Tuesday, July 6, 2010

How to Configure an $80 File Server in 45 Minutes

I use a modded Xbox and Xbox Media Center for playing media files across the network on my television and sound system. I also download large files, such as Linux ISOs, via BitTorrent. However, leaving my primary computer on all the time seemed like a waste of energy. I wanted a cheap, small headless machine that I could use as a Samba server and BitTorrent client so I could leave my workstation off when I wasn't using it.

I was in luck; HCDI Trading had a great "Fall Extravaganza" deal - a Dell OptiPlex GX150 Desktop for $47.91 (around $30 s/h). The specs:

  • Desktop Form Factor (it fits on a shelf in a media center)
  • Intel Celeron 700 MHz
  • 128MB RAM
  • 10GB HD
  • CD-Rom

Not cutting edge, but perfect for what I needed and priced very reasonably. Instead of using X11 and remote clients, I felt that web interfaces would be optimal out of simplicity. If I configured the server as a LAMP (Linux, Apache, MySQL, and PHP), then I could use web GUIs for configuration, administration, and file transfers.

Out of the popular Linux distributions, I'm most familiar with Gentoo and Ubuntu. Gentoo didn't seem appropriate; it would take hours and hours to get everything set up, configured and updated (even from a Stage 3 install), and any performance gain from the optimization would be negligible. Ubuntu has a history of reliability and ease of set up, so I used the Edgy Eft release as the basis my server.

The Applications

  • Remote Console - OpenSSH
    • The free version of the SSH connectivity tools. Use it for secure access to the console.
  • Web Server - Apache
    • It's free, fast, well supported and documented
  • Network Fileshare - Samba
    • Creates file shares that can be easily accessed from Windows. Notoriously obnoxious to configure by hand, the Samba Web Administration Tool (SWAT) simplifies the process immensely.
  • Database - MySQL
    • The popular free database server. While I can use a command line to administrate MySQL, I prefer using phpMyAdmin.
  • P2P Client - BitTorrent

Directions

  1. Download and burn the Ubuntu Server 6.10 i386 install CD.
  2. Install Ubuntu Server
    1. When asked about Software Installation, choose LAMP.
    2. When the installation is finished, take out the CD (I flip it over and leave it in the drive for the next time I need it) and reboot.
  3. Logon as the regular user that you created during setup.
  4. Give root a password.
    sudo passwd root
  5. We're going to modify the file sources list to allow us access to all the required files. Uncomment the edgy universe, edgy-security main restricted, edgy-security universe, edgy multiverse, edgy-backports main restricted universe multiverse, and edgy-commercial main. In addition, Ubuntu leaves the install CD as one of the file sources after installation, so you'll need to remove the CD from the source list. Comment out the CD from the list and save.
    sudo nano -w /etc/apt/sources.list
  6. Update apt-get to use the new sources.
    sudo apt-get update
  7. Install OpenSSH server
    sudo apt-get install openssh-server
  8. Turn off the computer.
    sudo shutdown -h now
  9. Disconnect the monitor, keyboard, and set the computer wherever it's going to lurk for the next couple years.
  10. Turn on the computer and wait about a minute for it to boot.
  11. Using your preferred SSH client, connect to the machine. I prefer PuTTY for Windows.
  12. Install Samba and SWAT.
    sudo apt-get install samba smbfs swat xinetd
  13. Create a Samba user.
    sudo smbpasswd -a username
  14. Add SWAT to the xinet configuration and save.
    sudo nano -w /etc/xinetd.d/swat
    # description: SAMBA SWAT
    service swat
    {
    disable = no
    socket_type = stream
    protocol = tcp
    #should use a more limited user here
    user = root
    wait = no
    server = /usr/sbin/swat
    }
  15. Reload Xinetd with the new configuration.
    sudo dpkg-reconfigure xinetd
  16. Verify that SWAT is up and running. If you don't get any response, start Googling.
    sudo netstat -tap | grep swat
  17. Make the shared directory for TorrentFlux.
    sudo mkdir /share
    sudo mkdir /share/incoming
    sudo chmod 777 /share/incoming/
  18. Create the Samba share. Using a web browser, go to http://hostname:901/shares
    1. path /share/incoming
    2. valid users - username
    3. read only - no
    4. Commit Changes
  19. Install phpMyAdmin
    sudo apt-get install phpmyadmin
  20. Connect to phpMyadmin using a web browser - http://hostname/phpmyadmin
    1. Secure MySQL / phpMyAdmin
      1. login: root (no pass)
      2. Privileges
        1. Add a password to both root accounts
      1. Create a database for TorrentFlux
        1. Databases - Create New - torrentflux
  21. Install TorrentFlux
    sudo apt-get install torrentflux
    1. Ignore the libphp-adodb message.
    2. Configure database for torrentflux with dbconfig-common? - yes
    3. Password - blank (generates random)
    4. Restart Apache? - Yes
  22. Configure TorrentFlux using a web browser - http://hostname/torrentflux/
    1. Username / Password - root / root
    2. Configuration
      1. Path - /share/incoming/
      2. Save
    3. Click My Profile
      1. Update your password
    4. Create a user account
      1. Admin - New User
  23. Remove the Apache default documents from web root.
    sudo rm -Rf /var/www/apache2-default/
  24. Create a basic web launchpad for easy access your services.
  25. sudo nano -w /var/www/index.php



Post Mortem

Overall, the installation and configuration took me about 45 minutes to go from zero to a fully functional system. The Xbox can see the share, and I've got a new playground for web development. I would suggest cleaning up and securing that home page a bit; research .htaccess files to lock it down.

I put my file server in my media center underneath my router, and it takes up about the same space as a VCR. For $80, I feel that I've made an excellent investment.

Tuesday, April 27, 2010

Interesting blogs

http://www.cse.msu.edu/~wangbo1/ns2/

http://peelmeagrape.net/eoin/swarm/

http://tagus.inesc-id.pt/~pestrela/ns2/ns2_tips.html

http://www.ict.csiro.au/staff/ren.liu/ns-2/simulate-802_11.htm

http://elmurod.net/wps/?p=72

How to Display RTS/CTS Packet Type in NS2 Trace File

http://www.cse.msu.edu/~wangbo1/ns2/

There are several posts on NS2 user mailing list explaining how to display RTS/CTS packet type in NS2 trace file. Here is an example. However, I failed to adopt them in NS 2.26 with new trace format. Here is my solution which just uses code written for old trace format. You only need to modify ns-2.26/trace/cmu-trace.cc in the following way and I think this solution can also work for NS versions later than 2.26.

void
CMUTrace::format_mac(Packet *p, const char *why, int offset)
{

......


/*
sprintf(pt_->buffer() + offset,
"-Ma %x -Md %x -Ms %x -Mt %x ",
mh->dh_duration,
ETHER_ADDR(mh->dh_da),
ETHER_ADDR(mh->dh_sa),
GET_ETHER_TYPE(mh->dh_body));
*/



sprintf(pt_->buffer() + offset,
"-Ma %x -Md %x -Ms %x -Mt %s ",
mh->dh_duration,
ETHER_ADDR(mh->dh_da),
ETHER_ADDR(mh->dh_sa),
((ch->ptype() == PT_MAC) ? (
(mh->dh_fc.fc_subtype == MAC_Subtype_RTS) ? "RTS" :
(mh->dh_fc.fc_subtype == MAC_Subtype_CTS) ? "CTS" :
(mh->dh_fc.fc_subtype == MAC_Subtype_ACK) ? "ACK" :
"UNKN"
) : packet_info.name(ch->ptype())));



return;


}

Here is an exmaple NS2 trace file which displays the RTS/CTS/ACK information of MAC layer frame.

r -t 0.003583515 -Hs 0 -Hd -2 ... -Nl MAC -Nw --- -Ma 0 -Md 0 -Ms 0 -Mt ACK
s -t 0.003853348 -Hs 8 -Hd -2 ...-Nl MAC -Nw --- -Ma 2ff -Md 0 -Ms 8 -Mt RTS
r -t 0.004125515 -Hs 0 -Hd -2 ... -Nl MAC -Nw --- -Ma 2ff -Md 0 -Ms 8 -Mt RTS
s -t 0.004135515 -Hs 0 -Hd -2 ... -Nl MAC -Nw --- -Ma 1fd -Md 8 -Ms 0 -Mt CTS
r -t 0.004383682 -Hs 8 -Hd -2 ... -Nl MAC -Nw --- -Ma 1fd -Md 8 -Ms 0 -Mt CTS

Thursday, April 22, 2010

Notes

Google .. blackhole attack ns2 aodv ..
Examples
http://elmurod.net/wps/?p=196

http://cdsweb.cern.ch/record/1206617?ln=de
http://arxiv.org/pdf/0909.2371.pdf


Some code here : ?
Google .. prevent blackhole ns2 ..
? http://www.nesmd.com/shtml/18947.shtml ?

in reference to:

"Google .. blackhole attack ns2 aodv .. Examples http://elmurod.net/wps/?p=196 http://cdsweb.cern.ch/record/1206617?ln=de http://arxiv.org/pdf/0909.2371.pdf Some code here : ? Google .. prevent blackhole ns2 .. ? http://www.nesmd.com/shtml/18947.shtml ?"
- simulating blackhole attack in ns-29 (view on Google Sidewiki)

Example to create BH node

I want to make a node behave as blackhole node, using the above technique. The above method drops all packets that go through it, but it doesn’t kind of forces a packet to go through it by sending a high sequence number.
For that, i have done this within this fn. AODV::recvRequest(Packet *p):

if(malicious==true) {
sendReply(rq->rq_src, // IP Destination
1, // Hop Count
index, // Dest IP Address
4294967295, // Max. Dest Sequence Num if the node is malicious
MY_ROUTE_TIMEOUT, // Lifetime
rq->rq_timestamp); // timestamp
}
else {
sendReply(rq->rq_src, // IP Destination
1, // Hop Count
index, // Dest IP Address
seqno, // Dest Sequence Num
MY_ROUTE_TIMEOUT, // Lifetime
rq->rq_timestamp); // timestamp
}

and removed
if(malicious == true) {
drop(p, DROP_RTR_ROUTE_LOOP);
}

from rt_resolve() fn.

Now, what changes do i need to make, so that the code can distinguish data packets and management packets, so that it could drop the data packet

if you want to drop only data packets you need to check packet type reger HDR_CMN for more info.

Ns2 code

# Preamble
set ns [new Simulator]

# Tell the simulator to use dynamic routing
$ns rtproto DV
Agent/rtProto/DV set advertInterval 1.0
Agent/rtProto/DV set INFINITY 100

# Set ns color indices
$ns color 0 blue
$ns color 1 black
$ns color 2 turquoise
$ns color 3 orange
$ns color 4 gold
$ns color 5 red


# Open trace files
set f [open SIM.tr w]
$ns trace-all $f
set nf [open SIM.nam w]
$ns namtrace-all $nf


proc finish {} {
global ns f nf
$ns flush-trace
close $nf
close $f
puts "running nam"
exit 0
}


# Create 13 routers
for {set i 0} {$i < 13} {incr i} {
set r($i) [$ns node]
}

# Create 30 pc
for {set i 13} {$i < 43} {incr i} {
set pc($i) [$ns node]
}

for {set i 0} {$i < 13} {incr i} {
$r($i) shape hexagon
$r($i) color green
}

for {set i 13} {$i < 43} {incr i} {
$pc($i) color blue
}





# Create duplex links
$ns duplex-link $r(0) $pc(13) 1Mb 50ms DropTail
$ns duplex-link $r(0) $pc(14) 1Mb 50ms DropTail
$ns duplex-link $r(0) $pc(15) 1Mb 50ms DropTail
$ns duplex-link $r(0) $r(1) 2Mb 50ms DropTail
$ns duplex-link $r(0) $r(5) 2Mb 50ms DropTail
$ns duplex-link $r(1) $pc(16) 1Mb 50ms DropTail
$ns duplex-link $r(1) $pc(17) 1Mb 50ms DropTail
$ns duplex-link $r(1) $pc(18) 1Mb 50ms DropTail
$ns duplex-link $r(1) $r(2) 2Mb 50ms DropTail
$ns duplex-link $r(1) $r(6) 10Mb 50ms DropTail
$ns duplex-link $r(2) $pc(19) 1Mb 50ms DropTail
$ns duplex-link $r(2) $pc(20) 1Mb 50ms DropTail
$ns duplex-link $r(2) $pc(21) 1Mb 50ms DropTail
$ns duplex-link $r(2) $r(3) 2Mb 50ms DropTail
$ns duplex-link $r(2) $r(7) 10Mb 50ms DropTail
$ns duplex-link $r(3) $pc(22) 1Mb 50ms DropTail
$ns duplex-link $r(3) $pc(23) 1Mb 50ms DropTail
$ns duplex-link $r(3) $pc(24) 1Mb 50ms DropTail
$ns duplex-link $r(3) $r(4) 2Mb 50ms DropTail
$ns duplex-link $r(3) $r(8) 10Mb 50ms DropTail
$ns duplex-link $r(4) $pc(25) 1Mb 50ms DropTail
$ns duplex-link $r(4) $pc(26) 1Mb 50ms DropTail
$ns duplex-link $r(4) $pc(27) 1Mb 50ms DropTail
$ns duplex-link $r(4) $r(9) 2Mb 50ms DropTail
$ns duplex-link $r(5) $pc(28) 1Mb 50ms DropTail
$ns duplex-link $r(5) $pc(29) 1Mb 50ms DropTail
$ns duplex-link $r(5) $pc(30) 1Mb 50ms DropTail
$ns duplex-link $r(5) $r(6) 10Mb 50ms DropTail
$ns duplex-link $r(6) $r(7) 10Mb 50ms DropTail
$ns duplex-link $r(6) $r(10) 10Mb 50ms DropTail
$ns duplex-link $r(7) $r(8) 10Mb 50ms DropTail
$ns duplex-link $r(7) $r(11) 10Mb 50ms DropTail
$ns duplex-link $r(8) $r(9) 10Mb 50ms DropTail
$ns duplex-link $r(8) $r(12) 10Mb 50ms DropTail
$ns duplex-link $r(9) $pc(31) 1Mb 50ms DropTail
$ns duplex-link $r(9) $pc(32) 1Mb 50ms DropTail
$ns duplex-link $r(9) $pc(33) 1Mb 50ms DropTail
$ns duplex-link $r(10) $r(11) 2Mb 50ms DropTail
$ns duplex-link $r(10) $pc(34) 1Mb 50ms DropTail
$ns duplex-link $r(10) $pc(35) 1Mb 50ms DropTail
$ns duplex-link $r(10) $pc(36) 1Mb 50ms DropTail
$ns duplex-link $r(11) $r(12) 2Mb 50ms DropTail
$ns duplex-link $r(11) $pc(37) 1Mb 50ms DropTail
$ns duplex-link $r(11) $pc(38) 1Mb 50ms DropTail
$ns duplex-link $r(11) $pc(39) 1Mb 50ms DropTail
$ns duplex-link $r(12) $pc(40) 1Mb 50ms DropTail
$ns duplex-link $r(12) $pc(41) 1Mb 50ms DropTail
$ns duplex-link $r(12) $pc(42) 1Mb 50ms DropTail



# Orient the nodes
$ns duplex-link-op $r(0) $pc(13) orient 165deg
$ns duplex-link-op $r(0) $pc(14) orient 180deg
$ns duplex-link-op $r(0) $pc(15) orient 195deg
$ns duplex-link-op $r(0) $r(1) orient 0deg
$ns duplex-link-op $r(0) $r(5) orient 270deg
$ns duplex-link-op $r(1) $pc(16) orient 105deg
$ns duplex-link-op $r(1) $pc(17) orient 90deg
$ns duplex-link-op $r(1) $pc(18) orient 75deg
$ns duplex-link-op $r(1) $r(2) orient 0deg
$ns duplex-link-op $r(1) $r(6) orient 270deg
$ns duplex-link-op $r(2) $pc(19) orient 105deg
$ns duplex-link-op $r(2) $pc(20) orient 90deg
$ns duplex-link-op $r(2) $pc(21) orient 75deg
$ns duplex-link-op $r(2) $r(3) orient 0deg
$ns duplex-link-op $r(2) $r(7) orient 270deg
$ns duplex-link-op $r(3) $pc(22) orient 105deg
$ns duplex-link-op $r(3) $pc(23) orient 90deg
$ns duplex-link-op

in reference to: [ns] black hole simulation (view on Google Sidewiki)

Wednesday, April 21, 2010

Printing Routing Table in AODV


Actually you can find all this information in trace file which NS2 made, however using following code simplifies getting required informaiton during running time. The code stores

Add following code to aodv.h after void rt_down(aodv_rt_entry *rt);

void rt_print(nsaddr_t node_id);

Add following code to aodv.cc after void AODV::rt_down(aodv_rt_entry *rt)

void AODV::rt_print(nsaddr_t node_id) {
FILE * dumpFile;
char dumpFileName[50] = "rtable.txt";
dumpFile = fopen(dumpFileName, 'a');
aodv_rt_entry *rt;
fprintf(dumpFile, "=======================================================");
for (rt=rtable.head();rt; rt = rt->rt_link.le_next) {
fprintf(dumpFile, "NODE: %i\t %.4lf\t %i\t %i\t %i\t %i\t %i\t %.4lf\t %d \n", node_id, CURRENT_TIME, rt->rt_dst, rt->rt_nexthop, rt->rt_hops, rt->rt_seqno, rt->rt_expire, rt->rt_flags)
}
fclose(dumpFile);
}

The function (rt_print) can be used anywhere in AODV. For example, I am using the function, when route request generated node receives route reply message (RREP).

if (ih->daddr() == index) { // If I am the original source
// Update the route discovery latency statistics
// rp->rp_timestamp is the time of request origination
rt_print(index); // print this nodes whole routing table
rt->rt_disc_latency[(unsigned char)rt->hist_indx] = (CURRENT_TIME - rp->rp_timestamp)
/ (double) rp->rp_hop_count;
// increment indx for next time
rt->hist_indx = (rt->hist_indx + 1) % MAX_HISTORY;
}