Networking January 25, 2017 Goals for Today Background

More information

RDMA and Hardware Support

RDMA and Hardware Support RDMA and Hardware Support SIGCOMM Topic Preview 2018 Yibo Zhu Microsoft Research 1 The (Traditional) Journey of Data How app developers see the network Under the hood This architecture had been working

More information

UDP10G-IP reference design manual

UDP10G-IP reference design manual UDP10G-IP reference design manual Rev1.2 22-Mar-18 1 Introduction Comparing to TCP, UDP provides a procedure to send messages with a minimum of protocol mechanism, but the data cannot guarantee to arrive

More information

Ethernet transport protocols for FPGA

Ethernet transport protocols for FPGA Ethernet transport protocols for FPGA Wojciech M. Zabołotny Institute of Electronic Systems, Warsaw University of Technology Previous version available at: https://indico.gsi.de/conferencedisplay.py?confid=3080

More information

BlazePPS (Blaze Packet Processing System) CSEE W4840 Project Design

BlazePPS (Blaze Packet Processing System) CSEE W4840 Project Design BlazePPS (Blaze Packet Processing System) CSEE W4840 Project Design Valeh Valiollahpour Amiri (vv2252) Christopher Campbell (cc3769) Yuanpei Zhang (yz2727) Sheng Qian ( sq2168) March 26, 2015 I) Hardware

More information

P a g e i. Project Report on TCP over SoNIC. Project Report. TCP over SoNIC. Group Members. Abhishek Kumar Maurya

P a g e i. Project Report on TCP over SoNIC. Project Report. TCP over SoNIC. Group Members. Abhishek Kumar Maurya P a g e i Project Report TCP over SoNIC Group Members Name Abhishek Kumar Maurya Amarinder Singh Bindra Gaurav Aggarwal Email ID am2633@cornell.edu ab2546@cornell.edu ga286@cornell.edu P a g e ii Table

More information

FAQ. Release rc2

FAQ. Release rc2 FAQ Release 19.02.0-rc2 January 15, 2019 CONTENTS 1 What does EAL: map_all_hugepages(): open failed: Permission denied Cannot init memory mean? 2 2 If I want to change the number of hugepages allocated,

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

NFS/RDMA over 40Gbps iwarp Wael Noureddine Chelsio Communications

NFS/RDMA over 40Gbps iwarp Wael Noureddine Chelsio Communications NFS/RDMA over 40Gbps iwarp Wael Noureddine Chelsio Communications Outline RDMA Motivating trends iwarp NFS over RDMA Overview Chelsio T5 support Performance results 2 Adoption Rate of 40GbE Source: Crehan

More information

Multi-Gigabit Transceivers Getting Started with Xilinx s Rocket I/Os

Multi-Gigabit Transceivers Getting Started with Xilinx s Rocket I/Os Multi-Gigabit Transceivers Getting Started with Xilinx s Rocket I/Os Craig Ulmer cdulmer@sandia.gov July 26, 2007 Craig Ulmer SNL/CA Sandia is a multiprogram laboratory operated by Sandia Corporation,

More information

Feature Rich Flow Monitoring with P4

Feature Rich Flow Monitoring with P4 Feature Rich Flow Monitoring with P4 John Sonchack University of Pennsylvania 1 Outline Introduction: Flow Records Design and Implementation: P4 Accelerated Flow Record Generation Benchmarks and Optimizations

More information

Input / Output. Kevin Webb Swarthmore College April 12, 2018

Input / Output. Kevin Webb Swarthmore College April 12, 2018 Input / Output Kevin Webb Swarthmore College April 12, 2018 xkcd #927 Fortunately, the charging one has been solved now that we've all standardized on mini-usb. Or is it micro-usb? Today s Goals Characterize

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

FlexNIC: Rethinking Network DMA

FlexNIC: Rethinking Network DMA FlexNIC: Rethinking Network DMA Antoine Kaufmann Simon Peter Tom Anderson Arvind Krishnamurthy University of Washington HotOS 2015 Networks: Fast and Growing Faster 1 T 400 GbE Ethernet Bandwidth [bits/s]

More information

Exception-Less System Calls for Event-Driven Servers

Exception-Less System Calls for Event-Driven Servers Exception-Less System Calls for Event-Driven Servers Livio Soares and Michael Stumm University of Toronto Talk overview At OSDI'10: exception-less system calls Technique targeted at highly threaded servers

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

TOE10G-IP with CPU reference design

TOE10G-IP with CPU reference design TOE10G-IP with CPU reference design Rev1.1 6-Feb-19 1 Introduction TCP/IP is the core protocol of the Internet Protocol Suite for networking application. TCP/IP model has four layers, i.e. Application

More information

On the cost of tunnel endpoint processing in overlay virtual networks

On the cost of tunnel endpoint processing in overlay virtual networks J. Weerasinghe; NVSDN2014, London; 8 th December 2014 On the cost of tunnel endpoint processing in overlay virtual networks J. Weerasinghe & F. Abel IBM Research Zurich Laboratory Outline Motivation Overlay

More information

Prof. Hakim Weatherspoon CS 3410, Spring 2015 Computer Science Cornell University. See: Online P&H Chapter 6.5 6

Prof. Hakim Weatherspoon CS 3410, Spring 2015 Computer Science Cornell University. See: Online P&H Chapter 6.5 6 Prof. Hakim Weatherspoon CS 3410, Spring 2015 Computer Science Cornell University See: Online P&H Chapter 6.5 6 Project3 submit souped up bot to CMS Project3 Cache Race Games night Monday, May 4 th, 5pm

More information

LegUp: Accelerating Memcached on Cloud FPGAs

LegUp: Accelerating Memcached on Cloud FPGAs 0 LegUp: Accelerating Memcached on Cloud FPGAs Xilinx Developer Forum December 10, 2018 Andrew Canis & Ruolong Lian LegUp Computing Inc. 1 COMPUTE IS BECOMING SPECIALIZED 1 GPU Nvidia graphics cards are

More information

The Missing Piece of Virtualization. I/O Virtualization on 10 Gb Ethernet For Virtualized Data Centers

The Missing Piece of Virtualization. I/O Virtualization on 10 Gb Ethernet For Virtualized Data Centers The Missing Piece of Virtualization I/O Virtualization on 10 Gb Ethernet For Virtualized Data Centers Agenda 10 GbE Adapters Built for Virtualization I/O Throughput: Virtual & Non-Virtual Servers Case

More information

Slides on cross- domain call and Remote Procedure Call (RPC)

Slides on cross- domain call and Remote Procedure Call (RPC) Slides on cross- domain call and Remote Procedure Call (RPC) This classic paper is a good example of a microbenchmarking study. It also explains the RPC abstraction and serves as a case study of the nuts-and-bolts

More information

DCS-ctrl: A Fast and Flexible Device-Control Mechanism for Device-Centric Server Architecture

DCS-ctrl: A Fast and Flexible Device-Control Mechanism for Device-Centric Server Architecture DCS-ctrl: A Fast and Flexible ice-control Mechanism for ice-centric Server Architecture Dongup Kwon 1, Jaehyung Ahn 2, Dongju Chae 2, Mohammadamin Ajdari 2, Jaewon Lee 1, Suheon Bae 1, Youngsok Kim 1,

More information

Networking at the Speed of Light

Networking at the Speed of Light Networking at the Speed of Light Dror Goldenberg VP Software Architecture MaRS Workshop April 2017 Cloud The Software Defined Data Center Resource virtualization Efficient services VM, Containers uservices

More information

Light: A Scalable, High-performance and Fully-compatible User-level TCP Stack. Dan Li ( 李丹 ) Tsinghua University

Light: A Scalable, High-performance and Fully-compatible User-level TCP Stack. Dan Li ( 李丹 ) Tsinghua University Light: A Scalable, High-performance and Fully-compatible User-level TCP Stack Dan Li ( 李丹 ) Tsinghua University Data Center Network Performance Hardware Capability of Modern Servers Multi-core CPU Kernel

More information

IsoStack Highly Efficient Network Processing on Dedicated Cores

IsoStack Highly Efficient Network Processing on Dedicated Cores IsoStack Highly Efficient Network Processing on Dedicated Cores Leah Shalev Eran Borovik, Julian Satran, Muli Ben-Yehuda Outline Motivation IsoStack architecture Prototype TCP/IP over 10GE on a single

More information

Joint ITU-T/IEEE Workshop on Carrier-class Ethernet

Joint ITU-T/IEEE Workshop on Carrier-class Ethernet Joint ITU-T/IEEE Workshop on Carrier-class Ethernet Proposed 100GE PCS and Electrical Interface Gary Nicholl, Senior Technical Leader Cisco Systems 1 Outline The following slides present a proposal for

More information

A Next Generation Home Access Point and Router

A Next Generation Home Access Point and Router A Next Generation Home Access Point and Router Product Marketing Manager Network Communication Technology and Application of the New Generation Points of Discussion Why Do We Need a Next Gen Home Router?

More information

I/O, today, is Remote (block) Load/Store, and must not be slower than Compute, any more

I/O, today, is Remote (block) Load/Store, and must not be slower than Compute, any more I/O, today, is Remote (block) Load/Store, and must not be slower than Compute, any more Manolis Katevenis FORTH, Heraklion, Crete, Greece (in collab. with Univ. of Crete) http://www.ics.forth.gr/carv/

More information

FPGA Augmented ASICs: The Time Has Come

FPGA Augmented ASICs: The Time Has Come FPGA Augmented ASICs: The Time Has Come David Riddoch Steve Pope Copyright 2012 Solarflare Communications, Inc. All Rights Reserved. Hardware acceleration is Niche (With the obvious exception of graphics

More information

Xilinx Answer QDMA Performance Report

Xilinx Answer QDMA Performance Report Xilinx Answer 71453 QDMA Performance Report Important Note: This downloadable PDF of an Answer Record is provided to enhance its usability and readability. It is important to note that Answer Records are

More information

Developing deterministic networking technology for railway applications using TTEthernet software-based end systems

Developing deterministic networking technology for railway applications using TTEthernet software-based end systems Developing deterministic networking technology for railway applications using TTEthernet software-based end systems Project n 100021 Astrit Ademaj, TTTech Computertechnik AG Outline GENESYS requirements

More information

CoVisor: A Composi,onal Hypervisor for So6ware- Defined Networks

CoVisor: A Composi,onal Hypervisor for So6ware- Defined Networks CoVisor: A Composi,onal Hypervisor for So6ware- Defined Networks Xin Jin Jennifer Gossels, Jennifer Rexford, David Walker 1 So6ware- Defined Networking Centralized control with open APIs OpenFlow Applica,on

More information

VSSIM: Virtual Machine based SSD Simulator

VSSIM: Virtual Machine based SSD Simulator 29 th IEEE Conference on Mass Storage Systems and Technologies (MSST) Long Beach, California, USA, May 6~10, 2013 VSSIM: Virtual Machine based SSD Simulator Jinsoo Yoo, Youjip Won, Joongwoo Hwang, Sooyong

More information

PASTE: A Network Programming Interface for Non-Volatile Main Memory

PASTE: A Network Programming Interface for Non-Volatile Main Memory PASTE: A Network Programming Interface for Non-Volatile Main Memory Michio Honda (NEC Laboratories Europe) Giuseppe Lettieri (Università di Pisa) Lars Eggert and Douglas Santry (NetApp) USENIX NSDI 2018

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

MoonGen. A Scriptable High-Speed Packet Generator. Paul Emmerich. January 31st, 2016 FOSDEM Chair for Network Architectures and Services

MoonGen. A Scriptable High-Speed Packet Generator. Paul Emmerich. January 31st, 2016 FOSDEM Chair for Network Architectures and Services MoonGen A Scriptable High-Speed Packet Generator Paul Emmerich January 31st, 216 FOSDEM 216 Chair for Network Architectures and Services Department of Informatics Paul Emmerich MoonGen: A Scriptable High-Speed

More information

NIC-PCIE-4RJ45-PLU PCI Express x4 Quad Port Copper Gigabit Server Adapter (Intel I350 Based)

NIC-PCIE-4RJ45-PLU PCI Express x4 Quad Port Copper Gigabit Server Adapter (Intel I350 Based) NIC-PCIE-4RJ45-PLU PCI Express x4 Quad Port Copper Gigabit Server Adapter (Intel I350 Based) Quad-port Gigabit Ethernet server adapters designed with performance enhancing features and new power management

More information

PASTE: Fast End System Networking with netmap

PASTE: Fast End System Networking with netmap PASTE: Fast End System Networking with netmap Michio Honda, Giuseppe Lettieri, Lars Eggert and Douglas Santry BSDCan 2018 Contact: @michioh, micchie@sfc.wide.ad.jp Code: https://github.com/micchie/netmap/tree/stack

More information

Low Latency 40- and 100-Gbps Ethernet MAC and PHY MegaCore Function User Guide

Low Latency 40- and 100-Gbps Ethernet MAC and PHY MegaCore Function User Guide Low Latency 40- and 100-Gbps Ethernet MAC and PHY MegaCore Function User Guide Subscribe Last updated for Quartus Prime Design Suite: 16.0 UG-01172 101 Innovation Drive San Jose, CA 95134 www.altera.com

More information

An Intelligent NIC Design Xin Song

An Intelligent NIC Design Xin Song 2nd International Conference on Advances in Mechanical Engineering and Industrial Informatics (AMEII 2016) An Intelligent NIC Design Xin Song School of Electronic and Information Engineering Tianjin Vocational

More information

Data Rate Adaptation in EPoC

Data Rate Adaptation in EPoC Data Rate Adaptation in EPoC Marek Hajduczenia, PhD ZTE Corporation marek.hajduczenia@zte.pt Andrea Garavaglia Qualcomm Inc. andreag@qti.qualcomm.com IEEE 802.3bn Orlando, FL 18-21 March 2013 Summary This

More information

CO405H. Department of Compu:ng Imperial College London. Computing in Space with OpenSPL Topic 14: Networking DFEs

CO405H. Department of Compu:ng Imperial College London. Computing in Space with OpenSPL Topic 14: Networking DFEs CO405H Computing in Space with OpenSPL Topic 14: Networking DFEs Oskar Mencer Georgi Gaydadjiev Department of Compu:ng Imperial College London h#p://www.doc.ic.ac.uk/~oskar/ h#p://www.doc.ic.ac.uk/~georgig/

More information

Prof. Hakim Weatherspoon CS 3410, Spring 2015 Computer Science Cornell University. See: Online P&H Chapter 6.5 6

Prof. Hakim Weatherspoon CS 3410, Spring 2015 Computer Science Cornell University. See: Online P&H Chapter 6.5 6 Prof. Hakim Weatherspoon CS 3410, Spring 2015 Computer Science Cornell University See: Online P&H Chapter 6.5 6 Prelim2 Topics Lecture: Lectures 10 to 24 Data and Control Hazards (Chapters 4.7 4.8) RISC/CISC

More information

Atacama: An Open Experimental Platform for Mixed-Criticality Networking on Top of Ethernet

Atacama: An Open Experimental Platform for Mixed-Criticality Networking on Top of Ethernet Atacama: An Open Experimental Platform for Mixed-Criticality Networking on Top of Ethernet Gonzalo Carvajal 1,2 and Sebastian Fischmeister 1 1 University of Waterloo, ON, Canada 2 Universidad de Concepcion,

More information

10GE network tests with UDP. Janusz Szuba European XFEL

10GE network tests with UDP. Janusz Szuba European XFEL 10GE network tests with UDP Janusz Szuba European XFEL Outline 2 Overview of initial DAQ architecture Slice test hardware specification Initial networking test results DAQ software UDP tests Summary 10GE

More information

INT 1011 TCP Offload Engine (Full Offload)

INT 1011 TCP Offload Engine (Full Offload) INT 1011 TCP Offload Engine (Full Offload) Product brief, features and benefits summary Provides lowest Latency and highest bandwidth. Highly customizable hardware IP block. Easily portable to ASIC flow,

More information

10 Gigabit Ethernet 10GBase-W PHY. 1 Introduction. Product Brief Version May 2005

10 Gigabit Ethernet 10GBase-W PHY. 1 Introduction. Product Brief Version May 2005 1 Introduction Initially, 10 Gigabit Ethernet is used by network managers to provide high-speed, local backbone interconnection between large-capacity switches, as it enables Internet Service Providers

More information

Improving Packet Processing Performance of a Memory- Bounded Application

Improving Packet Processing Performance of a Memory- Bounded Application Improving Packet Processing Performance of a Memory- Bounded Application Jörn Schumacher CERN / University of Paderborn, Germany jorn.schumacher@cern.ch On behalf of the ATLAS FELIX Developer Team LHCb

More information

100GE and 40GE PCS Proposal

100GE and 40GE PCS Proposal 100GE and 0GE PCS Proposal Mark Gustlin, Gary Nicholl, Oded Trainin IEEE HSSG September 2007 Supporters Pete Anslow - Nortel Med Belhadj Cortina Brad Booth AMCC Frank Chang - Vitesse Chris Cole Finisar

More information

1GbEth. Access Switch. Workgroup Switch. 10MbEth. Figure 1: Enterprise LAN Topology Example

1GbEth. Access Switch. Workgroup Switch. 10MbEth. Figure 1: Enterprise LAN Topology Example 1 Introduction Ethernet is available in different speeds (10/100/1000 and 10000Mbps) and provides connectivity to meet a wide range of needs from desktop to switches. MorethanIP IP solutions provide a

More information

CN-100 Network Analyzer Product Overview

CN-100 Network Analyzer Product Overview CN-100 Network Analyzer Product Overview CN-100 network analyzers offer an extremely powerful yet cost effective solution for today s complex networking requirements. Test Ethernet or ATM networks with

More information

Open Source Traffic Analyzer

Open Source Traffic Analyzer Open Source Traffic Analyzer Daniel Turull June 2010 Outline 1 Introduction 2 Background study 3 Design 4 Implementation 5 Evaluation 6 Conclusions 7 Demo Outline 1 Introduction 2 Background study 3 Design

More information

Extreme Low Latency 10G Ethernet IP Solution Product Brief (HTK-ELL10G-ETH-FPGA)

Extreme Low Latency 10G Ethernet IP Solution Product Brief (HTK-ELL10G-ETH-FPGA) The 10Gbps Extreme Low Latency Ethernet IP solution offers a fully integrated IEEE802.3 compliant package for NIC (Network Interface Card) and Ethernet switching applications. This industry leading extreme

More information

ntop Users Group Meeting

ntop Users Group Meeting ntop Users Group Meeting PF_RING Tutorial Alfredo Cardigliano Overview Introduction Installation Configuration Tuning Use cases PF_RING Open source packet processing framework for

More information

Programmable NICs. Lecture 14, Computer Networks (198:552)

Programmable NICs. Lecture 14, Computer Networks (198:552) Programmable NICs Lecture 14, Computer Networks (198:552) Network Interface Cards (NICs) The physical interface between a machine and the wire Life of a transmitted packet Userspace application NIC Transport

More information

How to Build a 100 Gbps DDoS Traffic Generator

How to Build a 100 Gbps DDoS Traffic Generator How to Build a 100 Gbps DDoS Traffic Generator DIY with a Single Commodity-off-the-shelf Server (COTS) Surasak Sanguanpong Surasak.S@ku.ac.th DISCLAIMER THE FOLLOWING CONTENTS HAS BEEN APPROVED FOR APPROPIATE

More information

Building Gigabit Interfaces in Altera Transceiver Devices

Building Gigabit Interfaces in Altera Transceiver Devices Building Gigabit Interfaces in Altera Transceiver Devices Course Description In this course, you will learn how you can build high-speed, gigabit interfaces using the 28- nm embedded transceivers found

More information

WAN-compatible 10 Gigabit Ethernet Tutorial. Opticon Burlingame, CA July 31, 2000

WAN-compatible 10 Gigabit Ethernet Tutorial. Opticon Burlingame, CA July 31, 2000 WAN-compatible 10 Gigabit Ethernet Tutorial Opticon 2000 Burlingame, CA July 31, 2000 Carrier and Service Provider Applications Location A 10GbE Metro Metropolitan Optical Networks Dark Fiber Interconnect

More information

Introduction to the OpenCAPI Interface

Introduction to the OpenCAPI Interface Introduction to the OpenCAPI Interface Brian Allison, STSM OpenCAPI Technology and Enablement Speaker name, Title Company/Organization Name Join the Conversation #OpenPOWERSummit Industry Collaboration

More information

Precision and Accuracy of Packet Generators Who tests the testers?

Precision and Accuracy of Packet Generators Who tests the testers? Precision and Accuracy of Packet Generators Who tests the testers? Paul Emmerich, Sebastian Gallenmüller, Alexander Frank, Georg Carle emmericp@net.in.tum.de Technical University of Munich Chair of Network

More information

GUARANTEED END-TO-END LATENCY THROUGH ETHERNET

GUARANTEED END-TO-END LATENCY THROUGH ETHERNET GUARANTEED END-TO-END LATENCY THROUGH ETHERNET Øyvind Holmeide, OnTime Networks AS, Oslo, Norway oeyvind@ontimenet.com Markus Schmitz, OnTime Networks LLC, Texas, USA markus@ontimenet.com Abstract: Latency

More information

How Might Recently Formed System Interconnect Consortia Affect PM? Doug Voigt, SNIA TC

How Might Recently Formed System Interconnect Consortia Affect PM? Doug Voigt, SNIA TC How Might Recently Formed System Interconnect Consortia Affect PM? Doug Voigt, SNIA TC Three Consortia Formed in Oct 2016 Gen-Z Open CAPI CCIX complex to rack scale memory fabric Cache coherent accelerator

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

10 Gigabit Ethernet 10GBase-R PCS Core. 1 Introduction. Product Brief Version August 2004

10 Gigabit Ethernet 10GBase-R PCS Core. 1 Introduction. Product Brief Version August 2004 1 Introduction Initially, 10 Gigabit Ethernet (10 GbE) is used by network managers to provide high-speed, local backbone interconnection between large-capacity switches, as it enables Internet Service

More information

Low Latency 40- and 100-Gbps Ethernet MAC and PHY MegaCore Function User Guide

Low Latency 40- and 100-Gbps Ethernet MAC and PHY MegaCore Function User Guide Low Latency 40- and 100-Gbps Ethernet MAC and PHY MegaCore Function User Guide Subscribe Last updated for Quartus Prime Design Suite: 16.0 UG-01172 101 Innovation Drive San Jose, CA 95134 www.altera.com

More information

A Disseminated Distributed OS for Hardware Resource Disaggregation Yizhou Shan

A Disseminated Distributed OS for Hardware Resource Disaggregation Yizhou Shan LegoOS A Disseminated Distributed OS for Hardware Resource Disaggregation Yizhou Shan, Yutong Huang, Yilun Chen, and Yiying Zhang Y 4 1 2 Monolithic Server OS / Hypervisor 3 Problems? 4 cpu mem Resource

More information

Investigation on Technical Feasibility of FEC Architecture with 1X400Gbps or 4X100Gbps

Investigation on Technical Feasibility of FEC Architecture with 1X400Gbps or 4X100Gbps Investigation on Technical Feasibility of FEC Architecture 1X400Gbps or 4X Xinyuan Wang, Tongtong Wang, Wenbin Yang HUAWEI TECHNOLOGIES CO., LTD. IEEE 802.3bs 400 GbE Task Force Contributor and Supporter

More information

Data Link Layer. Our goals: understand principles behind data link layer services: instantiation and implementation of various link layer technologies

Data Link Layer. Our goals: understand principles behind data link layer services: instantiation and implementation of various link layer technologies Data Link Layer Our goals: understand principles behind data link layer services: link layer addressing instantiation and implementation of various link layer technologies 1 Outline Introduction and services

More information

Moneta: A High-performance Storage Array Architecture for Nextgeneration, Micro 2010

Moneta: A High-performance Storage Array Architecture for Nextgeneration, Micro 2010 Moneta: A High-performance Storage Array Architecture for Nextgeneration, Non-volatile Memories Micro 2010 NVM-based SSD NVMs are replacing spinning-disks Performance of disks has lagged NAND flash showed

More information

International Journal of Engineering Trends and Technology (IJETT) Volume 6 Number 3- Dec 2013

International Journal of Engineering Trends and Technology (IJETT) Volume 6 Number 3- Dec 2013 High Speed Design of Ethernet MAC * Bolleddu Alekya 1 P. Bala Nagu 2 1 PG Student (M. Tech), Dept. of ECE, Chirala Engineering College, Chirala, A.P, India. 2 Associate Professor, Dept. of ECE, Chirala

More information

A HT3 Platform for Rapid Prototyping and High Performance Reconfigurable Computing

A HT3 Platform for Rapid Prototyping and High Performance Reconfigurable Computing A HT3 Platform for Rapid Prototyping and High Performance Reconfigurable Computing Second International Workshop on HyperTransport Research and Application (WHTRA 2011) University of Heidelberg Computer

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