In my attempt to move my traffic shaping to my router from an internal
box, I have come up with a new set of rules. I would appreciate any
feedback the list members might have.
What I am trying to do is shape my outbound bandwidth so that my que
doesn''t get full. I run a ftp server 24/7, do normal email, some ssh,
web surfing and some downloading. I have a cable modem with 10mbit
download and 1mbit upload. The router is a linux box running IPCOP and
eth1 faces the internet and eth0 faces the LAN.
I want to set my cap at 105kbps (slightly less than 1mbit) for all
outbound traffic. I want ftp traffic to send at the full 105kbps
without affecting my download speed or ssh ability. I also want the ftp
to send at the full 105kbps UNLESS I am sending an email. I think the
following rules will do that.
#!/bin/bash
# clear out the chain and setup a new chain
iptables -t mangle -D OUTPUT -o eth1 -j BW-OUT 2> /dev/null > /dev/null
iptables -t mangle -F BW-OUT 2> /dev/null > /dev/null
iptables -t mangle -X BW-OUT 2> /dev/null > /dev/null
iptables -t mangle -N BW-OUT
iptables -t mangle -I POSTROUTING -o eth1 -j BW-OUT
# mark packets: 2 is active ftp and passive ftp, 1 is ACK for downloads
and email and everything else
iptables -t mangle -A BW-OUT -p tcp -m length --length :64 -j MARK
--set-mark 1
iptables -t mangle -A BW-OUT -m tcp -p tcp --dport 25 -j MARK --set-mark 1
iptables -t mangle -A BW-OUT -p tcp --sport 59999 -j MARK --set-mark 2
iptables -t mangle -A BW-OUT -p tcp --sport 50000:51000 -j MARK
--set-mark 2
iptables -t mangle -A BW-OUT -m mark --mark 0 -j MARK --set-mark 1
# clear the qdisc
tc qdisc del dev eth1 root
#add the root qdisk
tc qdisc add dev eth1 root handle 1: htb default 10
#add main rate limit class and 2 leafs
tc class add dev eth1 parent 1: classid 1:1 htb rate 105kpbs ceil 105kpbs
tc class add dev eth1 parent 1:1 classid 1:10 htb rate 50kbps ceil
105kbps prio 0
tc class add dev eth1 parent 1:1 classid 1:11 htb rate 50kbps ceil
105kbps prio 1
#filter traffic into classes
tc filter add dev eth1 parent 1:0 prio 0 protocol ip handle 1 fw flowid
1:10
tc filter add dev eth1 parent 1:0 prio 1 protocol ip handle 2 fw flowid
1:11
So, in a nutshell. This should shape outbound ftp to 105kpbs. If I
download, the ACK packets should go to the front of the que. If I send
a large email, the email should borrow speed from the ftp. Web surfing
should be snappy and ssh also beacuse they are in the higher priority que.
Any comments?
nix4me
