[Iccrg] congestion control and visability to the user
michawe at ifi.uio.no
Tue Mar 30 11:51:45 BST 2010
On Mar 29, 2010, at 5:34 PM, Mikael Abrahamsson wrote:
> On Mon, 29 Mar 2010, michawe at ifi.uio.no wrote:
>> My notion was that it's not, usually, because of standard TCP
>> being "lame" (as Matt rightfully called it).
> This is the first time in my 15+ year of networking that I've heard
> this view of the world. I couldn't disagr ee more. Only time I
> wasn't able to saturate my access link was when I was using Win95 or
> so, and it didn't s upport window scaling and my internet connection
> was 100+ megabit/s and the ftp server was 10ms away. I could n't
> disagree more with this view of the world.
How can you be so sure? The longer the delay, and the greater
the bandwidth, the harder it gets for standard TCP to saturate the
bottleneck. See the abstract of RFC 3649 (2nd paragraph), which
lays it out quite nicely - what do you think of this, is it unrealistic,
or wrong in your opinion?
Surely, like everyone else I know, you have once downloaded
something from some sever somewhere on the net
at much less than your internet connection's speed - then, how do
you know where the bottleneck comes from - is it a slow access
link of the server, or the server itself being slow, or TCP not
saturating the bottleneck (wherever it is) well? How would you know?
(because I believe that it's hard to know about these things, I
agree with your suggestion to develop some means to make
these things more visible - and Web100, for instance (as Matt
mentioned) wouldn't give you a good answer to that either.
(yes you could deduce things from looking at cwnd in Web100,
but where the bottleneck is you wouldn't know) )
> Again, ISPs only get away with congesting their cores if there isn't
> enough competition in the market, and that's not a technical
> problem, it's usually a political problem. Trying to handle this
> dysfunctional situation by technical means is bound to fail. Why
> would someone invest in expensive bw capacity/congestion/flow/
> fairness technology if said upgrade due to bw constraints is already
> too expensive?I don't get it.
There's nothing expensive to an ISP about changing congestion
control functionality in end hosts.
> In my experience, the biggest congestion problem is in the upstream
> direction from my CPE. Any normal user would have an ADSL modem or
> alike there from some low-end manufacturer who won't invest a dollar
> more to do anything more than FIFO because they're acting in an
> extremely tight marketplace where every cent counts. There is
> already plenty of tech out there to help in this siuation but
> they've chosen to do nothing. Fair-queue is one which would hugely
> benefit all ADSL users (or anyone with slow uplinks). This is not
> being done, and yet people are talking about inveting eve nmore
> complicated mechanisms?
There's nothing expensive to a vendor of an ADSL modem about
changing congestion control functionality in end hosts.
More information about the Iccrg