I''ve taken a look at the list archives, the Advanced Routing HOWTO, and done a quick netsearch, but still don''t know how to solve my problem. I''m currently living in university housing, and they provide an network connection. The only limit on transfer speed is link saturation, however there are monthly and daily limits for all data transfered to and from the internet. Is it currently possible to shape traffic so that there is: - no limit on local transfers - 5G/month, 200/day, 800/night limit on external transfers - the ability to still do transfers as fast as the link will allow (until it bumps up against transfer limits). Any help would be appreciated... a thread I might have missed, a webpage, anything. Thanks in advance, Nathan
On Wed, 2001-11-14 at 12:37, Don Cohen wrote:> I don''t see to be ale to post to the list, so a personal response: > > I''m currently living in university housing, and they provide an network > connection. The only limit on transfer speed is link saturation, > however there are monthly and daily limits for all data transfered to > and from the internet. Is it currently possible to shape traffic so > that there is: > - no limit on local transfers > - 5G/month, 200/day, 800/night limit on external transfers > - the ability to still do transfers as fast as the link will allow > (until it bumps up against transfer limits). > > The university does enforce these limits? > I guess they just stop forwarding when the limits are reached. > So you don''t need to enforce them. You just want to allocate > them more evenly over time? The easy way would be to limit the > rate at a finer granularity. Otherwise we get into details of > the limit - like when do the limited period start/end.They track, but never stop forwarding; they just start charging at $0.05/meg ($50/gig). I''d rather avoid that. What I really want to do is allow uneven allocation, but stay inside the limits. After looking at the docs, I think I need something like a Token Bucket Filter, but at large scale: hours and days rather than seconds. The only problem is that I haven''t seen anything that even says if this sort of thing is possible.
I''ve set up bandwidth limiter in my Linux Redhat 7.1 Box, but i
don''t feel
that it works, anybody know how to show that this is really work ? some kind
of benchmarking tools ;)
-- ady --
email: ady <at> ebdesk.com
adiwicaksono <at> yahoo.com
ady <at> students.if.itb.ac.id
homepage: http://ady97.hypermart.net/
Ady Wicaksono a écrit :> > I''ve set up bandwidth limiter in my Linux Redhat 7.1 Box, but i don''t feel > that it works, anybody know how to show that this is really work ? some kind > of benchmarking tools ;)Hi, a simple test I did was to limit my station''s bandwidth, and start a long ftp transfer. My ftp client was telling me the current transfer rate. Then I remove the limit and see the transfer rate increase. -- Marc Delisle Service de l''informatique Collège de Sherbrooke, Québec
Thanks .... Hmm.... now i see that CBQ is really work :), On Thursday 15 November 2001 08:41 pm, Marc Delisle wrote:> Ady Wicaksono a écrit : > > I''ve set up bandwidth limiter in my Linux Redhat 7.1 Box, but i don''t > > feel that it works, anybody know how to show that this is really work ? > > some kind of benchmarking tools ;) > > Hi, > > a simple test I did was to limit my station''s bandwidth, and start a long > ftp transfer. My ftp client was telling me the current transfer rate. Then > I remove the limit and see the transfer rate increase.-- -- ady -- email: ady <at> ebdesk.com adiwicaksono <at> yahoo.com ady <at> students.if.itb.ac.id homepage: http://ady97.hypermart.net/
I''ve got a slackware8 system with kernel 2.4.13 self-installed. I selected all the advanced routing/etc options (i figured i''d rather have a slightly bloated kernel and more options than having to recompile) as include. I couldn''t find the ip tool on my system, so i downloaded the latest from the location in the HOWTO. When compiling it, I hit this: --------------- make[1]: Entering directory `/usr/src/iproute2/lib'' gcc -D_GNU_SOURCE -O2 -Wstrict-prototypes -Wall -g -I../include-glibc -inclu de ../include-glibc/glibc-bugs.h -I/usr/src/linux/include -I../include -DRESOLV E_HOSTNAMES -c -o ll_proto.o ll_proto.c ll_proto.c:36: `ETH_P_ECHO'' undeclared here (not in a function) ll_proto.c:36: initializer element is not constant ll_proto.c:36: (near initialization for `llproto_names[1].id'') make[1]: *** [ll_proto.o] Error 1 make[1]: Leaving directory `/usr/src/iproute2/lib'' make: *** [all] Error 2 --------------- Do I have something missing/wrong? I set this path in the makefile and commented out the 2 lines for if you are using bind, since I don''t have bind and am using glibc. --------------- #options if you compile with libc5, and without a bind>=4.9.4 libresolv LDLIBSADDLIB=inet_ntop.o inet_pton.o --------------- Any help is greatly appreciated.
> Do I have something missing/wrong? > I set this path in the makefile and commented out the 2 lines for if you are > using bind, since I don''t have bind and am using glibc.It''s because of a change in recent kernels. Apparently ETH_P_ECHO was bogus anyways, and was removed. Search google groups for more details. The fix is simple. In lib/ll_proto.c, just delete line 36. This line: __PF(ECHO,echo) Cheers, Jason.