|
|
Sponsored Content
Special Forums
IP Networking
TCP Packet size
Post 302523124 by olle on Wednesday 18th of May 2011 03:16:29 AM
05-18-2011
mcnamara: No, there is unfortunately no way to get around using another chip.
And I should have mentioned that the MSS given from this device/chip is at 536 bytes. And this communication takes place at a local network. And there are no ICMP messages indicating that this MSS is set too high.
From a Windows application, which too communicates with this device, the packet size used is 536 bytes (well, only when needed of course). The chip is terrible slow anyway, but doing all this extra ACK'ing is making it much worse.
I'm passing buffers to send() which, some of them, are quite a lot bigger than 536 bytes and still the biggest TCP packet seen has a data length of 268 bytes. And as said the window size announced is always at 536 bytes from the chip. Still I fail to see how this makes any sense.
DGPickett: I can see that send() is swallowing chunks bigger than 268 bytes. I have tried playing around with TCP options like NODELAY, and changing socket buffer sizes. And have tried to change some of the settings found at /proc/sys/net related to TCP, but still nothing have changed this behavior.
Thank you for helping, and more tips are welcome
And I should have mentioned that the MSS given from this device/chip is at 536 bytes. And this communication takes place at a local network. And there are no ICMP messages indicating that this MSS is set too high.
From a Windows application, which too communicates with this device, the packet size used is 536 bytes (well, only when needed of course). The chip is terrible slow anyway, but doing all this extra ACK'ing is making it much worse.
I'm passing buffers to send() which, some of them, are quite a lot bigger than 536 bytes and still the biggest TCP packet seen has a data length of 268 bytes. And as said the window size announced is always at 536 bytes from the chip. Still I fail to see how this makes any sense.
DGPickett: I can see that send() is swallowing chunks bigger than 268 bytes. I have tried playing around with TCP options like NODELAY, and changing socket buffer sizes. And have tried to change some of the settings found at /proc/sys/net related to TCP, but still nothing have changed this behavior.
Thank you for helping, and more tips are welcome
|
|
10 More Discussions You Might Find Interesting
1. IP Networking
Hi,
Is there any way that i can directly take out the IP packet and see its contents.
Waiting for your answer ..............
Bye (4 Replies)
Discussion started by: manjunath
4 Replies
2. UNIX for Advanced & Expert Users
Hi,
When i try this command
ping ukblx151.ukhx.astrazeneca.net -n 3 | grep icmp
it gives following error :mad::
0821-077 ping: illegal packet size.
But when i give command
ping ukblx151.ukhx.astrazeneca.net
It returns correct output.:D
Could you please help? (2 Replies)
Discussion started by: vishal_ranjan
2 Replies
3. HP-UX
how to get network packet size
I would like get network output rate(kb/sec)
I type command "netstat -i"
Ipkts Ierrs Opkts Oerrs 653387 0 678202 0
but i didn't know what is it packet size ,
how could i get it? (1 Reply)
Discussion started by: alert0919
1 Replies
4. IP Networking
Hello,
I'm having an issue with TCP sockets.
When the TCP connection is terminated on one end, TCP packet with RST flag set is being sent to the sender. All the packets sent so far were carrying the DSCP 'AF21' set by me. But packet with RST flag is carrying DSCP '0'.
Is this expected or... (0 Replies)
Discussion started by: Solace
0 Replies
5. Programming
Hi all,
I need to change the source port number of an outgoing TCP packet. First I have to bind the socket to a particular port(suppose 9001) but when I send the TCP packet I want to change the source port number lets say to 9002 still letting the socket to be bound to the same old port (9001).... (0 Replies)
Discussion started by: anuragrai134
0 Replies
6. UNIX for Advanced & Expert Users
We're running Solaris 7 on FDDI n/w on an E6500 host and wish to use MTU (packet size) > 1500, more like 3072 bytes to begin with and possibly up to 4096 bytes.
Linux has /etc/network/interfaces. Does ANYONE remember the equivalent in Unix? When I do ifconfig eth0 mtu 4000, I get the error... (0 Replies)
Discussion started by: sharique
0 Replies
7. Solaris
Greetings, I'm stuck in a time warp using ancient machines from the prehistoric era that should be rightfully displayed in the Smithsonian.
We're running Solaris 7 on FDDI n/w on an E6500 host and wish to use MTU (packet size) > 1500, more like 3072 bytes to begin with and possibly up to 4096... (9 Replies)
Discussion started by: sharique
9 Replies
8. AIX
(5 Replies)
Discussion started by: Vishal_dba
5 Replies
9. Solaris
I have problem with oracle solaris 10 running on oracle sparc T4-2 server.
Os information: 5.10 Generic_150400-03 sun4v sparc sun4v
Output from tcpstat.d script
TCP bytes: out outRetrans in inDup inUnorder
6833763 7300 98884 0... (2 Replies)
Discussion started by: insatiable1610
2 Replies
10. Solaris
I'd like to use SPARC Solaris10 to 'wake up' an old PC on same LAN.
what is the simplest way of doing it? (preferably without installing new software)
thanks. (14 Replies)
Discussion started by: orange47
14 Replies
LEARN ABOUT OPENDARWIN
synos
synos(1) Mail Avenger 0.8.3 synos(1) NAME
synos - guess operating system from TCP SYN fingerprint SYNOPSIS
synos [--mtu mtu] [--db path] syn-fingerprint DESCRIPTION
synos takes a SYN fingerprint, in the format described for the CLIENT_SYNFP environment variable in the avenger(1) man page, and outputs a guess as to the type of the client operating system. synos makes use of the OpenBSD SYN fingerprint database (which is also repackaged with Mail Avenger). OPTIONS --mtu val Certain operating systems set the initial TCP window size based on the maximum transmission unit, or MTU, of the network. For such operating systems, synos usually checks the window size using both the client's MSS option plus 40 bytes (for TCP and IP headers), or a hard-coded MTU, which defaults to 1,500 bytes. If either value works, the fingerprint is considered to match the operating system. You can change the value 1,500 by specifying this option. A value of 0 tells synos to use only the value derived from the MSS option. --db file Specifies an alternate location for the SYN fingerprint database. FILES
/usr/local/share/pf.os Default location of SYN fingerprint database. SEE ALSO
avenger(1), asmtpd(8) The Mail Avenger home page: <http://www.mailavenger.org/>. The OpenBSD home page: <http://www.openbsd.org/>. BUGS
The operating system type is determined by heuristics that are not always reliable. Moreover, not all operating systems can be distinguished. The database may not even contain a client's particular operating system and version. It is not hard to fool synos deliberately by changing TCP socket options or injecting raw packets onto the network. AUTHOR
David Mazieres Mail Avenger 0.8.3 2012-04-05 synos(1)