03/12/08Nuova Systems Inc. Page 1 TCP Issues in the Data Center Tom Lyon The Future of TCP: Train-wreck or Evolution? Stanford University 2008-04-01.

Slides:



Advertisements
Similar presentations
Congestion Control and Fairness Models Nick Feamster CS 4251 Computer Networking II Spring 2008.
Advertisements

Computer Networking Lecture 20 – Queue Management and QoS.
1 TCP Congestion Control. 2 TCP Segment Structure source port # dest port # 32 bits application data (variable length) sequence number acknowledgement.
Congestion Control Created by M Bateman, A Ruddle & C Allison As part of the TCP View project.
Congestion Control: TCP & DC-TCP Swarun Kumar With Slides From: Prof. Katabi, Alizadeh et al.
Congestion Control An Overview -Jyothi Guntaka. Congestion  What is congestion ?  The aggregate demand for network resources exceeds the available capacity.
Presenter - Eric Wang CS577 Advanced Computer Networks
Congestion control in data centers
Defense: Christopher Francis, Rumou duan Data Center TCP (DCTCP) 1.
Open Issues on TCP for Mobile Computing Ibrahim Matta Computer Science, Boston University Vassilis Tsaoussidis Computer Science, Northeastern University.
SEDCL: Stanford Experimental Data Center Laboratory.
High-performance bulk data transfers with TCP Matei Ripeanu University of Chicago.
1 TCP Transport Control Protocol Reliable In-order delivery Flow control Responds to congestion “Nice” Protocol.
ACN: Congestion Control1 Congestion Control and Resource Allocation.
Lecture 5: Congestion Control l Challenge: how do we efficiently share network resources among billions of hosts? n Last time: TCP n This time: Alternative.
L13: Sharing in network systems Dina Katabi Spring Some slides are from lectures by Nick Mckeown, Ion Stoica, Frans.
Jennifer Rexford Fall 2014 (TTh 3:00-4:20 in CS 105) COS 561: Advanced Computer Networks TCP.
On Horrible TCP Performance over Underwater Links Balaji Prabhakar Abdul Kabbani, Balaji Prabhakar Stanford University.
5/8/2006 Nicole SAN Protocols 1 Storage Networking Protocols Nicole Opferman CS 526.
Ch. 28 Q and A IS 333 Spring Q1 Q: What is network latency? 1.Changes in delay and duration of the changes 2.time required to transfer data across.
IWARP Ethernet Key to Driving Ethernet into the Future Brian Hausauer Chief Architect NetEffect, Inc.
ICTCP: Incast Congestion Control for TCP in Data Center Networks∗
CS332 Ch. 28 Spring 2014 Victor Norman. Access delay vs. Queuing Delay Q: What is the difference between access delay and queuing delay? A: I think the.
The Effects of Systemic Packets Loss on Aggregate TCP Flows Thomas J. Hacker May 8, 2002 Internet 2 Member Meeting.
Curbing Delays in Datacenters: Need Time to Save Time? Mohammad Alizadeh Sachin Katti, Balaji Prabhakar Insieme Networks Stanford University 1.
3: Transport Layer3b-1 Principles of Congestion Control Congestion: r informally: “too many sources sending too much data too fast for network to handle”
Transport Layer 4 2: Transport Layer 4.
Detail: Reducing the Flow Completion Time Tail in Datacenter Networks SIGCOMM PIGGY.
These materials are licensed under the Creative Commons Attribution-Noncommercial 3.0 Unported license (
CS540/TE630 Computer Network Architecture Spring 2009 Tu/Th 10:30am-Noon Sue Moon.
Raj Jain The Ohio State University R1: Performance Analysis of TCP Enhancements for WWW Traffic using UBR+ with Limited Buffers over Satellite.
Trends In Network Industry - Exploring Possibilities for IPAC Network Steven Lo.
The NE010 iWARP Adapter Gary Montry Senior Scientist
CSE679: Computer Network Review r Review of the uncounted quiz r Computer network review.
EE 122: Congestion Control and Avoidance Kevin Lai October 23, 2002.
TCP Trunking: Design, Implementation and Performance H.T. Kung and S. Y. Wang.
Vertical Optimization Of Data Transmission For Mobile Wireless Terminals MICHAEL METHFESSEL, KAI F. DOMBROWSKI, PETER LANGENDORFER, HORST FRANKENFELDT,
Infiniband Bart Taylor. What it is InfiniBand™ Architecture defines a new interconnect technology for servers that changes the way data centers will be.
Knowing when to deploy DAS NAS and SAN can optimize t Availability t Scalability t Performance t Manageability t Cost effectiveness.
SANs Today Increasing port count Multi-vendor Edge and Core switches
Ethernet. Ethernet standards milestones 1973: Ethernet Invented 1983: 10Mbps Ethernet 1985: 10Mbps Repeater 1990: 10BASE-T 1995: 100Mbps Ethernet 1998:
Mr. P. K. GuptaSandeep Gupta Roopak Agarwal
TCP continued. Discussion – TCP Throughput TCP will most likely generate the saw tooth type of traffic. – A rough estimate is that the congestion window.
© Janice Regan, CMPT 128, CMPT 371 Data Communications and Networking Congestion Control 0.
Thoughts on the Evolution of TCP in the Internet Sally Floyd PFLDnet 2004 February 16, 2004.
CS640: Introduction to Computer Networks Aditya Akella Lecture 15 TCP Congestion Control.
CIS679: TCP and Multimedia r Review of last lecture r TCP and Multimedia.
TCP/IP1 Address Resolution Protocol Internet uses IP address to recognize a computer. But IP address needs to be translated to physical address (NIC).
1 Three ways to (ab)use Multipath Congestion Control Costin Raiciu University Politehnica of Bucharest.
Day 13 Intro to MANs and WANs. MANs Cover a larger distance than LANs –Typically multiple buildings, office park Usually in the shape of a ring –Typically.
Chapter 28 Q and A IS 333 Spring A quiz question Q: What is network latency? 1.Changes in delay and duration of the changes 2.time required to transfer.
ICTCP: Incast Congestion Control for TCP in Data Center Networks By: Hilfi Alkaff.
1 Building big router from lots of little routers Nick McKeown Assistant Professor of Electrical Engineering and Computer Science, Stanford University.
Congestion Control.
UNIT-V Transport Layer protocols for Ad Hoc Wireless Networks
Congestion Control and Resource Allocation
Introduction to Networks
Flow and Congestion Control
Queuing and Queue Management
Storage Networking Protocols
So far, On the networking side, we looked at mechanisms to links hosts using direct linked networks and then forming a network of these networks. We introduced.
RDMA over Commodity Ethernet at Scale
Lecture 16, Computer Networks (198:552)
Congestion Control Reasons:
TCP Congestion Control
Beyond FTP & hard drives: Accelerating LAN file transfers
Congestion Control (from Chapter 05)
Congestion Control and Resource Allocation
Lecture 5, Computer Networks (198:552)
Lecture 6, Computer Networks (198:552)
Presentation transcript:

03/12/08Nuova Systems Inc. Page 1 TCP Issues in the Data Center Tom Lyon The Future of TCP: Train-wreck or Evolution? Stanford University

03/12/08Nuova Systems Inc. Page 2 TCP: Not Just for “The Internet”  Essentially all network software relies on TCP/IP semantics  “The network is the data center”  In the data center, gigabits are “free”  10 5 times cheaper than WAN bandwidth  Terabit class switches  10Gb endpoints  TCP needs:  High bandwidth  Low Latency  Predictability & Fairness

03/12/08Nuova Systems Inc. Page 3 Storage Networks  Storage Access slowly evolving from hardware bus to open network  NAS vs SAN  NFS & CIFS vs SCSI's many flavors  Ethernet vs Fibre Channel vs Infiniband

03/12/08Nuova Systems Inc. Page 4 Storage Networks: Ethernet vs EtherNot  iSCSI, NFS, CIFS  TCP & Ethernet  Congestion Loss  Stream Oriented  Software Transport  High CPU overhead  SCSI-FCP, SCSI-SRP  F.C. and Infiniband  Credit Flow Control  Block Oriented  Hardware Transport  Low CPU overhead

03/12/08Nuova Systems Inc. Page 5 Storage Networks: Convergence  Data Center Ethernet  Choice of congestion classes  Lossy vs lossless  Choice of storage transports  TCP or F.C. (FCOE)‏  Choice of hardware or software transport  TOE w TCP, software FCOE,...

03/12/08Nuova Systems Inc. Page 6 TCP: Time Out of Joint  TCP was standardized in a much slower world  ½ Second minimum retransmit timeout  20 micro-second RTT achievable today!  Fast re-transmit algorithm only works for streams – more data being sent  Most data center traffic is request/response – often single packets  Packet loss hurts because TCP won't (not can't) respond fast enough

03/12/08Nuova Systems Inc. Page 7 Congestion in the Data Center  Gigantic, non-blocking switches are the norm  Hundreds of ports, terabits of throughput  Buffers and buffer management are the most costly part of the switch  Link based flow control (“pause”) allows switch to push congestion back to its upstream neighbors  If the upstream neighbor is the source server, then the congestion “Goes away”  Or does it?

03/12/08Nuova Systems Inc. Page 8 Servers and Gigabits  Any current x86 server can easily saturate a 1Gb Ethernet link with TCP traffic  Many current servers can saturate 10Gb Ethernet links!  Lossless classes cause the pipe to fill faster  What happens when the first hop, the server's own Ethernet link, is the point of congestion?

03/12/08Nuova Systems Inc. Page 9 TCP and the Fat Pipe  If TCP doesn't “see” congestion (loss or ECN) then it will continue to increase its window to try to get more bandwidth in the network  Lossless network => high throughput  But... a single streaming connection will consume all available buffers  Newer connections will have a hard time getting buffers => extreme unfairness  The server needs good congestion management

03/12/08Nuova Systems Inc. Page 10 Servers, Ethernet, and Queues  “Everyone” knows that big, simple FIFO queues are a bad idea in routers  What do servers have today? - big, simple FIFO queues!  The queues are owned and maintained by the Ethernet NIC hardware  Horrible unfairness can be demonstrated with only 2 TCP connections  Many servers deal with 1000s of TCP connections

03/12/08Nuova Systems Inc. Page 11 Connection Size vs Throughput – idle 1G link

03/12/08Nuova Systems Inc. Page 12 Connection Size vs Throughput – busy 1G link – competing with a single “hog” connection UNFAIR!

03/12/08Nuova Systems Inc. Page 13 Improving Server Congestion Management  Omitted due to event rules!

03/12/08Nuova Systems Inc. Page 14 TCP: Rock or Hard Place?  With lossy Ethernet, TCP bandwidth can collapse due to stupidly high timeouts  => Unpredictable performance  With lossless Ethernet, TCP fairness can collapse due to stupid queuing policies  => Unpredictable performance  Data Center Managers hate unpredictability  Ethernet standards have evolved, TCP needs to catch up  TCP and Ethernet implementations must improve

03/12/08Nuova Systems Inc. Page 15 Why does this matter?  The Earth is being paved by data centers  Google, Microsoft, NSA, Walmart, Facebook,...  Improving TCP means more overall efficiency in the data center  Heat, CO 2, and radioactive waste are becoming measurable by-products of TCP inefficiency  Fix TCP => Save the World!