[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Good HFSC explanation

Apologies for resurrecting this thread, but...
--- jared r r spiegel <jrrs[email protected]> wrote:
> On Fri, Feb 11, 2005 at 03:39:17PM +0000, Bob wrote:
> > Is there a clear HFSC explanation somewhere, with real simple
> examples? 
> > Preferably that apply directly to PF which uses three SC types, not
> two.
>   of note:
> http://marc.theaimsgroup.com/?l=openbsd-pf&m=105691519510241&w=2
> http://marc.theaimsgroup.com/?l=openbsd-pf&m=107936788832658&w=2
> http://marc.theaimsgroup.com/?l=openbsd-pf&m=110488079304643&w=2
Based on the jared's reply in the third link, what Trevor Talbot said
in the first link, and based on not seeing anything contradictory, is
it correct then to say that:
If you have specified the 3 service curves, the "bandwidth" keyword is
redundant and/or unnecessary?
I am still bothered by the "priority" keyword/notion.  It seems to me
that even with fully-specified realtime and linkshare curves, you have
really set up a two priority system.
Assume there is no backlog and you have 2 packets that are received and
placed into 2 different queues that have not reached their "realtime"
bandwidth limit.  It would make sense that in a CBQ-style system, the
packet falling into a queue with "realtime" bandwidth available _and_
higher priority would be delayed less than the other packet in this
So with fully-specified service curves, does HFSC as implemented here
in fact superimpose CBQ-style hierarchical priorities ontop, or do the
service curve specifications somehow mean that also giving "priority"
doesn't makes sense?
I took jared's reply in the third link above to indicate the latter was
true, but I don't immediately see why it would be.
John R.
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around