Throughput Issues for High-Speed Wide-Area Networks
|
|
- Junior Banks
- 7 years ago
- Views:
Transcription
1 Throughput Issues for High-Speed Wide-Area Networks Brian L. Tierney ESnet Lawrence Berkeley National Laboratory (HEPIX Oct 30, 2009)
2 Why does the Network seem so slow?
3 TCP Performance Issues Getting good TCP performance over highlatency high-bandwidth networks is not easy! You must keep the TCP window full the size of the window is directly related to the network latency Easy to compute max throughput: Throughput = buffer size / latency eg: 64 KB buffer / 40 ms path = 1.6 KBytes (12.8 Kbits) / sec
4 Setting the TCP buffer sizes It is critical to use the optimal TCP send and receive socket buffer sizes for the link you are using. Recommended size to fill the pipe 2 x Bandwidth Delay Product (BDP) Recommended size to leave some bandwidth for others around 20% of (2 x BPB) =.4 * BDP Default TCP buffer sizes are way too small for today s high speed networks Until recently, default TCP send/receive buffers were typically 64 KB tuned buffer to fill LBL to BNL link: 10 MB 150X bigger than the default buffer size with default TCP buffers, you can only get a small % of the available bandwidth!
5 Buffer Size Example Optimal Buffer size formula: buffer size = 20% * (2 * bandwidth * RTT) ping time (RTT) = 50 ms Narrow link = 500 Mbps (62 MBytes/sec) e.g.: the end-to-end network consists of all 1000 BT ethernet and OC-12 (622 Mbps) TCP buffers should be:.05 sec * 62 * 2 * 20% = 1.24 MBytes
6 Sample Buffer Sizes LBL to (50% of pipe) SLAC (RTT = 2 ms, narrow link = 1000 Mbps) : 256 KB BNL: (RTT = 80 ms, narrow link = 1000 Mbps): 10 MB CERN: (RTT = 165 ms, narrow link = 1000 Mbps): 20.6 MB Note: default buffer size is usually only 64 KB, and default maximum autotuning buffer size for is often only 256KB e.g.: FreeBSD 7.2 Autotuning default max = 256 KB times too small! Home DSL, US to Europe (RTT = 150, narrow link = 2 Mbps): 38 KB Default buffers are OK.
7 Socket Buffer Autotuning To solve the buffer tuning problem, based on work at LANL and PSC, Linux OS added TCP Buffer autotuning Sender-side TCP buffer autotuning introduced in Linux 2.4 Receiver-side autotuning added in Linux 2.6 Most OS s now include TCP autotuning TCP send buffer starts at 64 KB As the data transfer takes place, the buffer size is continuously re-adjusted up max autotune size Current OS Autotuning default maximum buffers Linux 2.6: 256K to 4MB, depending on version FreeBSD 7: 256K Windows Vista: 16M Mac OSX 10.5: 8M
8 Autotuning Settings Linux 2.6 net.core.rmem_max = net.core.wmem_max = # autotuning min, default, and max number of bytes to use net.ipv4.tcp_rmem = net.ipv4.tcp_wmem = FreeBSD 7.0 net.inet.tcp.sendbuf_auto=1 net.inet.tcp.recvbuf_auto=1 net.inet.tcp.sendbuf_max= net.inet.tcp.recvbuf_max= Windows Vista netsh interface tcp set global autotunninglevel=normal max buffer fixed at 16MB OSX 10.5 ( Self-Tuning TCP) kern.ipc.maxsockbuf= For more info, see:
9 More Problems: TCP congestion control Path = LBL to CERN (Geneva) OC-3, (in 2000), RTT = 150 ms average BW = 30 Mbps
10 Parallel Streams Can Help
11 Parallel Streams Issues Potentially unfair Places more load on the end hosts But they are necessary when you don t have root access, and can t convince the sysadmin to increase the max TCP buffers graph from Tom Dunigan, ORNL
12 TCP Response Function Well known fact that TCP Reno does not scale to highspeed networks Average TCP congestion window = 1.2 p segments p = packet loss rate see: What this means: For a TCP connection with 1500-byte packets and a 100 ms round-trip time filling a 10 Gbps pipe would require a congestion window of 83,333 packets, a packet drop rate of at most one drop every 5,000,000,000 packets. requires at most one packet loss every 6000s, or 1h:40m to keep the pipe full
13 Proposed TCP Modifications Many proposed alternate congestion control algorithms: BIC/CUBIC HTCP: (Hamilton TCP) Scalable TCP Compound TCP And several more
14 Linux Results 800 TCP Results Note that BIC reaches Max throughput MUCH faster Linux 2.6, BIC TCP Linux 2.4 Linux 2.6, BIC off RTT = 67 ms time slot (5 second intervals)
15 Comparison of Various TCP Congestion Control Algorithms
16 Selecting TCP Congestion Control in Linux To determine current configuration: sysctl -a grep congestion net.ipv4.tcp_congestion_control = cubic net.ipv4.tcp_available_congestion_control = cubic reno Use /etc/sysctl.conf to set to any available congested congestion control. Supported options (may need to enabled by default in your kernel): CUBIC, BIC, HTCP, HSTCP, STCP, LTCP, more..
17 Summary so far To optimize TCP throughput, do the following: Use a newer OS that supports TCP buffer autotuning Increase the maximum TCP autotuning buffer size Use a few parallel streams if possible Use a modern congestion control algorithm
18 PERFSONAR
19 High Performance Networking Most of the R&E community has access to 10 Gbps networks. Naive users with the right tools should be able to easily get: 200 Mbps/per stream between properly maintained systems 2 Gbps aggregate rates between significant computing resources Most users are not experiencing this level of performance There is widespread frustration involved in the transfer of large data sets between facilities, or from the data s facility of origin back to the researcher s home institution. From the BES network requirements workshop: doc/bes Net Req Workshop 2007 Final Report.pdf We can increase performance by measuring the network and reporting problems!
20 Network Troubleshooting is a Multi-Domain Problem Source Campus Education Backbone Destination Campus S D JET Network Regional
21 Where are common problems? Source Campus Congested or faulty links between domains Education Backbone Latency dependant problems inside domains with small RTT Destination Campus S D JET Network Regional
22 Soft Network Failures Soft failures are where basic connectivity functions, but high performance is not possible. TCP was intentionally designed to hide all transmission errors from the user: As long as the TCPs continue to function properly and the internet system does not become completely partitioned, no transmission errors will affect the users. (From IEN 129, RFC 716) Some soft failures only affect high bandwidth long RTT flows. Hard failures are easy to detect & fix, soft failures can lie hidden for years.
23 Common Soft Failures Small Queue Tail Drop Switches not able to handle the long packet trains prevalent in long RTT sessions and cross traffic at the same time Un-intentional Rate Limiting Process Switching on Cisco 6500 devices due to faults, acl s, or mis-configuration Security Devices E.g.: 10X improvement by turning off Cisco Reflexive ACL Random Packet Loss Bad fibers or connectors Low light levels due to amps/interfaces failing Duplex mismatch
24 Local testing will not find all problems Source Campus S Performance is poor when RTT exceeds 20 ms R&E Backbone Performance is good when RTT is < 20 ms Destination Campus D Regional Regional Switch with small buffers
25 Addressing the Problem: perfsonar Developing an open web-services based framework for collecting, managing and sharing network measurements Deploying the framework across the science community Encouraging people to deploy known good measurement points near domain boundaries Using the framework to find & correct soft network failures.
26 perfsonar Deployments Internet2 ESnet Argonne National Lab Brookhaven National Lab Fermilab National Energy Research Scientific Computing Center Pacific Northwest National Lab University of Michigan, Ann Arbor Indiana University Boston University University of Texas Arlington Oklahoma University, Norman Michigan Information Technology Center William & Mary University of Wisconsin Madison Southern Methodist University, Dallas University of Texas Austin Vanderbilt University APAN GLORIAD JGN2PLUS KISTI Korea Monash University, Melbourne NCHC, HsinChu, Taiwan Simon Fraser University GEANT GARR HUNGARNET PIONEER SWITCH CCIN2P3 CERN CNAF DE-KIT NIKHEF/SARA PIC RAL TRIUMF
27 perfsonar Architecture Interoperable network measurement middleware (SOA): Modular Web services-based Decentralized Locally controlled Integrates: Network measurement tools and data archives Data manipulation Information Services Discovery Topology Authentication and authorization Based on: Open Grid Forum Network (OGF) Network Measurement Working Group (NM- WG) schema Formalizing the specification of perfsonar protocols in the OGF NMC-WG Network topology description work in the OGF NML-WG
28 Main perfsonar Services Lookup Service gls Global service used to find services hls Home service for registering local perfsonar metadata Measurement Archives SNMP MA Interface Data psb MA -- Scheduled bandwidth and latency data Measurement Points BWCTL OWAMP PINGER Troubleshooting Tools NDT NPAD Topology Service
29 PerfSONAR-PS Deployment Status Currently deployed in over 130 locations 45 bwctl and owamp servers 40 psb MA 15 SNMP MA 5 gls and 135 hls US Atlas Deployment Monitoring all Tier 1 to Tier 2 connections For current list of services, see: 29
30 Deploying a perfsonar measurement host in under 30 minutes Using the PS Performance Toolkit is very simple Boot from CD Use command line tool to configure what disk partition to use for persistent data Network address and DNS User and root passwords Use Web GUI to configure Select which services to run Select remote measurement points for bandwidth and latency tests Configure Cacti to collect SNMP data from key router interfaces 30
31 SAMPLE RESULTS
32 Example: US Atlas LHC Tier 1 to Tier 2 Center Data transfer problem Couldn t exceed 1 Gbps across a 10GE end to end path that included 5 administrative domains Used perfsonar tools to localize problem Identified problem device An unrelated domain had leaked a full routing table to the router for a short time causing FIB corruption. The routing problem was fixed, but the router started process switching some flows after that. Fixed Rebooting device fixed the symptoms of the problem Better BGP filters configured to prevent reoccurrence 32
33 Sample Results: Fixed a soft failure
34 Sample Results: Bulk Data Transfer between DOE Supercomputer Centers Users were having problems moving data between supercomputer centers One user was: waiting more than an entire workday for a 33 GB input file perfsonar Measurement tools were installed Regularly scheduled measurements were started Numerous choke points were identified & corrected Dedicate wide area transfer nodes were setup Tuned for Wide Area Transfers Now moving 40 TB in less than 3 days 34
35 Importance of Regular Testing You cant wait for users to report problems and then fix them Soft failure can go unreported for years Problems that get fixed have a way of coming back System defaults coming back after hardware/software upgrade New employees not knowing why the previous employee set things up that way perfsonar makes is easy to collect, archive, and alert on throughput information
36 Example: NERSC <-> ORNL Testing
37 More Information Also see:
Network Monitoring with the perfsonar Dashboard
Network Monitoring with the perfsonar Dashboard Andy Lake Brian Tierney ESnet Advanced Network Technologies Group TIP2013 Honolulu HI January 15, 2013 Overview perfsonar overview Dashboard history and
More informationDeploying distributed network monitoring mesh
Deploying distributed network monitoring mesh for LHC Tier-1 and Tier-2 sites Phil DeMar, Maxim Grigoriev Fermilab Joe Metzger, Brian Tierney ESnet Martin Swany University of Delaware Jeff Boote, Eric
More informationTCP Tuning Techniques for High-Speed Wide-Area Networks. Wizard Gap
NFNN2, 20th-21st June 2005 National e-science Centre, Edinburgh TCP Tuning Techniques for High-Speed Wide-Area Networks Distributed Systems Department Lawrence Berkeley National Laboratory http://gridmon.dl.ac.uk/nfnn/
More informationPath Monitoring Tools Deployment Planning for U.S. T123. Jeff Boote Internet2/R&D. July 22, 2008 Jt Techs, UNL, Lincoln NE
Path Monitoring Tools Deployment Planning for U.S. T123 Jeff Boote Internet2/R&D July 22, 2008 Jt Techs, UNL, Lincoln NE Implementation Considerations Constraints 1. Different LHC participants are interested
More informationLHC Path Monitoring Tools. Jeff Boote Internet2/R&D. May 27, 2008 US ATLAS T2/T3 Workshop at UM
LHC Path Monitoring Tools Deployment Planning Jeff Boote Internet2/R&D May 27, 2008 US ATLAS T2/T3 Workshop at UM Overview Value of network performance monitoring and diagnostic tools Recommended tools
More informationTier3 Network Issues. Richard Carlson May 19, 2009 rcarlson@internet2.edu
Tier3 Network Issues Richard Carlson May 19, 2009 rcarlson@internet2.edu Internet2 overview Member organization with a national backbone infrastructure Campus & Regional network members National and International
More informationTCP loss sensitivity analysis ADAM KRAJEWSKI, IT-CS-CE
TCP loss sensitivity analysis ADAM KRAJEWSKI, IT-CS-CE Original problem IT-DB was backuping some data to Wigner CC. High transfer rate was required in order to avoid service degradation. 4G out of... 10G
More informationOSG CAMPUS INFRASTRUCTURES SERIES
OSG CAMPUS INFRASTRUCTURES SERIES Methods for Network Troubleshooting in Distributed High Throughput Environments --- Webinar Shawn McKee OSG Networking / University of Michigan Physics January 25, 2013
More informationHigh Performance Bulk Data Transfer
High Performance Bulk Data Transfer Brian Tierney and Joe Metzger, ESnet Joint Techs, Columbus OH, July, 2010 Why does the network seem so slow? Wizard Gap Slide from Matt Mathis, PSC Today s Talk This
More informationGARUDA - NKN Partner's Meet 2015 Big data networks and TCP
GARUDA - NKN Partner's Meet 2015 Big data networks and TCP Brij Kishor Jashal Email brij.jashal@tifr.res.in Garuda-NKN meet 10 Sep 2015 1 Outline: Scale of LHC computing ( as an example of Big data network
More informationNetwork performance monitoring Insight into perfsonar
Network performance monitoring Insight into perfsonar Szymon Trocha, Poznań Supercomputing and Networking Center E-infrastructure Autumn Workshops, Chisinau, Moldova 9 September 2014 Agenda! Network performance
More informationScience DMZs Understanding their role in high-performance data transfers
Science DMZs Understanding their role in high-performance data transfers Chris Tracy, Network Engineer Eli Dart, Network Engineer ESnet Engineering Group Overview Bulk Data Movement a common task Pieces
More informationperfsonar: End-to-End Network Performance Verification
perfsonar: End-to-End Network Performance Verification Toby Wong Sr. Network Analyst, BCNET Ian Gable Technical Manager, Canada Overview 1. IntroducGons 2. Problem Statement/Example Scenario 3. Why perfsonar?
More informationIntegration of Network Performance Monitoring Data at FTS3
Integration of Network Performance Monitoring Data at FTS3 July-August 2013 Author: Rocío Rama Ballesteros Supervisor(s): Michail Salichos Alejandro Álvarez CERN openlab Summer Student Report 2013 Project
More informationComparing the Network Performance of Windows File Sharing Environments
Technical Report Comparing the Network Performance of Windows File Sharing Environments Dan Chilton, Srinivas Addanki, NetApp September 2010 TR-3869 EXECUTIVE SUMMARY This technical report presents the
More informationDICE Diagnostic Service
DICE Diagnostic Service Joe Metzger metzger@es.net Joint Techs Measurement Working Group January 27 2011 Background Each of the DICE collaborators are defining and delivering services to their users. A
More informationIperf Tutorial. Jon Dugan <jdugan@es.net> Summer JointTechs 2010, Columbus, OH
Iperf Tutorial Jon Dugan Summer JointTechs 2010, Columbus, OH Outline What are we measuring? TCP Measurements UDP Measurements Useful tricks Iperf Development What are we measuring? Throughput?
More informationIntroduction & Motivation
Introduction & Motivation WACREN Network Monitoring and Measurement Workshop Antoine Delvaux a.delvaux@man.poznan.pl perfsonar developer 30.09.2015 This document is a result of work by the perfsonar Project
More informationNetest: A Tool to Measure the Maximum Burst Size, Available Bandwidth and Achievable Throughput
Netest: A Tool to Measure the Maximum Burst Size, Available Bandwidth and Achievable Throughput Guojun Jin Brian Tierney Distributed Systems Department Lawrence Berkeley National Laboratory 1 Cyclotron
More informationperfsonar Overview Jason Zurawski, ESnet zurawski@es.net Southern Partnerships for Advanced Networking November 3 rd 2015
perfsonar Overview Jason Zurawski, ESnet zurawski@es.net Southern Partnerships for Advanced Networking November 3 rd 2015 This document is a result of work by the perfsonar Project (http://www.perfsonar.net)
More informationLHCONE Site Connections
LHCONE Site Connections Michael O Connor moc@es.net ESnet Network Engineering Asia Tier Center Forum on Networking Daejeon, South Korea September 23, 2015 Outline Introduction ESnet LHCONE Traffic Volumes
More informationApplication of perfsonar architecture in support of GRID monitoring
Application of perfsonar architecture in support of GRID monitoring Binczewski A, Lawenda M, Łapacz R, Trocha Sz INGRID 2007, S.Margherita Ligure, Italy Agenda Start point Network monitoring perfsonar
More informationMarch 10 th 2011, OSG All Hands Workshop Network Performance Jason Zurawski, Internet2. Performance Measurement Tools
March 10 th 2011, OSG All Hands Workshop Network Performance Jason Zurawski, Internet2 Performance Measurement Tools Agenda Tutorial Agenda: Network Performance Primer Why Should We Care? (15 Mins) GeMng
More informationQuestion: 3 When using Application Intelligence, Server Time may be defined as.
1 Network General - 1T6-521 Application Performance Analysis and Troubleshooting Question: 1 One component in an application turn is. A. Server response time B. Network process time C. Application response
More informationNetwork Testing Tools. Ettore Tamiro GARR - Network Operation Center PAP Workshop July 15, 2009
Network Testing Tools Ettore Tamiro GARR - Network Operation Center PAP Workshop July 15, 2009 This tutorial is based on a previous work by For the Rich Carlson NWS Workshop Jan 12-14, 2009 Installing
More informationTCP Labs. WACREN Network Monitoring and Measurement Workshop Antoine Delvaux a.delvaux@man.poznan.pl perfsonar developer 30.09.
TCP Labs WACREN Network Monitoring and Measurement Workshop Antoine Delvaux a.delvaux@man.poznan.pl perfsonar developer 30.09.2015 Hands-on session We ll explore practical aspects of TCP Checking the effect
More informationImplementing perfsonar in the South African National Research and Education Network
Implementing perfsonar in the South African National Research and Education Network Kevin DRAAI 1,2, Roderick MOOI 1,3 1 Council for Scientific and Industrial Research, P.O. Box 395, Pretoria 0001, South
More informationEVALUATING NETWORK BUFFER SIZE REQUIREMENTS
EVALUATING NETWORK BUFFER SIZE REQUIREMENTS for Very Large Data Transfers Michael Smitasin Lawrence Berkeley National Laboratory (LBNL) Brian Tierney Energy Sciences Network (ESnet) [ 2 ] Example Workflow
More informationNetwork Middleware Solutions
Network Middleware: Lambda Station, TeraPaths, Phoebus Matt Crawford GLIF Meeting; Seattle, Washington October 1-2, 2008 Lambda Station (I) Target: last-mile problem between local computing resources and
More informationE-Center: A Collaborative Platform for Wide Area Network Users
FERMILAB-CONF-12-827-CD International Conference on Computing in High Energy and Nuclear Physics 2012 (CHEP2012) IOP Publishing E-Center: A Collaborative Platform for Wide Area Network Users M Grigoriev
More informationI N T E R I M R E P O R T O N N E T W O R K P E R F O R M A N C E
I N T E R I M R E P O R T O N N E T W O R K P E R F O R M A N C E Document Filename: Activity: Partner(s): Lead Partner: Document classification: SA2 KTH, EENet, KBFI, VU, RTU, ITPA, NICH BNTU, UIIP NASB
More informationHow To Set Up A Network Measurement Toolkit For A Network Performance Test On A Network With A Network (Networking) On A Microsoft Ipa 2.5 (Netware) On An Ipa2 (Netcom) On Your Computer
Firewall Port Recommendations for the ps Performance Toolkit Prepared by the NTAC Performance Working Group November 2014 Edited by J. Zurawski (ESnet), A. Brown (Internet2), A. Lake (ESnet), K. Miller
More informationperfsonar MDM release 3.0 - Product Brief
perfsonar MDM release 3.0 - Product Brief In order to provide the fast, reliable and uninterrupted network communication that users of the GÉANT 2 research networks rely on, network administrators must
More informationProposal for a perfsonar Multi Domain Monitoring Service for LHCOPN
Proposal for a perfsonar Multi Domain Monitoring Service for LHCOPN Service Specification Version 10 Proposal for a perfsonar Multi Domain Monitoring Service for LHCOPN Document History Product/Version
More informationHands on Workshop. Network Performance Monitoring and Multicast Routing. Yasuichi Kitamura NICT Jin Tanaka KDDI/NICT APAN-JP NOC
Hands on Workshop Network Performance Monitoring and Multicast Routing Yasuichi Kitamura NICT Jin Tanaka KDDI/NICT APAN-JP NOC July 18th TEIN2 Site Coordination Workshop Network Performance Monitoring
More informationApplication Note. Windows 2000/XP TCP Tuning for High Bandwidth Networks. mguard smart mguard PCI mguard blade
Application Note Windows 2000/XP TCP Tuning for High Bandwidth Networks mguard smart mguard PCI mguard blade mguard industrial mguard delta Innominate Security Technologies AG Albert-Einstein-Str. 14 12489
More informationEnd-to-End Network/Application Performance Troubleshooting Methodology
End-to-End Network/Application Performance Troubleshooting Methodology Wenji Wu, Andrey Bobyshev, Mark Bowden, Matt Crawford, Phil Demar, Vyto Grigaliunas, Maxim Grigoriev, Don Petravick Fermilab, P.O.
More informationIntroduction to perfsonar
Introduction to perfsonar Loukik Kudarimoti, DANTE 27 th September, 2006 SEEREN2 Summer School, Heraklion Overview of this talk Answers to some basic questions The need for Multi-domain monitoring What
More informationD1.2 Network Load Balancing
D1. Network Load Balancing Ronald van der Pol, Freek Dijkstra, Igor Idziejczak, and Mark Meijerink SARA Computing and Networking Services, Science Park 11, 9 XG Amsterdam, The Netherlands June ronald.vanderpol@sara.nl,freek.dijkstra@sara.nl,
More informationAchieving the Science DMZ
Achieving the Science DMZ Eli Dart, Network Engineer ESnet Network Engineering Group Joint Techs, Winter 2012 Baton Rouge, LA January 22, 2012 Outline of the Day Motivation Services Overview Science DMZ
More informationperfsonar Multi-Domain Monitoring Service Deployment and Support: The LHC-OPN Use Case
perfsonar Multi-Domain Monitoring Service Deployment and Support: The LHC-OPN Use Case Fausto Vetter, Domenico Vicinanza DANTE TNC 2010, Vilnius, 2 June 2010 Agenda Large Hadron Collider Optical Private
More informationVisualizations and Correlations in Troubleshooting
Visualizations and Correlations in Troubleshooting Kevin Burns Comcast kevin_burns@cable.comcast.com 1 Comcast Technology Groups Cable CMTS, Modem, Edge Services Backbone Transport, Routing Converged Regional
More informationCampus Network Design Science DMZ
Campus Network Design Science DMZ Dale Smith Network Startup Resource Center dsmith@nsrc.org The information in this document comes largely from work done by ESnet, the USA Energy Sciences Network see
More informationAugust 9 th 2011, OSG Site Admin Workshop Jason Zurawski Internet2 Research Liaison NDT
August 9 th 2011, OSG Site Admin Workshop Jason Zurawski Internet2 Research Liaison NDT Agenda Tutorial Agenda: Network Performance Primer Why Should We Care? (30 Mins) IntroducOon to Measurement Tools
More informationHigh-Speed TCP Performance Characterization under Various Operating Systems
High-Speed TCP Performance Characterization under Various Operating Systems Y. Iwanaga, K. Kumazoe, D. Cavendish, M.Tsuru and Y. Oie Kyushu Institute of Technology 68-4, Kawazu, Iizuka-shi, Fukuoka, 82-852,
More informationUsing Fuzzy Logic Control to Provide Intelligent Traffic Management Service for High-Speed Networks ABSTRACT:
Using Fuzzy Logic Control to Provide Intelligent Traffic Management Service for High-Speed Networks ABSTRACT: In view of the fast-growing Internet traffic, this paper propose a distributed traffic management
More informationMonitoring the US ATLAS Network Infrastructure with perfsonar-ps
Monitoring the US ATLAS Network Infrastructure with perfsonar-ps For the ATLAS Collaboration, Shawn McKee 1, Andrew Lake 2, Philippe Laurens 3, Horst Severini 4, Tomasz Wlodek 5, Stephen Wolff 6 and Jason
More informationRFC 6349 Testing with TrueSpeed from JDSU Experience Your Network as Your Customers Do
RFC 6349 Testing with TrueSpeed from JDSU Experience Your Network as Your Customers Do RFC 6349 is the new transmission control protocol (TCP) throughput test methodology that JDSU co-authored along with
More informationDynamic Circuit Network (DCN) / perfsonar Shared Infrastructure
Dynamic Circuit Network (DCN) / perfsonar Shared Infrastructure Tom Lehman USC/ISI Network Cyberinfrastructure Application Classes: Bulk Transport 2-Way Interactive Video Real-Time Communications others.
More informationperfsonar MDM updates for LHCONE: VRF monitoring, updated web UI, VM images
perfsonar MDM updates for LHCONE: VRF monitoring, updated web UI, VM images Domenico Vicinanza DANTE, Cambridge, UK perfsonar MDM Product Manager domenico.vicinanza@dante.net LHCONE Meeting Oslo 20-21
More informationTCP tuning guide for distributed application on wide area networks 1.0 Introduction
TCP tuning guide for distributed application on wide area networks 1.0 Introduction Obtaining good TCP throughput across a wide area network usually requires some tuning. This is especially true in high-speed
More informationGlobus Striped GridFTP Framework and Server. Raj Kettimuthu, ANL and U. Chicago
Globus Striped GridFTP Framework and Server Raj Kettimuthu, ANL and U. Chicago Outline Introduction Features Motivation Architecture Globus XIO Experimental Results 3 August 2005 The Ohio State University
More informationRTT 60.5 msec receiver window size: 32 KB
Real-World ARQ Performance: TCP Ex.: Purdue UCSD Purdue (NSL): web server UCSD: web client traceroute to planetlab3.ucsd.edu (132.239.17.226), 30 hops max, 40 byte packets 1 switch-lwsn2133-z1r11 (128.10.27.250)
More informationESnet Support for WAN Data Movement
ESnet Support for WAN Data Movement Eli Dart, Network Engineer ESnet Science Engagement Group Joint Facilities User Forum on Data Intensive Computing Oakland, CA June 16, 2014 Outline ESnet overview Support
More informationNetwork monitoring with perfsonar. Duncan Rand Imperial College London
Network monitoring with perfsonar Duncan Rand Imperial College London A little history: Gridmon Cast your minds back to GridPP16 at QMUL about 6 years ago; we saw a not too dissimilar network monitoring
More informationOverview of Network Measurement Tools
Overview of Network Measurement Tools Jon M. Dugan Energy Sciences Network Lawrence Berkeley National Laboratory NANOG 43, Brooklyn, NY June 1, 2008 Networking for the Future of Science
More informationMarch 10 th 2011, OSG All Hands Mee6ng, Network Performance Jason Zurawski Internet2 NDT
March 10 th 2011, OSG All Hands Mee6ng, Network Performance Jason Zurawski Internet2 NDT Agenda Tutorial Agenda: Network Performance Primer Why Should We Care? (15 Mins) GeNng the Tools (10 Mins) Use of
More informationMeasuring IP Performance. Geoff Huston Telstra
Measuring IP Performance Geoff Huston Telstra What are you trying to measure? User experience Responsiveness Sustained Throughput Application performance quality Consistency Availability Network Behaviour
More informationFebruary 9 th 2010, AIMS 2010 Ma4 Zekauskas, ma4@internet2.edu
February 9 th 2010, AIMS 2010 Ma4 Zekauskas, ma4@internet2.edu Outline Mo?va?on What is perfsonar? Who is involved in perfsonar? Who is adop?ng perfsonar? Why might perfsonar be useful to you For more
More informationQuantifying the Performance Degradation of IPv6 for TCP in Windows and Linux Networking
Quantifying the Performance Degradation of IPv6 for TCP in Windows and Linux Networking Burjiz Soorty School of Computing and Mathematical Sciences Auckland University of Technology Auckland, New Zealand
More informationTransport Layer Protocols
Transport Layer Protocols Version. Transport layer performs two main tasks for the application layer by using the network layer. It provides end to end communication between two applications, and implements
More informationApplications. Network Application Performance Analysis. Laboratory. Objective. Overview
Laboratory 12 Applications Network Application Performance Analysis Objective The objective of this lab is to analyze the performance of an Internet application protocol and its relation to the underlying
More informationAchieving Mainframe-Class Performance on Intel Servers Using InfiniBand Building Blocks. An Oracle White Paper April 2003
Achieving Mainframe-Class Performance on Intel Servers Using InfiniBand Building Blocks An Oracle White Paper April 2003 Achieving Mainframe-Class Performance on Intel Servers Using InfiniBand Building
More informationFinal for ECE374 05/06/13 Solution!!
1 Final for ECE374 05/06/13 Solution!! Instructions: Put your name and student number on each sheet of paper! The exam is closed book. You have 90 minutes to complete the exam. Be a smart exam taker -
More informationTCP AND ATM IN WIDE AREA NETWORKS
TCP AND ATM IN WIDE AREA NETWORKS Benjamin J. Ewy, Joseph B. Evans, Victor S. Frost, Gary J. Minden Telecommunications & Information Sciences Laboratory Department of Electrical Engineering & Computer
More informationDeploying distributed network monitoring mesh for LHC Tier-1 and Tier-2 sites
Deploying distributed network monitoring mesh for LHC Tier-1 and Tier-2 sites Jeff Boote 1, Eric Boyd 1, Aaron Brown 1, Maxim Grigoriev 2, Joe Metzger 3, Phil DeMar 2, Martin Swany 4, Brian Tierney 3,
More informationNetwork testing with iperf
Network testing with iperf Bill Farrow bill.farrow@gmail.com Why use iperf? What does it run on? TCP example UDP example Use Case: Network Server Appliance Use Case: Embedded Video Streaming Use Case:
More informationVoice Quality Measurement in perfsonar
Proceedings of the Asia-Pacific Advanced Network 2010 v. 30, p. 104-109 http://dx.doi.org/10.7125/apan.30.15 ISSN 2227-3026 Voice Quality Measurement in perfsonar Pin-Hsuan Chen 1, *, Che-Nan Yang 2 NCHC
More informationPerformance Measurement of Wireless LAN Using Open Source
Performance Measurement of Wireless LAN Using Open Source Vipin M Wireless Communication Research Group AU KBC Research Centre http://comm.au-kbc.org/ 1 Overview General Network Why Network Performance
More informationImproving our Evaluation of Transport Protocols. Sally Floyd Hamilton Institute July 29, 2005
Improving our Evaluation of Transport Protocols Sally Floyd Hamilton Institute July 29, 2005 Computer System Performance Modeling and Durable Nonsense A disconcertingly large portion of the literature
More informationTCP Adaptation for MPI on Long-and-Fat Networks
TCP Adaptation for MPI on Long-and-Fat Networks Motohiko Matsuda, Tomohiro Kudoh Yuetsu Kodama, Ryousei Takano Grid Technology Research Center Yutaka Ishikawa The University of Tokyo Outline Background
More informationChapter 1 Reading Organizer
Chapter 1 Reading Organizer After completion of this chapter, you should be able to: Describe convergence of data, voice and video in the context of switched networks Describe a switched network in a small
More informationCCNA R&S: Introduction to Networks. Chapter 5: Ethernet
CCNA R&S: Introduction to Networks Chapter 5: Ethernet 5.0.1.1 Introduction The OSI physical layer provides the means to transport the bits that make up a data link layer frame across the network media.
More informationCisco Application Networking for IBM WebSphere
Cisco Application Networking for IBM WebSphere Faster Downloads and Site Navigation, Less Bandwidth and Server Processing, and Greater Availability for Global Deployments What You Will Learn To address
More informationHigh Performance Data-Transfers in Grid Environment using GridFTP over InfiniBand
High Performance Data-Transfers in Grid Environment using GridFTP over InfiniBand Hari Subramoni *, Ping Lai *, Raj Kettimuthu **, Dhabaleswar. K. (DK) Panda * * Computer Science and Engineering Department
More informationNetwork Performance and Possible Improvement For Medical Demonstration
Network Performance and Possible Improvement For Medical Demonstration APAN-JP/KDDI/NICT/TEIN2-JP Jin Tanaka Medical Workshop in 27 th APAN Kaohsiung Meeting March 5 2009 Kaohsiung Taiwan 1 Who am I? Jin
More informationTop-Down Network Design
Top-Down Network Design Chapter Nine Developing Network Management Strategies Copyright 2010 Cisco Press & Priscilla Oppenheimer 29 Network Management Design A good design can help an organization achieve
More informationVMWARE WHITE PAPER 1
1 VMWARE WHITE PAPER Introduction This paper outlines the considerations that affect network throughput. The paper examines the applications deployed on top of a virtual infrastructure and discusses the
More informationNetwork Performance Tools
Network Performance Tools Jeff Boote Internet2/R&D June 1, 2008 NANOG 43/ Brooklyn, NY Overview BWCTL OWAMP NDT/NPAD BWCTL: What is it? A resource alloca=on and scheduling daemon for arbitra=on of iperf
More informationUsing Linux Traffic Control on Virtual Circuits J. Zurawski Internet2 zurawski@internet2.edu February 25 nd 2013
Using Linux Traffic Control on Virtual Circuits J. Zurawski Internet2 zurawski@internet2.edu February 25 nd 2013 1. Abstract Research and Education (R&E) networks have experimented with the concept of
More informationIntroduc)on & Mo)va)on
Introduc)on & Mo)va)on This document is a result of work by the perfsonar Project (hdp://www.perfsonar.net) and is licensed under CC BY- SA 4.0 (hdps://crea)vecommons.org/licenses/by- sa/4.0/). Event Presenter,
More informationNetwork Probe. Figure 1.1 Cacti Utilization Graph
Network Probe Description The MCNC Client Network Engineering group will install several open source network performance management tools on a computer provided by the LEA or charter school to build a
More informationNetwork congestion control using NetFlow
Network congestion control using NetFlow Maxim A. Kolosovskiy Elena N. Kryuchkova Altai State Technical University, Russia Abstract The goal of congestion control is to avoid congestion in network elements.
More informationAllocating Network Bandwidth to Match Business Priorities
Allocating Network Bandwidth to Match Business Priorities Speaker Peter Sichel Chief Engineer Sustainable Softworks psichel@sustworks.com MacWorld San Francisco 2006 Session M225 12-Jan-2006 10:30 AM -
More informationHow To Connect To Bloomerg.Com With A Network Card From A Powerline To A Powerpoint Terminal On A Microsoft Powerbook (Powerline) On A Blackberry Or Ipnet (Powerbook) On An Ipnet Box On
Transport and Security Specification 15 July 2015 Version: 5.9 Contents Overview 3 Standard network requirements 3 Source and Destination Ports 3 Configuring the Connection Wizard 4 Private Bloomberg Network
More informationUsing LiveAction Software for Successful VoIP Deployments How to quickly and accurately deploy QoS for VoIP networks
LiveAction Application Note Using LiveAction Software for Successful VoIP Deployments How to quickly and accurately deploy QoS for VoIP networks September 2012 http://www.actionpacked.com Table of Contents
More informationANI Network Testbed Update
ANI Network Testbed Update Brian Tierney, ESnet, Joint Techs, Columbus OH, July, 2010 ANI: Advanced Network Initiative Project Start Date: September, 2009 Funded by ARRA for 3 years Designed, built, and
More informationProcedure: You can find the problem sheet on Drive D: of the lab PCs. 1. IP address for this host computer 2. Subnet mask 3. Default gateway address
Objectives University of Jordan Faculty of Engineering & Technology Computer Engineering Department Computer Networks Laboratory 907528 Lab.4 Basic Network Operation and Troubleshooting 1. To become familiar
More informationLinux TCP Implementation Issues in High-Speed Networks
Linux TCP Implementation Issues in High-Speed Networks D.J.Leith Hamilton Institute, Ireland www.hamilton.ie 1. Implementation Issues 1.1. SACK algorithm inefficient Packets in flight and not yet acknowledged
More informationPerformance Analysis of IPv4 v/s IPv6 in Virtual Environment Using UBUNTU
Performance Analysis of IPv4 v/s IPv6 in Virtual Environment Using UBUNTU Savita Shiwani Computer Science,Gyan Vihar University, Rajasthan, India G.N. Purohit AIM & ACT, Banasthali University, Banasthali,
More informationImproving Scientific Outcomes at the APS with a Science DMZ
Improving Scientific Outcomes at the APS with a Science DMZ Jason Zurawski zurawski@es.net Science Engagement Engineer, ESnet Lawrence Berkeley National Laboratory GlobusWorld 2015 April 15 th, 2015 Outline
More informationCleaning Encrypted Traffic
Optenet Documentation Cleaning Encrypted Traffic Troubleshooting Guide iii Version History Doc Version Product Date Summary of Changes V6 OST-6.4.300 01/02/2015 English editing Optenet Documentation
More informationNetwork measurement II. Sebastian Castro NZRS 27 th May 2015 Victoria University
Network measurement II Sebastian Castro NZRS 27 th May 2015 Victoria University Agenda Network measurement in the wild Common pitfalls in methodologies Network measurement datasets Stories Our work AMP
More informationWiFiPerf User Guide 1.5
WiFiPerf User Guide 1.5 AccessAgility LLC 2012 AccessAgility LLC. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any
More information4 High-speed Transmission and Interoperability
4 High-speed Transmission and Interoperability Technology 4-1 Transport Protocols for Fast Long-Distance Networks: Comparison of Their Performances in JGN KUMAZOE Kazumi, KOUYAMA Katsushi, HORI Yoshiaki,
More informationPerformance Analysis of Switches in High Speed LAN
Performance Analysis of Switches in High Speed LAN Muhammad Imran, Asif Shaikh, Syed S. Rizvi, Khaled M. Elleithy Computer Science and Engineering Department, University of Bridgeport, Bridgeport, CT {mimran,
More informationUse of Alternate Path WAN Circuits at Fermilab
Use of Alternate Path WAN Circuits at Fermilab Phil DeMar, Andrey Bobyshev, Matt Crawford, Vyto Grigaliunas Fermilab, PO BOX 500, Batavia, IL 60510, USA demar@fnal.gov Abstract. Fermilab hosts the American
More informationCCNA Discovery 4.0.3.0 Networking for Homes and Small Businesses Student Packet Tracer Lab Manual
4.0.3.0 Networking for Homes and Small Businesses Student Packet Tracer Lab Manual This document is exclusive property of Cisco Systems, Inc. Permission is granted to print and copy this document for non-commercial
More informationTCP Window Size for WWAN Jim Panian Qualcomm
TCP Window Size for WWAN Jim Panian Qualcomm June 17, 2004 PCCA 1 TCP Window Size for WWAN Role of TCP Windowing Calculation of Optimum Window Size WWAN Window Size Requirements WWAN and Long Fat Networks
More informationUsing Netflow data for forecasting
Using Netflow data for forecasting Les Cottrell SLAC and Fawad Nazir NIIT, Presented at the CHEP06 Meeting, Mumbai India, February 2006 www.slac.stanford.edu/grp/scs/net/talk06/icfachep06.ppt Partially
More information