Re: Nonlinear limits on the number or rate of NSURLConnections
Re: Nonlinear limits on the number or rate of NSURLConnections
- Subject: Re: Nonlinear limits on the number or rate of NSURLConnections
- From: Jerry Krinock <email@hidden>
- Date: Wed, 26 Aug 2009 13:54:09 -0700
On 2009 Aug 26, at 06:19, Hamish Allan wrote:
On Wed, Aug 26, 2009 at 8:00 AM, Jerry Krinock<email@hidden> wrote:
1. The shutoff seems to be triggered by too many connections, not
by too
many bits per second. Sending, say, 5 requests per second and
receiving,
say, one 1500-byte packet with each one, is a rather modest 60 Kb/
sec.
That sounds exactly like the sort of naive heuristic a sniffer might
have for detecting P2P traffic.
Thanks, Hamish. I didn't know that much about P2P until this morning,
but I believe you're saying that an ISP would use the high rate of
connection requests to distinguish between "allowed" streaming video
from, say cnn.com and "bad" BitTorrent/P2P connecting to many servers
in a "swarm". That makes sense.
2. The shutoff is not just limiting my bit rate, it's a complete
shutdown
for a minute or so. Like "You were bad so we're going to drop all
of your
packets for a minute or so."
And that sound like exactly the sort of response ISPs are starting to
make. "Ah, well, if you were running BitTorrent, no wonder your
internet is broken!"
Very interesting. So, I'm updating my app to sense when connections
are being dropped and adapt accordingly.
Also, I ran this test a couple times:
http://broadband.mpi-sws.org/transparency/bttest.php
but both times it concluded with "Our tool was unable to measure your
link. The connection to the measurement server was lost". I suspect
that the type of manipulation is different than what the designers of
this test were expecting, but that's surprising since this behavior
has been around for, I'd say, at least a year or more.
On 2009 Aug 26, at 13:31, Scott Ribe wrote:
Could it have something to do with your router getting temporarily
wedged,
using some resources faster than they're recycled? I'd suspect that,
considering how many routers have what I'd call a "low-robustness"
IP stack.
Also make sure the router doesn't have its own throttling or some kind
DoS-protection enabled.
It's an old Apple "Dual Ethernet" Airport, the same one that I had in
2005 when I did not see this behavior. Of course, its firmware has
probably been updated a couple times since then, now 4.0.9. I just
looked through its configuration and definitely there are no settings
regarding throttling or denial-of-service. Unless someone can confirm
that this unit cannot be responsible, I guess I'll have to bypass it
to be sure.
On 2009 Aug 26, at 07:55, Jeff Johnson wrote:
I've noticed the same kind of behavior with my own app Vienna <http://www.vienna-rss.org/vienna2.php
> and would be very interested in the answer to your question,
though unfortunately I don't have an answer to it myself.
I will try to do more testing of this in the next week.
Great. I think we know what we're looking for now.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Macnetworkprog mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden