Cisco whinge thread :)

s0ck

s0ck

s0ck

Soldato
Joined
18 Oct 2002
Posts
5,299
Something to read if you're bored :)

I've tried three routers now.

I've gateway'd through a 2500 (no DSL WIC), a 1700 and an 800 series router.
All of these add 10-12ms to my ICMP packets and in game pings are roughly double that of a 30 quid SOHO router.
I have to surmise that the following:

1. DSL WICs add unnecessary latency
or
2. Something in the IOS/firmware of all three needs tweaking

I ruled out the 2500 as being old and tried a 1700. Same problem. I then said to myself, "gee, this one is pretty old too" and acquired an 877; they're all the same though.

I'm now back on my el cheapo router because a 25 ping in Quakelive is preferable to a 65 ping.


The only thing I can possibly come up with is store/forward is used by Cisco and cut through is used by cheap stuff? I suspect this isn't true though and they all use the former.

I'd be really interested to hear anyone else's findings on this or suggestions for improvement.
 
hmmm, I can't say that I've noticed problems on mine (but then my 877 is plugged into a cable modem for now). On your 877, try adjusting the MTU on the Dialer / ATM interface
Depending on the IOS version it'll be either
mtu 1492
ip mtu 1492
ip adjust-mss 1492

then validate with a show interface and look for the MTU statement. You'll have to shut /no shut the interface.

Cisco's should have CEF enabled by default which allows for faster routing of packets.

That's all I can think of for the moment :)
 
Something must be wrong there to be honest. When testing my lab through 4 routers (also all 2500 series) I never had latency that high. how exactly are you ruling out the routers?

- Pea0n
 
Chief Wiggum said:
hmmm, I can't say that I've noticed problems on mine (but then my 877 is plugged into a cable modem for now). On your 877, try adjusting the MTU on the Dialer / ATM interface
Depending on the IOS version it'll be either
mtu 1492
ip mtu 1492
ip adjust-mss 1492

then validate with a show interface and look for the MTU statement. You'll have to shut /no shut the interface.

Cisco's should have CEF enabled by default which allows for faster routing of packets.

That's all I can think of for the moment :)
Click to expand...

I have a feeling I've turned CEF off on the 877 (can't remember why immediately) but it would have been at defaults for the others, will check.
Will try the MTU stuff too, cheers :D

Pea0n said:
Something must be wrong there to be honest. When testing my lab through 4 routers (also all 2500 series) I never had latency that high. how exactly are you ruling out the routers?

- Pea0n
Click to expand...

paradigm said:
Not got latency that high on either my 2651XM (Cable) or my 877 (ADSL).
Click to expand...

If you've got some time on your hands (I know, who has eh?) it would be great if you could ping www.jolt.co.uk through your Cisco equipment and then try it on a SOHO unit (with no Cisco in between).

I will be very surprised if you don't notice an extra 10ms on your Cisco kit (final hop).
 
Don't know if it's just me but is Jolt just pinging high anyway?

50Mbit Virgin Cable here, currently using a Juniper SSG20.

Code: 
Pinging jolt.co.uk [84.234.17.86] with 32 bytes of data:
Reply from 84.234.17.86: bytes=32 time=52ms TTL=119
Reply from 84.234.17.86: bytes=32 time=46ms TTL=119
Reply from 84.234.17.86: bytes=32 time=22ms TTL=119
Reply from 84.234.17.86: bytes=32 time=24ms TTL=119

Ping statistics for 84.234.17.86:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 22ms, Maximum = 52ms, Average = 36ms

That seemed to be fairly typical of my results, vs:

Code: 
Pinging bbc.co.uk [212.58.224.138] with 32 bytes of data:
Reply from 212.58.224.138: bytes=32 time=18ms TTL=118
Reply from 212.58.224.138: bytes=32 time=17ms TTL=118
Reply from 212.58.224.138: bytes=32 time=15ms TTL=118
Reply from 212.58.224.138: bytes=32 time=20ms TTL=118

Ping statistics for 212.58.224.138:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 15ms, Maximum = 20ms, Average = 17ms

And most other decent sites that I tried.
 
--- www.jolt.co.uk ping statistics ---
250 packets transmitted, 250 received, 0% packet loss
rtt min/avg/max/mdev = 2.377/5.600/12.756/2.622 ms

There first 4 hops there are Cisco...

or from home...

Success Rate is 100 percent (250/250), round-trip time min/avg/max=13/17/19 ms

There's a few Cisco hops in there too and it's over DSL. I'm afraid you either have a faulty WIC or something's wrong in your config. Mountains of business ISPs use these devices for CPE, if there was a problem it would have been found and fixed long long ago.
 
And for comparison, from a Juniper Edge router (going over a 10Gig LINX connection)...

[email protected]4> ping www.jolt.co.uk rapid count 1000
PING www.jolt.co.uk (84.234.17.86): 56 data bytes
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
--- www.jolt.co.uk ping statistics ---
1000 packets transmitted, 1000 packets received, 0% packet loss
round-trip min/avg/max/stddev = 2.668/3.568/48.743/2.344 ms

[email protected]4> traceroute www.jolt.co.uk
traceroute to www.jolt.co.uk (84.234.17.86), 30 hops max, 40 byte packets
1 ge-3-3-0.mpr1.lhr3.uk.above.net (195.66.226.76) 1.144 ms 1.217 ms 1.550 ms
2 so-1-0-0.mpr1.lhr2.uk.above.net (64.125.28.38) 36.841 ms 2.231 ms 2.153 ms
3 79.141.39.212.available.above.net (79.141.39.212) 31.967 ms 197.847 ms 3.809 ms
4 bs1-core.netrino.co.uk (84.234.19.5) 3.511 ms 3.329 ms 3.165 ms
5 bruce.ferrago.net (84.234.17.86) 3.445 ms 3.372 ms 2.935 ms

Looks fine to me. But it's a bad time to test, LINX is having a lie down today, just had to disable our 224 peerings...
 
Yamahahahahaha said:
You've done something funny to the config. Wipe it and start over.

If that still doesn't fix it, put the newest modem firmware on the 877. (4.0.0.19 afaik?)
Click to expand...

Quite possibly! I don't really know what I'm doing :D

bigredshark said:
Success Rate is 100 percent (250/250), round-trip time min/avg/max=13/17/19 ms

There's a few Cisco hops in there too and it's over DSL.
Click to expand...

What router is that m8? 2800? :p

I'm afraid you either have a faulty WIC or something's wrong in your config. Mountains of business ISPs use these devices for CPE, if there was a problem it would have been found and fixed long long ago.
Click to expand...

That's the thing, it's not a problem as such. I can't see 10ms ever been enough for someone to pick up in a business sense; interactive traffic would have been near identical.
It's only when you get someone anal about their uber gaming pingz, y0, does it get noticed. Plus, back in the day, SOHO stuff might have been no better.

Perhaps I have unrealistic expectations about a 300 quid, older, Cisco router vs a very recent mass market £30 unit?
 
s0ck said:
What router is that m8? 2800? :p

That's the thing, it's not a problem as such. I can't see 10ms ever been enough for someone to pick up in a business sense; interactive traffic would have been near identical.
It's only when you get someone anal about their uber gaming pingz, y0, does it get noticed. Plus, back in the day, SOHO stuff might have been no better.

Perhaps I have unrealistic expectations about a 300 quid, older, Cisco router vs a very recent mass market £30 unit?
Click to expand...

The router running the DSL is a 2821 with a standard WIC1-ADSL.

Believe me businesses would notice the latency, far more than you would, the amount of money some companies are prepared to spend on reducing latency is frightening. We used to regularly get complaints about high latency when they were getting an average of 35ms (which, if the line is interleaved isn't terrible at all). Seriously, there is no way it's a widespread Cisco bug that hasn't been noticed, if it was on one particular platform it would be an outside chance but that doesn't seem to be the case here.

Have you checked your LAN side? Could be a LAN duplex mismatch too or a faulty interface/cable unless you've tried pinging from the router and got the same result.

Post your config if you can? and who's your ISP?
 
Current ISP is Sky. Last ISP was Enta. Same on both. Config is about is basic as it gets. Will have a squint tonight.

Has the WIC1-ADSL been revised over time? Has the chipset changed, etc?
 
s0ck said:
Current ISP is Sky. Last ISP was Enta. Same on both. Config is about is basic as it gets. Will have a squint tonight.

Has the WIC1-ADSL been revised over time? Has the chipset changed, etc?
Click to expand...

There are variations but they're functionally identical. It's sort of deprecated now in favour of the ADSL2+ HWIC that's available anyway - but that won't work in your routers.

It's also possible you've just got really unlucky in terms of the IOS version you're using, bugs aren't unheard of there...
 
bigredshark said:
But it's a bad time to test, LINX is having a lie down today, just had to disable our 224 peerings...
Click to expand...

I bet there are some providers don't have the spare capacity over other routes to take down their LINX connections :o :p

Everything was slow via my home ISP yesterday (Facebook, BBC etc) - the LINX traffic graph was funny though, like a cliff.
 
daz said:
I bet there are some providers don't have the spare capacity over other routes to take down their LINX connections :o :p

Everything was slow via my home ISP yesterday (Facebook, BBC etc) - the LINX traffic graph was funny though, like a cliff.
Click to expand...

I do know a few providers in that position, and a few more who're dumb and only have presence on the one network rather than both.

They lost an 80Gbps ISL at one point, which had to hurt. Looking better now but I seemed to trigger an avalanche of shutdowns when I announced we were taking down our sessions this morning. We'll likely be down until tomorrow morning now, as I suspect will others...
 
Code: 
Microsoft Windows [Version 6.1.7600]
Copyright (c) 2009 Microsoft Corporation.  All rights reserved.

C:\Windows\system32>ping www.jolt.co.uk

Pinging www.jolt.co.uk [84.234.17.86] with 32 bytes of data:
Reply from 84.234.17.86: bytes=32 time=14ms TTL=121
Reply from 84.234.17.86: bytes=32 time=14ms TTL=121
Reply from 84.234.17.86: bytes=32 time=14ms TTL=121
Reply from 84.234.17.86: bytes=32 time=14ms TTL=121

Ping statistics for 84.234.17.86:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 14ms, Maximum = 14ms, Average = 14ms

C:\Windows\system32>route add 84.234.17.86 mask 255.255.255.255 10.1.1.1
 OK!

C:\Windows\system32>ping jolt.co.uk

Pinging jolt.co.uk [84.234.17.86] with 32 bytes of data:
Reply from 84.234.17.86: bytes=32 time=15ms TTL=121
Reply from 84.234.17.86: bytes=32 time=20ms TTL=121
Reply from 84.234.17.86: bytes=32 time=20ms TTL=121
Reply from 84.234.17.86: bytes=32 time=16ms TTL=121

Ping statistics for 84.234.17.86:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 15ms, Maximum = 20ms, Average = 17ms

First route is Cable, via a 2651XM. Second route is ADSL via an 877.
 
Tracerts would be more valuable to be honest as its useless knowing if a ping is higher unless you know what is causing it. Even from the same PC as above it could be totally different routes through different ISPs

- Pea0n
 
Traceroutes are actually rubbish troubleshooting tools for that, all they tell you is the path, the time it takes a device to respond to ICMP and the time it takes to forward traffic are far from the same thing. I get better response from the BBC than I do from some of our older routers along the path, because they forward traffic in the CFEB but ICMP requires routing engine input to process.
 
The entire point of the tracert is that it tells you which hop is giving the latency. He says it is the router. A tracert will show where large amounts are being added, even a path ping would assist with this. A simple ping itself give naff all except the total, and that latency can come from anywhere as it tells you nothing, just the total. It could come from the local router, it could come form the last hop or somewhere in the middle. no way to tell...

- Pea0n
 
Pinging www.jolt.co.uk [84.234.17.86] with 32 bytes of data:
Reply from 84.234.17.86: bytes=32 time=14ms TTL=121
Reply from 84.234.17.86: bytes=32 time=14ms TTL=121
Reply from 84.234.17.86: bytes=32 time=13ms TTL=121
Reply from 84.234.17.86: bytes=32 time=14ms TTL=121

Ping statistics for 84.234.17.86:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 13ms, Maximum = 14ms, Average = 13ms

From an 877 running IOS 150-1
 
You must log in or register to reply here.
Back
Top Bottom