Changes between Version 29 and Version 30 of MEBSandbox/UDTExample


Ignore:
Timestamp:
05/25/12 17:03:38 (12 years ago)
Author:
Mark Berman
Comment:

Moved out of sandbox.

Legend:

Unmodified
Added
Removed
Modified
  • MEBSandbox/UDTExample

    v29 v30  
    1 [[PageOutline]]
    2 = Sample Experiment: UDT Evaluation =
     1= Page Moved =
    32
    4 == Purpose ==
    5 
    6 In this experiment, you will compare the performance of UDT (UDP-based Data Transfer) to TCP (transmission control protocol) in a simple network under varying conditions.
    7 
    8 This experiment is based on work carried out at the University of Missouri - Kansas City, by Sunae Shin, Kaustubh Dhondge, and Baek-Young Choi. Their paper "Understanding the Performance of TCP and UDP-based Data Transfer Protocols using EMULAB" was presented at the First GENI Research and Educational Experiment Workshop (GREE2012), March 15-16, 2012, Los Angeles, CA.
    9 
    10 == Prerequisites ==
    11 
    12 Before beginning this experiment, you should be prepared with the following.
    13 
    14  * You have GENI credentials to obtain GENI resources. (If not, see **link**).
    15  * You are able to use '''Flack''' to request GENI resources.  (If not, see [https://www.protogeni.net/trac/protogeni/wiki/FlackTutorial the Flack tutorial]).
    16  * You are comfortable using '''ssh''' and executing basic commands using a UNIX shell.
    17 
    18 == Setup ==
    19 
    20  * Download the attached [attachment:udt.rspec?format=raw rspec file] and save it on your machine. (Make sure to save in raw format.)
    21  * Start Flack, create a new slice, load rspec ''udt.rspec'' and submit for sliver creation. Your sliver should look something like this:
    22 
    23 [[Image(UDTExperimentSliver.png, 50%)]]
    24 
    25 == Techniques ==
    26 
    27 You will use the following techniques during this experiment.
    28 
    29 === File Transfer Using UDT ===
    30 
    31 Follow these steps to perform a file transfer using UDT.
    32 
    33  * Log into pc1 and pc2 in separate windows.
    34 
    35  * On pc1, start a UDT file transfer server, using this command:
    36 {{{
    37 % pc1:~% /local/udt4/app/sendfile
    38 server is ready at port: 9000
    39 }}}
    40 
    41  * On pc2, start a UDT file transfer client, using this command:
    42 {{{
    43 pc2:~% /local/udt4/app/recvfile pc1 9000 /local/datafiles/sm.10M /dev/null
    44 }}}
    45 You should see output like the following in your pc1 window, showing the results of the file transfer. Note the transfer rate.
    46 {{{
    47 new connection: 192.168.2.2:55839
    48 speed = 7.14472Mbits/sec
    49 }}}
    50 
    51  * There are three data files available for transfer tests: '''/local/datafiles/sm.10M''' is 10MB, '''/local/datafiles/med.100M''' is 100MB, and '''/local/datafiles/lg.1G''' is 1000MB. Leave your transfer server running on pc1, and try transferring each of these files in turn by typing the appropriate commands on pc2. Keep track of the transfer rates in each case.
    52 
    53  * You can leave your UDT server running or stop it with ctrl-C.
    54 
    55 === File Transfer Using FTP ===
    56 
    57 Follow these steps to perform a file transfer using FTP.
    58 
    59  * For a TCP-based (FTP) transfer, there's already a FTP server running on pc1. Log into pc2, and start an ftp client:
    60 
    61 (You type ''ftp pc1'', the user name ''anonymous'', and any password you want, although your e-mail address is traditional.)
    62 
    63 {{{
    64 pc2:~% ftp pc1
    65 Connected to PC1-lan1.
    66 220 (vsFTPd 2.3.2)
    67 Name (pc1:mberman): anonymous
    68 331 Please specify the password.
    69 Password:
    70 230 Login successful.
    71 Remote system type is UNIX.
    72 Using binary mode to transfer files.
    73 ftp>
    74 }}}
    75 
    76  * Still on pc2, request a file transfer. Note the reported file size, transfer time, and transfer rate.
    77 
    78 {{{
    79 ftp> get med.100M /dev/null
    80 local: /dev/null remote: med.100M
    81 200 PORT command successful. Consider using PASV.
    82 150 Opening BINARY mode data connection for med.100M (104857600 bytes).
    83 226 Transfer complete.
    84 104857600 bytes received in 8.91 secs (11491.9 kB/s)
    85 }}}
    86 
    87  * You can perform additional transfers with additional get commands. When you're done, exit the ftp client with the quit command.
    88 
    89 === Viewing and Adjusting link characteristics ===
    90 
    91 In this experiment, you'll be changing the characteristics of the link and measuring how they affect UDT and TCP performance.
    92 
    93  * Log into your delay node as you do with any other node. Then, on your delay node, use this command:
    94 {{{
    95 %sudo ipfw pipe show
    96 }}}
    97 
    98 You'll get something like this:
    99 {{{
    100 60111: 100.000 Mbit/s    1 ms   50 sl. 1 queues (1 buckets) droptail
    101     mask: 0x00 0x00000000/0x0000 -> 0x00000000/0x0000
    102 BKT Prot ___Source IP/port____ ____Dest. IP/port____ Tot_pkt/bytes Pkt/Byte Drp
    103   0 ip    207.167.175.72/0       195.123.216.8/6        7     1060  0    0   0
    104 60121: 100.000 Mbit/s    1 ms   50 sl. 1 queues (1 buckets) droptail
    105     mask: 0x00 0x00000000/0x0000 -> 0x00000000/0x0000
    106 BKT Prot ___Source IP/port____ ____Dest. IP/port____ Tot_pkt/bytes Pkt/Byte Drp
    107   0 ip   207.167.176.224/0         195.124.8.8/6        8     1138  0    0   0
    108 }}}
    109 
    110 This information shows the internal configuration of the "pipes" used to emulate network characteristics.
    111 (Your output may look different, depending on the version of ipfw installed on your delay node.
    112 In any case, the information you need is on the first line of output for each pipe.)
    113 
    114 You'll want to make note of the two pipe numbers, one for each direction of traffic along your link.
    115 In the example above, they are 60111 and 60121.
    116 
    117 There are three link characteristics we'll manipulate in this experiment: bandwidth, delay, and packet loss rate.
    118 You'll find their values listed in the ipfw output above.
    119 The link bandwidth appears on the first line immediately after the pipe number. It's 100Mbps in the example shown above.
    120 The next value shown is the delay, 1 ms in the example above.
    121 The packet loss rate (PLR) is omitted if it's zero, as shown above. If non-zero, you'll see something like '''plr 0.000100'''
    122 immediately after the "50 sl." on the first output line.
    123 
    124 It is possible to adjust the parameters of the two directions of your link separately, to emulate asymmetric links.
    125 In this experiment, however, we are looking at symmetric links, so we'll always change the settings on both pipes together.
    126 
    127 Here are the command sequences you'll need to change your link parameters.
    128 In each case, you'll need to provide the correct pipe numbers, if they're different from the example.
    129 
    130  * To change bandwidth (100M means 100Mbits/s):
    131 {{{
    132 sudo ipfw pipe 60111 config bw 100M
    133 sudo ipfw pipe 60121 config bw 100M
    134 }}}
    135 
    136  * Request a bandwidth of zero to use the full capacity of the link (unlimited):
    137 {{{
    138 sudo ipfw pipe 60111 config bw 0
    139 sudo ipfw pipe 60121 config bw 0
    140 }}}
    141 
    142  * To change link delay (delays are measured in ms):
    143 {{{
    144 sudo ipfw pipe 60111 config delay 10
    145 sudo ipfw pipe 60121 config delay 10
    146 }}}
    147 
    148  * To change packet loss rate (rate is a probability, so 0.001 means 0.1% packet loss):
    149 {{{
    150 sudo ipfw pipe 60111 config plr .0001
    151 sudo ipfw pipe 60121 config plr .0001
    152 }}}
    153 
    154  * You can combine settings for bandwidth, delay, and loss by specifying more than one in a single ipfw command. We'll use this form in the procedure below.
    155 
    156 == Experiment Procedure ==
    157 
    158  * Set your link parameters to use maximum bandwidth, no delay, no packet loss:
    159 
    160 {{{
    161 sudo ipfw pipe 60111 config bw 0 delay 0 plr 0
    162 sudo ipfw pipe 60121 config bw 0 delay 0 plr 0
    163 }}}
    164 
    165  * Verify with
    166 
    167 {{{
    168 sudo ipfw pipe show
    169 60111: unlimited    0 ms   50 sl. 1 queues (1 buckets) droptail
    170     mask: 0x00 0x00000000/0x0000 -> 0x00000000/0x0000
    171 BKT Prot ___Source IP/port____ ____Dest. IP/port____ Tot_pkt/bytes Pkt/Byte Drp
    172   0 ip    207.167.175.72/0       195.123.216.8/6        7     1060  0    0   0
    173 60121: unlimited    0 ms   50 sl. 1 queues (1 buckets) droptail
    174     mask: 0x00 0x00000000/0x0000 -> 0x00000000/0x0000
    175 BKT Prot ___Source IP/port____ ____Dest. IP/port____ Tot_pkt/bytes Pkt/Byte Drp
    176   0 ip   207.167.176.224/0         195.124.8.8/6        8     1138  0    0   0
    177 }}}
    178 
    179 Note that bandwidth is set to ''unlimited'', delay to ''0 ms'', and no PLR is shown.
    180 
    181  * Using this initial setting, try a few UDT transfers, including the larger files. Now try FTP transfers. Record the transfer sizes and rates.
    182 
    183  * Now change the link parameters to reduce the available bandwidth to 10Mbps:
    184 
    185 {{{
    186 sudo ipfw pipe 60111 config bw 10M delay 0 plr 0
    187 sudo ipfw pipe 60121 config bw 10M delay 0 plr 0
    188 }}}
    189 
    190  * Repeat your file transfers with the new settings. As before, note the transfer sizes and rates, as well as the link settings.
    191 
    192  * Continue with additional trials, varying each of the three link parameters over a range sufficient to observe meaningful performance differences. Record your data.
    193 
    194 == What to hand in ==
    195 
    196  1. Your raw data and appropriate graphs illustrating changes in performance for the two transfer protocols with differing link parameters.
    197  1. Your analysis. Here are some questions to consider.
    198   a. Does one protocol outperform the other?
    199   b. Under what conditions are performance differences most clearly seen? Why?
    200   c. What shortcomings in the experiment design may affect your results? How might you improve the experiment design?
    201   d. What interesting characteristics of the transfer protocols are not measured in this experiment? How might you design an experiment to investigate these?
     3This page has moved to [wiki:UDTExampleExperiment].