Discussion:
Comcast latency
(too old to reply)
mack
2008-04-30 04:43:46 UTC
Permalink
Has anyone else noticed a significant increase in latency within Comcast's network?

--
LR Mack McBride
Network Administrator
Alpha Red, Inc.
Steven M. Bellovin
2008-04-30 12:14:18 UTC
Permalink
On Tue, 29 Apr 2008 23:43:46 -0500
Post by mack
Has anyone else noticed a significant increase in latency within Comcast's network?
On one quick test, it looks normal to me from my house.


--Steve Bellovin, http://www.cs.columbia.edu/~smb
Jim Popovitch
2008-04-30 13:46:38 UTC
Permalink
Post by Steven M. Bellovin
On Tue, 29 Apr 2008 23:43:46 -0500
Post by mack
Has anyone else noticed a significant increase in latency within Comcast's network?
On one quick test, it looks normal to me from my house.
Looks can sometimes be deceiving. ;-) I've seen Comcast drop packets
left and right, but show 8Mbs/2Mbs on speed test sites. Other times
I've seen the opposite, zero PL but extremely high latency (seconds,
double digits!)... all the while non-dynamic web pages (probably
cached upstream) loaded just fine.

Look, I've always been a big fan of Comcast.... but at this point I
suspect they have just about over-engineered their network.

-Jim P.
John Osmon
2008-04-30 23:31:48 UTC
Permalink
Post by Jim Popovitch
Looks can sometimes be deceiving. ;-) I've seen Comcast drop packets
left and right, but show 8Mbs/2Mbs on speed test sites. Other times
I've seen the opposite, zero PL but extremely high latency (seconds,
double digits!)... all the while non-dynamic web pages (probably
cached upstream) loaded just fine.
I had mrt running across a Comcast connection that indicated a
14000 ms RTT for one packet out of several hundred. I wasn't sure
that was a trustworthy measurement, but it sure suprised me.

Another time, I had a VOIP conversation fall apart once with somone on a
Comcast link. When the jitter finally fell into a reasonable
range, the call sounded *great* again -- but there was a
4 second latency that had been introduced. It was wild. I kept
wondering -- "where was this buffered? Are we on a satellite
backup route?" We kept the conversation going for a while just for the
sheer novelty.

Overall, my data points on Comcast point to a pretty solid
service -- but the outliers are pretty impressive in their
own right.
Big Wave Dave
2008-04-30 14:26:03 UTC
Permalink
As luck would have it, I have been running Smokeping from my home
connection (Comcast Business) pointed towards Comcast Residential (and
other broadband providers) routers for a few weeks. I'm not sure when
you saw the latency increase, but if it has been in the past few
weeks, I might actually have useful data for you. If you (or others)
are interested in access, please contact me off-list.

Dave
Post by mack
Has anyone else noticed a significant increase in latency within Comcast's network?
--
LR Mack McBride
Network Administrator
Alpha Red, Inc.
_______________________________________________
NANOG mailing list
http://mailman.nanog.org/mailman/listinfo/nanog
Tim Thompson
2008-05-01 13:14:05 UTC
Permalink
Post by mack
Has anyone else noticed a significant increase in latency within
Comcast's network?
Here, in New England, I normally see 20% packet loss between Comcast and Level3 in NewYork using MTR. It has been as high as 60% in the past so this is an improvement.

-Tim Thompson
James Michael Keller
2008-05-01 14:25:11 UTC
Permalink
Post by Tim Thompson
Post by mack
Has anyone else noticed a significant increase in latency within
Comcast's network?
Here, in New England, I normally see 20% packet loss between Comcast and Level3 in NewYork using MTR. It has been as high as 60% in the past so this is an improvement.
-Tim Thompson
Remember that a number of ISP's 'core' routers are going to have ICMP
rate limits in place when one of their interfaces is the target. While
you may see high packet loss across an ISP link, unless you are getting
similar numbers from all hops past that point you aren't looking at real
packet loss, since the transit packets are getting through fine. It's
just the routers themselves that are ignoring requests or discarding
responses in favor of pushing routed packets.
--
James Michael Keller
Mike Fedyk
2008-05-01 21:32:07 UTC
Permalink
-----Original Message-----
Another time, I had a VOIP conversation fall apart once with
somone on a
Comcast link. When the jitter finally fell into a reasonable
range, the call sounded *great* again -- but there was a 4
second latency that had been introduced. It was wild. I
kept wondering -- "where was this buffered? Are we on a
satellite backup route?" We kept the conversation going for
a while just for the
sheer novelty.
The delay was probably introduced by an adaptive jitter buffer on one of the
VoIP end points. It is supposed to reduce the buffer as network conditions
improve. When it "fell apart" the buffer was increased until it was long
enough to reorder all of the packets.

It would be interesting to see if a call between those same end points would
reduce the size of their buffers after the network stabalized again.

Mike
Tim Thompson
2008-05-02 12:53:29 UTC
Permalink
Post by James Michael Keller
Post by Tim Thompson
Here, in New England, I normally see 20% packet loss between Comcast
and Level3 in NewYork using MTR. It has been as high as 60% in the
past so this is an improvement.
Remember that a number of ISP's 'core' routers are going to have ICMP
rate limits in place when one of their interfaces is the target. While
you may see high packet loss across an ISP link, unless you are getting
similar numbers from all hops past that point you aren't looking at real
packet loss, since the transit packets are getting through fine. It's
just the routers themselves that are ignoring requests or discarding
responses in favor of pushing routed packets.
Yes, the routers after Level3's edge router don't show similar packet
loss, so this must be the edge router just de-prioritizing ICMP. Thank
you for pointing out my mistake.

-Tim
jamie
2008-05-02 21:40:39 UTC
Permalink
You first, mister chicken-with-his-head-cut-off.

What's your plan?
Since nobody mentioned it yet, there are now less than 1000 days projected
http://www.potaroo.net/tools/ipv4/
Do you have an IPv6 plan?
How long do you think it will be until Sarbanes Oxley and SAS 70 auditors
start requiring disclosure of IPv4 exhaustion as a business continuity
risk, as well as the presence or lack thereof of an IPv6 plan?
When do you plan on telling your customers? (afterwards?)
Ahhh, you don't have any customers that have to plan to buy equipment 2
years in advance. Ok, I understand.
Mike.
ps. 1000 days assumes no rush, speculation, or hoarding. Do people do
that?
pps. Of course these are provocative comments for amusement. :)
ppps. Or not if you don't have any kind of IPv6 plan. Sorry, sorry...
+----------------- H U R R I C A N E - E L E C T R I C -----------------+
| Mike Leber Wholesale IPv4 and IPv6 Transit 510 580 4100 |
| Hurricane Electric Web Hosting Colocation AS6939 |
+-----------------------------------------------------------------------+
_______________________________________________
NANOG mailing list
http://mailman.nanog.org/mailman/listinfo/nanog
--
Would you like a little bit of legal advice?
NEVER let a scientist use the words "unanticipated" and "immediate" in the
same sentence.
Okay? Okay.
Patrick W. Gilmore
2008-05-02 21:47:44 UTC
Permalink
Post by jamie
You first, mister chicken-with-his-head-cut-off.
What's your plan?
Mike owns Hurricane Electric. HE.net has the most v6 routes, peering,
and pretty much any other metric you can dream up. His .sig says
"Wholesale IPv4 and IPv6 Transit". What do you think his plan is?

More important question: Perhaps you should spend 15 seconds
researching things before you send obviously ignorant comments to 10K
of your not-so-close friends?
--
TTFN,
patrick
Post by jamie
Since nobody mentioned it yet, there are now less than 1000 days projected
http://www.potaroo.net/tools/ipv4/
Do you have an IPv6 plan?
How long do you think it will be until Sarbanes Oxley and SAS 70 auditors
start requiring disclosure of IPv4 exhaustion as a business
continuity
risk, as well as the presence or lack thereof of an IPv6 plan?
When do you plan on telling your customers? (afterwards?)
Ahhh, you don't have any customers that have to plan to buy
equipment 2
years in advance. Ok, I understand.
Mike.
ps. 1000 days assumes no rush, speculation, or hoarding. Do people do
that?
pps. Of course these are provocative comments for amusement. :)
ppps. Or not if you don't have any kind of IPv6 plan. Sorry,
sorry...
+----------------- H U R R I C A N E - E L E C T R I C
-----------------+
| Mike Leber Wholesale IPv4 and IPv6 Transit 510 580 4100 |
| Hurricane Electric Web Hosting Colocation
AS6939 |
+
-----------------------------------------------------------------------+
_______________________________________________
NANOG mailing list
http://mailman.nanog.org/mailman/listinfo/nanog
--
Would you like a little bit of legal advice?
NEVER let a scientist use the words "unanticipated" and "immediate" in the
same sentence.
Okay? Okay.
_______________________________________________
NANOG mailing list
http://mailman.nanog.org/mailman/listinfo/nanog
Sean Figgins
2008-05-02 21:49:40 UTC
Permalink
Since nobody mentioned it yet, there are now less than 1000 days projected
No worries, the Internet is going to end in 2010, and the world ends on
December 21, 2012. I don't think we'll be needing IPv6 in that case.

Has anyone ever figured out how to make multi-homing of customers who
only have a /64 assigned to them work? Are the routers on the going to
be able to handle the billion routing prefixed that will be introduced?
Are there any IP Management software packages that won't bankrupt the
world's economy for IPv6 charges?

Maybe the world really will end, and it's all due to IPv6!

-Sean
David Conrad
2008-05-02 21:57:36 UTC
Permalink
Post by Sean Figgins
Has anyone ever figured out how to make multi-homing of customers who
only have a /64 assigned to them work?
Same way you make multi-homing of customers who only have a IPv4 /32
assigned to them work, i.e., not well.
Post by Sean Figgins
Maybe the world really will end, and it's all due to IPv6!
Internet doomed, MPEG stalled at 11:00... :-)

Regards,
-drc
Joel Jaeggli
2008-05-02 22:08:22 UTC
Permalink
Post by Sean Figgins
Since nobody mentioned it yet, there are now less than 1000 days projected
No worries, the Internet is going to end in 2010, and the world ends on
December 21, 2012. I don't think we'll be needing IPv6 in that case.
Has anyone ever figured out how to make multi-homing of customers who
only have a /64 assigned to them work?
how are your /32 v4 announcements working out?

longest prefix I carry in my v6 table are a /48s...

There are only 28224 ASes in announced in the v4 routing system how
many non-agregatable announcements will they represent if they all
participate in v6 tomorrow?
Post by Sean Figgins
Are the routers on the going to
be able to handle the billion routing prefixed that will be introduced?
Are there any IP Management software packages that won't bankrupt the
world's economy for IPv6 charges?
Maybe the world really will end, and it's all due to IPv6!
-Sean
_______________________________________________
NANOG mailing list
http://mailman.nanog.org/mailman/listinfo/nanog
Chris McDonald
2008-05-02 21:55:48 UTC
Permalink
Mike and HE are all over that ipv6
Post by jamie
You first, mister chicken-with-his-head-cut-off.
What's your plan?
Since nobody mentioned it yet, there are now less than 1000 days projected
http://www.potaroo.net/tools/ipv4/
Do you have an IPv6 plan?
How long do you think it will be until Sarbanes Oxley and SAS 70 auditors
start requiring disclosure of IPv4 exhaustion as a business continuity
risk, as well as the presence or lack thereof of an IPv6 plan?
When do you plan on telling your customers? (afterwards?)
Ahhh, you don't have any customers that have to plan to buy equipment 2
years in advance. Ok, I understand.
Mike.
ps. 1000 days assumes no rush, speculation, or hoarding. Do people do
that?
pps. Of course these are provocative comments for amusement. :)
ppps. Or not if you don't have any kind of IPv6 plan. Sorry, sorry...
+----------------- H U R R I C A N E - E L E C T R I C -----------------+
| Mike Leber Wholesale IPv4 and IPv6 Transit 510 580 4100 |
| Hurricane Electric Web Hosting Colocation AS6939 |
+-----------------------------------------------------------------------+
_______________________________________________
NANOG mailing list
http://mailman.nanog.org/mailman/listinfo/nanog
--
Would you like a little bit of legal advice?
NEVER let a scientist use the words "unanticipated" and "immediate" in the
same sentence.
Okay? Okay.
_______________________________________________
NANOG mailing list
http://mailman.nanog.org/mailman/listinfo/nanog
--
Sent from Gmail for mobile | mobile.google.com
Patrick W. Gilmore
2008-05-03 04:42:03 UTC
Permalink
P.S.
10K of your not-so-close friends?
does this mean this list has 10.000 subscribers ?
I've heard all kinds of numbers, you can probably dig something out of
the archives.

But my understanding is there are far greater than 10K mailboxes which
receive NANOG, especially if you include exploders. Could someone
from the mail admin team confirm?
--
TTFN,
patrick
Tim Yocum
2008-05-03 04:48:35 UTC
Permalink
All,

Patrick is correct - the subscriber count is just north of 10k; likely
far greater readership considering web archives, remailers, etc.

- Tim
Post by Patrick W. Gilmore
P.S.
10K of your not-so-close friends?
does this mean this list has 10.000 subscribers ?
I've heard all kinds of numbers, you can probably dig something out of
the archives.
But my understanding is there are far greater than 10K mailboxes which
receive NANOG, especially if you include exploders. Could someone
from the mail admin team confirm?
--
TTFN,
patrick
_______________________________________________
NANOG mailing list
http://mailman.nanog.org/mailman/listinfo/nanog
JC Dill
2008-05-13 18:25:37 UTC
Permalink
Post by Tim Yocum
Patrick is correct - the subscriber count is just north of 10k; likely
far greater readership considering web archives, remailers, etc.
However... subscribership != readership. There are always many
subscribers who don't actively read every post, or every day. (I'm just
now catching up on 10 days of unread posts - it would be easy to declare
email bankruptcy and just mark it all read...) All you have to do is
post an important announcement (such as "the mailing list is moving to a
new server") and then notice how few read it to prove the truth of this
fact. :-) Also, note how few read the list's welcome message, read the
FAQ, etc.

jc
Patrick W. Gilmore
2008-05-13 18:36:33 UTC
Permalink
Post by JC Dill
Post by Tim Yocum
Patrick is correct - the subscriber count is just north of 10k; likely
far greater readership considering web archives, remailers, etc.
However... subscribership != readership. There are always many
subscribers who don't actively read every post, or every day.
And there are "subscribers" which are actually exploders with many
people (mailboxes?) behind them.

Either way, it's not a small number of people / mailboxes.
--
TTFN,
patrick
Loading...