Hi! This is my second post in this list with this problem and I need clearing. Situation is this: My host is a WLAN Access Point (AP) and I am using one network card. The host is going to generate packets to stations and those links with good uplink to the host are going to have priority in the downlink. Incomming Information in the uplink are going to be stored and used at egress to regulate downlink traffic. Incomming packets are only information given and flow quality is within the ttl-field. ----> ingress (Classifying, using information in ttl-field). Packet generated from packet generator. | ------>egress (regulating downlink traffic) After some tests using tcng and reading [1], [2], [3], [4], I have realised that: * I cannot use DSMARK instead of ingress, sins I only have one wireless network card (wlan0) in my host. * Using u32 filter, I dont know how to store link quality after classification in ingress. Q: Can I store the link quality in a variable/structure in ingress after classification and use this information at egress? * Using tcindex: I have to use hash-table or set_tcindex and use this information at egress. * RSVP does not scale, so my choice will be to use tcindex, where I classify flows into classes using tcindex. Is tcindex and rsvp possible solutions to this problem using tcng [5]? Or is there another better solution to this problem? [1] Leonardo Balliache http://www.opalsoft.net/qos/DS-210.htm [2] tcng Reference Manual http://linux-ip.net/gl/tcng/node165.html [3] http://almesberger.net/cv/papers.html [4] http://almesberger.net/cv/papers/dsid-01.ps.gz [5] http://linux-ip.net/gl/tcng/node165.html - Johan- _______________________________________________ LARTC mailing list / LARTC@xxxxxxxxxxxxxxx http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
