Networks II NS2 tutorialProjects

1
Networks II ---NS-2 tutorial/Projects

• Dustin Jiang
• 06/07/02

The materials are modified from 5th VINT/NS
simulation tutorial/workshop
2
Outline

• NS2 Overview
• NS2 Fundamentals
• Project I overview

3
What is NS2

• Discrete event simulator
• Packet-level
• Link layer and up
• Wired and wireless
• Platforms
• Most UNIX and UNIX-like systems
• Window 95/98/NT
• (Emulation only for FreeBSD for now)

4
Functionality of ns2
5
Resources

• Tcl (Tool Command Language)
• http//dev.scriptics.com/scripting
• OTcl (MIT Object Tcl)
• gammaotcl/doc/tutorial.html (in distribution)
• ns manual
• Included in distribution gamma2ns/doc
• http//www.isi.edu/nsnam/ns/ns-documentation.html
• ns tutorial
• http//perform.wpi.edu/NS/

6
Part I Fundamentals
7
ns Architecture

• Object-oriented (C, OTcl)
• C for data
• Per packet action
• OTcl for control
• Periodic or triggered action

8
OTcl and C The Duality
Pure OTcl objects
Pure C objects
C/OTcl split objects
C
OTcl
ns
9
Extending Tcl Interpreter

• OTcl object-oriented Tcl
• TclCL C and OTcl linkage
• Discrete event scheduler
• Data network components
• Link layer and up
• Emulation support

Tcl
C/C
ns-2
10
Hello World - Batch Mode

• simple.tcl
• set ns new Simulator
• ns at 1 puts \Hello World!\
• ns at 1.5 exit
• ns run
• swallow 74 ns simple.tcl
• Hello World!
• swallow 75

11
Basic tcl

• Command arg1 arg2 arg3
• set a 43
• set b 27
• proc test a b
• set c expr a b
• set d expr expr a - b c
• for set k 0 k lt 10 incr k
• if k lt 5
• puts k lt 5, pow expr pow(d, k)
• else
• puts k gt 5, mod expr d k
• test 43 27

12
Basic OTcl

• Class Mom
• Mom instproc greet
• self instvar age_
• puts age_ years old mom How are you doing?
• Class Kid -superclass Mom
• Kid instproc greet
• self instvar age_
• puts age_ years old kid Whats up, dude?

• set mom new Mom
• mom set age_ 45
• set kid new Kid
• kid set age_ 15
• mom greet
• kid greet

13
Elements of ns-2

• Create the event scheduler
• Turn on tracing
• Create network
• Setup routing
• Insert errors
• Create transport connection
• Create traffic
• Transmit application-level data

14
Creating Event Scheduler

• Create event scheduler
• set ns new Simulator
• Schedule events
• ns at lttimegt lteventgt
• lteventgt any legitimate ns/tcl commands
• ns at 1.0 ftp start
• Start scheduler
• ns run

15
Tracing

• Trace packets on all links
• Open the NAM trace file Open the Trace file
• set nf open out.nam w set tf open
  out.tr w
• ns namtrace-all nf ns
  trace-all tf
• Must appear immediately after creating scheduler
• Turn on tracing on specific links
• ns trace-queue n0 n1
• ns namtrace-queue n0 n1

16
Creating Network

• Nodes
• set n0 ns node
• set n1 ns node
• Links and queuing
• ns duplex-link n0 n1 ltbandwidthgt ltdelaygt
  ltqueue_typegt
• ltqueue_typegt DropTail, RED, CBQ, FQ, SFQ, DRR
• ns duplex-link-op no n1 OPTIONS

n0
n1
17
Setup Routing

• Unicast
• ns rtproto lttypegt
• lttypegt Static, Session, DV, cost, multi-path
• Multicast
• ns multicast (right after new Simulator)
• ns mrtproto lttypegt
• lttypegt CtrMcast, DM, ST, BST

18
Creating Connection UDP

• UDP
• set udp new Agent/UDP
• set null new Agent/Null
• ns attach-agent n0 udp
• ns attach-agent n1 null
• ns connect udp null

UDP
null
n0
n1
19
Creating Traffic On Top of UDP

• CBR
• set src new Application/Traffic/CBR
• src attach-agent udp0
• Exponential or Pareto on-off
• set src new Application/Traffic/Exponential
• set src new Application/Traffic/Pareto
• src attach-agent udp0

CBR
UDP
null
n0
n1
20
Creating Connection TCP

• TCP
• set tcp new Agent/TCP
• set tcpsink new Agent/TCPSink
• ns attach-agent n0 tcp
• ns attach-agent n1 tcpsink
• ns connect tcp tcpsink

TCP
sink
n0
n1
21
Creating Traffic On Top of TCP

• FTP
• set ftp new Application/FTP
• ftp attach-agent tcp
• Telnet
• set telnet new Application/Telnet
• telnet attach-agent tcp

FTP
TCP
sink
n0
n1
22
Post-processing general trace file
cat out.tr grep " 2 3 cbr " grep r column
1 10 awk 'dif 2 - old2 if(dif0) dif
1 if(dif gt 0) printf("d\tf\n", 2, (1 -
old1) / dif) old1 1 old2 2' gt
jitter.txt
23
Post-processing NAM trace file

• Event -t time -s from d to p pType e pSize -c
  flowId -i uid a flowId -x source destination
  seqNo flag sname
• C Ecn Echo P priority A congestion window
  reducded E Congestion Experiented F fast
  start tcp-fs and tcp-int M Ecn-capable-transpor
  t
• r -t 0.156112 -s 1 -d 2 -p tcp -e 1000 -c 2 -i
  177 -a 2 -x 1.0 5.0 52 ------- null
• -t 0.156112 -s 2 -d 3 -p tcp -e 1000 -c 2 -i
  177 -a 2 -x 1.0 5.0 52 ------- null
• d -t 0.156112 -s 2 -d 3 -p tcp -e 1000 -c 2 -i
  177 -a 2 -x 1.0 5.0 52 ------- null
• r -t 0.156184 -s 3 -d 4 -p tcp -e 1000 -c 1 -i
  101 -a 1 -x 0.0 4.0 33 ------- null
• -t 0.156184 -s 4 -d 3 -p ack -e 40 -c 1 -i 184
  -a 1 -x 4.0 0.0 33 ------- null
• - -t 0.156184 -s 4 -d 3 -p ack -e 40 -c 1 -i 184
  -a 1 -x 4.0 0.0 33 ------- null

24
Summary Generic Script Structure

• set ns new Simulator
• Turn on tracing
• Create topology
• Setup packet loss, link dynamics
• Create routing agents
• Create
• - multicast groups
• - protocol agents
• - application and/or setup traffic sources
• Post-processing procs
• Start simulation

25
Visualization Tools

• nam-1 (Network AniMator Version 1)
• Packet-level animation
• Well supported by ns
• xgraph
• Conversion from ns trace to xgraph format

26
nam

• Basic visualization
• Topology layout
• Animation control
• Synchronous replay
• Fine-tune layout
• TCP/SRM visualization
• Editor generate ns simulation scripts

27
ns?nam Interface

• Color
• Node manipulation
• Link manipulation
• Topology layout
• Protocol state
• Misc

28
nam Interface Color

• Color mapping
• ns color 40 red
• ns color 41 blue
• ns color 42 chocolate
• Color ? flow id association
• tcp0 set fid_ 40 red packets
• tcp1 set fid_ 41 blue packets

29
nam Interface Nodes

• Color
• node color red
• Shape (cant be changed after sim starts)
• node shape box circle, box, hexagon
• Marks (concentric shapes)
• ns at 1.0 n0 add-mark m0 blue box
• ns at 2.0 n0 delete-mark m0
• Label (single string)
• ns at 1.1 n0 label \web cache 0\

30
nam Interfaces Links

• Color
• ns duplex-link-op n0 n1 color "green"
• Label
• ns duplex-link-op n0 n1 label "abced"
• Dynamics (automatically handled)
• ns rtmodel Deterministic 2.0 0.9 0.1 n0 n1
• Asymmetric links not allowed

31
nam Interface Topo Layout

• Manual layout specify everything
• ns duplex-link-op n(0) n(1) orient right
• ns duplex-link-op n(1) n(2) orient right
• ns duplex-link-op n(2) n(3) orient right
• ns duplex-link-op n(3) n(4) orient 60deg
• If anything missing ? automatic layout

32
nam Interface Protocol State

• Monitor values of agent variables
• ns add-agent-trace srm0 srm_agent0
• ns monitor-agent-trace srm0
• srm0 tracevar C1_
• srm0 tracevar C2_
• ns delete-agent-trace tcp1

33
nam Interface Misc

• Annotation
• Add textual explaination to your sim
• ns at 3.5 "ns trace-annotate \packet drop\"
• Set animation rate
• ns at 0.0 "ns set-animation-rate 0.1ms"

34
xgraph

• The xgraph program draws a graph on an X display
  given data read from either data files.
• To run it xgraph dataFileName
• You can save the hardcopy of the graph as a
  postscript file.
• Xgraph is available on gamma/usr/local/ns-allinon
  e-2.1b8a/bin/xgraph

35
xgraph

• Format of the input data file
• Device Postscript
• Disposition To File
• FileOrDev /usr/tmp/hpgl.out
• TitleText Sample Xgraph Data
• "Alpha
• -5 29
• -4.9 27.91
• -------
• 5 19
• Notes a blank line is required between two data
  sets.
• "Beta
• -5 33
• -4.9 31.81
• -4.8 30.64
• ------

36
A simulation example
ftp
W32
tcp
sink
n0
n4
5Mb, 15ms
n2
n3
10Mb,2ms
10Mb,2ms
n1
n5
sink
tcp
ftp
W32
37
(No Transcript)
38
(No Transcript)
39
Part II Project I overview
40
Outline

• Queue management mechanisms
• TCP flavors
• Project assignments
• Project options and schedule

41
Fair queue

• Queue maintained at each output port
• Packet placed in queue for its flow
• Round robin servicing
• Skip empty queues
• Can have weighted fair queuing

42
RED Overview

• Detects congestion using avg. queue size
• Drop probability is a function of the avg. queue
  size. min_th, max_th, p_max
• No bias against bursty traffic
• No global synchronization
• A flows drop rate is proportional to its share
  of the bandwidth through the gateway

43
RED Approach

• RED queue management, roughly
• for each packet arrival
• calculate the new average queue size avg
• if min_th lt avg gtmax_th
• calculate probability pa
• with probability pa
• mark/drop the arriving packet
• else if max_th lt avg
• drop the arriving packet
• Variables
• avg average queue size
• Pa packet marking/dropping probability
• Parameters
• Min_th minimum threshold for queue
• Max_th maximum threshold for queue

44
RED Drawbacks

• There is no TCP-awareness
• Dropping packets from flows in proportion to
  their bandwidth doesnt result in fair sharing.
  e.g Fast link vs. a Slow link
• Does not consider the number of flows
• Discrimination against
• Large RTT flows w.r.t. Small RTT flows
• Adaptive flows w.r.t. Non-adaptive flows

45
SRED Stabilized RED

• Avg. Q size is not used for drop probability
• Does statistical estimation of active flows
• Drop probability is
  f (instantaneous Q size, active flows)
• Stabilizes buffer occupation at a level
  independent of the number of active flows
• Identifies misbehaving flows

46
SRED Approach

• Hit(t) 0 if no hit 1 if hit
• Hit frequency P at an instant t is defined as
  P(t) (1
  a)P(t 1) a Hit(t), 0 lt a lt 1
• a p/M, p is overwrite prob, M is flow table
  size
• psred(q) pmax if B/3 lt q lt B,
• pmax/4 if B/6 lt q lt B/3,
• 0 if 0 lt q lt B/6
  
  
  
• pzap psred(q) min (1, 1/(256 P(t))2)

47
FRED Flow RED

• Uses per-active-flow accounting
• Each flows loss rate depends on the flows
  buffer use
• Protects adaptive flows
• Isolates non-adaptive greedy flows
• Uses the RED approach with modifications

48
FRED Approach

• Maintain a flow-state table, one entry per active
  flow with fields qleni and strikei
• Parameters minq, maxq, avgcq
• Identify non-adaptive flows as
• qleni gt maxq
• avg gt maxth qleni gt 2 avgcq
• qleni gt avgcq strike gt 1
• Drop only from robust flows qleni gt minq

49
Comparison of various queue policies

• Fairness over time or over traffic load
• Average queue length over time or over traffic
  load
• Queue delay over time or over traffic load
• Packet less ratio over time or over traffic load
• Link utilization over time or over traffic load
• FQ (Fair Queue)
• RED Random Early Detection
• One of the following
• SRED Stabilized RED
• FRED
• REM
• BLUE
• AVQ Active Virtual Queue
• Or any kind of adaptive queue management scheme.

50
TCP Agents

• ns has several variants of TCP available
• -- Agent/TCP a tahoe'' TCP sender
• -- Agent/TCP/Reno a Reno'' TCP sender
• -- Agent/TCP/NewReno Reno with a modification
• -- Agent/TCP/Sack1 TCP with selective repeat
  (follows RFC2018)
• -- Agent/TCP/Vegas TCP Vegas
• -- Agent/TCP/Fack Reno TCP with forward
  acknowledge ment''
• The oneway TCP receiving agents currently
  supported are
• -- Agent/TCPSink TCP sink with one ACK per
  packet
• -- Agent/TCPSink/DelAck TCP sink with
  configurable delay per ACK
• -- Agent/TCPSink/Sack1 selective ACK sink
  (follows RFC2018)
• -- Agent/TCPSink/Sack1/DelAck Sack1 with DelAck
• The twoway experimental sender currently supports
  only a Reno form of TCP
• -- Agent/TCP/FullTcp

51
TCP Flavors

• TCP-Tahoe
• W1 adaptation on congestion
• TCP-Reno
• WW/2 adaptation on fast retransmit, W1 on
  timeout
• TCP-newReno
• TCP-Reno intelligent fast recovery
• TCP-Vegas, TCP-SACK

52
TCP Tahoe

• Slow-start
• Congestion control upon time-out or DUP-ACKs
• When the sender receives 3 duplicate ACKs for the
  same sequence number, sender infers a loss
• Congestion window reduced to 1 and slow-start
  performed again
• Simple
• Congestion control too aggressive

53
TCP Reno

• Tahoe
• Packet loss detected both through timeouts, and
  through DUP-ACKs
• Sender reduces window by half, the ssthresh is
  set to half of current window, and congestion
  avoidance is performed (window increases only by
  1 every round-trip time)
• Fast recovery ensures that pipe does not become
  empty
• Window cut-down to 1 (and subsequent slow-start)
  performed only on time-out

54
TCP New-Reno

• TCP-Reno with more intelligence during fast
  recovery
• In TCP-Reno, the first partial ACK will bring the
  sender out of the fast recovery phase
• Results in timeouts when there are multiple
  losses
• In TCP New-Reno, partial ACK is taken as an
  indication of another lost packet (which is
  immediately retransmitted).
• Sender comes out of fast recovery only after all
  outstanding packets (at the time of first loss)
  are ACKed

55
TCP SACK

• TCP (Tahoe, Reno, and New-Reno) uses cumulative
  acknowledgements
• When there are multiple losses, TCP Reno and
  New-Reno can retransmit only one lost packet per
  round-trip time
• What about TCP-Tahoe?
• SACK enables receiver to give more information
  to sender about received packets allowing sender
  to recover from multiple-packet losses faster and
  more efficiently

56
TCP SACK (Example)

• Assume packets 5-25 are transmitted
• Let packets 5, 12, and 18 be lost
• Receiver sends back a CACK5, and
  SACK(6-11,13-17,19-25)
• Sender knows that packets 5, 12, and 18 are lost
  and retransmits them immediately

57
Other TCP flavors

• TCP Vegas
• Uses round-trip time as an early-congestion-feedba
  ck mechanism
• Reduces losses
• TCP FACK
• Intelligently uses TCP SACK information to
  optimize the fast recovery mechanism further

58
TCP Agent Parameters

• Common configuration parameters and defaults for
  TCP agents
• Agent/TCP set window 20 max bound on window
  size
• Agent/TCP set windowInit 1 initial/reset
  value of cwnd
• Agent/TCP set windowOption 1 cong avoid
  algorithm (1 standard)
• Agent/TCP set windowConstant 4 used only when
  windowOption ! 1
• Agent/TCP set windowThresh 0.002 used in
  computing averaged window
• Agent/TCP set overhead 0 !0 adds random time
  between sends
• Agent/TCP set ecn 0 TCP should react to ecn
  bit
• Agent/TCP set packetSize 1000 packet size
  used by sender (bytes)
• Agent/TCP set bugFix true see documentation
• Agent/TCP set slowstartrestart true see
  documentation

59
TCP Agent Parameters

• Agent/TCP set tcpTick 0.1 timer granularity
  in sec (.1 is NONST ANDARD)
• Agent/TCP set maxrto 64 bound on RTO
  (seconds)
• Agent/TCP set dupacks 0 duplicate ACK counter
  
• Agent/TCP set ack 0 highest ACK received
• Agent/TCP set cwnd 0 congestion window
  (packets)
• Agent/TCP set awnd 0 averaged cwnd
  (experimental)
• Agent/TCP set ssthresh 0 slowstat threshold
  (packets)
• Agent/TCP set rtt 0 rtt sample
• Agent/TCP set srtt 0 smoothed (averaged) rtt
• Agent/TCP set rttvar 0 mean deviation of rtt
  samples
• Agent/TCP set backoff 0 current RTO backoff
  factor
• Agent/TCP set maxseq 0 max (packet) seq
  number sent

60
TCP Sink Agents

• Sinks for oneway TCP senders
• Types
• -- standard sinks, delayedACK sinks, SACK sinks
• Standard sinks
• -- generate one ACK per packet received
• -- ACK number overloaded in sequence number''
  packet field
• DelayedACK sinks
• -- same as standard, but with variable delay
  added between
• -- time to delay ACKs specified in seconds
• SACK sinks
• ACKs
• -- generates additional information for SACK
  capable sender
• -- configurable maxSackBlocks parameter

61
TwoWay TCP (FullTCP'')

• most TCP objects are oneway (and require a
  source/sink pair)
• real TCP can be bidirectional
• simultaneous two way data transfer alters TCP
  dynamics considerably
• the TCP/FullTcp agent
• -- follows closely to Reno'' TCP implementation
  in 4.4 BSD
• -- byte oriented transfers
• -- two way data supported
• -- most of the connection establishment/teardown
• -- symmetric only one agent type used for both
  sides

62
FullTCP Parameters

• Parameters and defaults
• Agent/TCP/FullTcp set segsperack 1 segs
  received before generating ACK
• Agent/TCP/FullTcp set segsize 536 segment
  size (MSS size for bulk xfers)
• Agent/TCP/FullTcp set tcprexmtthresh 3
  dupACKs thresh to trigger fast rexmt
• Agent/TCP/FullTcp set iss 0 initial send
  sequence number
• Agent/TCP/FullTcp set nodelay false disable
  senderside Nagle algorithm
• Agent/TCP/FullTcp set dataonsyn false send
  data on initial SYN?
• Agent/TCP/FullTcp set dupsegfix true avoid
  fast rxt due to dup segsacks
• Agent/TCP/FullTcp set dupackreset false
  reset dupACK ctr on !0 len data seg s containing
  dup ACKs
• Agent/TCP/FullTcp set interval 0.1 delayed
  ACK interval

63
Comparison of various TCP Flavors

• Fairness over time and over traffic load
• Responsiveness over time and over traffic load
• Link utilization over time and over traffic load
• Packet drop ratio over time and over traffic load
  
• Packet retransmission ratio over time and over
  traffic load
• Average queue size over time and over traffic
  load
• Tahoe (TCP with the slow start, congestion
  avoidance, and fast retransmit algorithm)
• Reno (Tahoe plus fast recovery algorithm)
• New Reno with ECN
• SACK (selective Acknowledgements)
• FACK (forward Acknowledgements)
• Rate-Halving with ECN

64
Topology
65
Final Report

• Full implementation design
• Comprehensive set of results
• Two separated reports
• Code
• Refer to the project description on course
  website for more information.