Connection closing improvements and explicit cleanup code

[dccp2tcp] / README

diff --git a/README b/README
index 520902e6f4b4dd12a0f0b58f25552dc2e38e467c..4364d412ac05d37971adac9904554ac7edcd6982 100644 (file)
--- a/README
+++ b/README
@@ -10,25 +10,35 @@ on linux anyway (we need the dccp header include file) and it comprises only one
 If you have problems, simply modify the make file as needed. Contact me, if major problems
 arise.
 
+In order to utilize this program effectively you will also need Tcptrace, which you can download
+from http://www.tcptrace.org and the version of xplot available from http://www.tcptrace.org under
+"Useful Companion Programs" (Note! This is not the xplot that is in the Ubuntu repositories).
+
 
 Usage is pretty simple:
 dccp2tcp dccp_file tcp_file [-d] [-y] [-g] [-s]
        -d is debug
        -y shifts the window line in tcptrace (yellow) to the highest received acknowledgment. Normally this line is just a constant amount more than the ack number(i.e. useless).
        -g shifts the ack line in tcptrace (green) to the highest received acknowledgment. Normally this line is the standard TCP ack number, which, for DCCP, translates to the highest contiguous acknowledgement in the ack vector.
-       -s converts the DCCP ack vector to TCP SACKS.
+       -s converts the DCCP ack vector to TCP SACKS. Specify -s twice to only see those Ack vectors with a loss interval in them. This is convenient way to see loss events.
 
 For typical usage, you probably just want -s.
 
+Once you run dccp2tcp, you will then want to run tcptrace on the tcp_file to generate graphs. The command should be something like this:
+tcptrace -lGt tcp_file
+
+You can then use xplot to view the resulting graphs, like this:
+xplot a2b_tsg.xpl a2b_rtt.xpl
+
 
 This program does have several important limitations:
-       1)Supports only a single DCCP contection per capture.
+       1)Supports only a single DCCP connection per capture.
        2)Source Port can't equal Destination Port. If the two are equal, this program will get very confused.
        3)DCCP MUST use 48 bit sequence numbers.
        4)Checksums are not computed (they are zeroed).
        5)Only converts those packet types used by the Linux implementation at this date
        6)DCCP Ack packets show up as TCP packets containing one byte
-       7)Very little error checking of packet headers, it's research software afterall
+       7)Very little error checking of DCCP
 
 
 
@@ -46,7 +56,7 @@ You should have received a copy of the GNU General Public License
 along with dccp2tcp.  If not, see <http://www.gnu.org/licenses/>.
 
 
-Copyright 2010
+Copyright 2011
 Samuel Jero
 Computer Science student, Ohio University
 sj323707@ohio.edu