3 years ago i've found a company devoted to networking
Now I'm proud to announce the successor of
Sequre ISP is light years ahead of
Sequre ISP comes as Free Software (AGPLv3) or, if you're interested in our experience and support mounting ISPs, with a commercial license.
So i invite you to go to sequreisp home page and the
Or you can just check the
Regards.
Luciano Ruete, author of
New
This release should be functional,
highly is needed.
To get it up&running edit
Then run:
#
download from:
![]()
or git-[pull|clone] from:
http://www.lugmen.org.ar/~luciano/git-repo/htb-gen/.git
Some changlelog:
-command line options updated
-
-
-General docs update
02.06.2007 02:24
time to make a public decent documented and generalized release.
But right now i think that is better to put the stuff here, so others can
enjoy the notorious improvements (and maybe someone wants to help out)
Lets go to the hacks:
I have made 2 flavors of
config files where touched and some documentation update was made in place.
* First flavor (
![]()
Per host p2p percent of rate assignation
Named ISP/LAN and clients in the
Code simplification
pfifo_fast for prio class
Compatibility with bash v2
tc batch mode support, now both iptables and tc are batched,
huge speed impact on large setups and yet tc and iptables
command in the source are transparent readables
![]()
Grained prio/non_prio per host definition, you can setup per client:
prio_udp_ports
prio_protos (as esp,gre,igmp or ie even udp to include all udp traffic)
prio_helpers (netfilter helpers)
An
data entry
innline graphics per client
* Bonus
Then there is a
images each time you visit the page. The images are stored in /var/lib/rrd/img so apache need
to read and write in this directory.
The graphics are per client and have different color(ligth/dark green) for prio/non_prio traffic. :-)
There is a git repository
contribute sends diffs against the
#
Good luck, and plz mail me any
management tool. You can set up/down portions of bandwith for each
host or network, that goes trough your router/firewall.
Prioritary traffic(web, mail, gaming, ftp, voip, streaming) is
preferred over Junk traffic(kazaa, emule, etc). Also dynamic
bandwith borrow and
htb boundaries.
A
is possible.
All bash based so it can be used in embedded routers/firewalls
(wired/wireless).
Two backend are aviable:
The firstone is
iface_down="eth1" # Server LAN ifaceiface_up="eth0" # Server INET iface
total_rate_down=1024 #Total download bw
total_rate_up=512 #Total upload bw
Te secondone is
# down down up up# min max min max
#ip (rate) (ceil) (rate) (ceil)
192.168.1.2 0 64 0 32
192.168.1.3 0 128 0 64
192.168.1.4 0 256 0 128
10.0.0.1/30 256 512 128 256
200.80.22.2 256 256 256 256
As you see
-Is extremly easy to maintain a large amount of hosts/networks
-ips from different newtorks are allowed(they have to pass FORWARD anyway)
-newtork syntax is allowed, bw will be assigned to the network
-fixed rate support, i can garantee some bw.
-A 0(zero) in rate means that it will be automagickly calculated based on the
ceil weigth, and the unassigned bandwidth
After this you run '
![]()
![]()
![]()
Beta version(see News section at top for details)
![]()
Archive at http://www.praga.org.ar/dev/
-QoS htb kernel support
-iproute2 tc
-iptables
-
easy addapted if you know a litle bit of iptables):
-Prio traffic
By thefault the host bandwidth is shared betwen this two kind of traffic, but
the script grants that anytime that i use prio traffic it will climb up to
90%, till that 'prio traffic' ends. This % can be modified if you want, see
the 'rate_dfl_percet' value in conf. Also is posible to save junk bandwith
assigning only a % of host ceil to the dfl traffic, see 'ceil_dfl_percet'.
This will help to have several host sharing bw without almost any complaint.
ones) that connects two or more networks.
In general there are two networks: Internet and a LAN(or a set of public IPs)
At any place where is a shared Internet access for multiple users (Small or
medium range ISP and WISP, neighbourhood wireless/wired LAN, university, etc)