50M Think Broadband Ping Monitor Results (POST YOURS) - Page 238

qasdfdsaq · 03-04-2013, 04:51

Quote:

Originally Posted by Qtx

How do you get packetloss like that without the maximum latency increasing? Is it something to be with it being intermittent on and off over the course of every few seconds?

The maximum latency is limited by the maximum number of retries and delay time configured on the system (which signs seem to indicate have been reduced)

---------- Post added at 04:48 ---------- Previous post was at 04:46 ----------

Quote:

Originally Posted by Sephiroth

when wireless was ON (whether or not used) the ramping effect occurred; when wireless was OFF there was no ramping. We deduced (with no absolute proof) that the hib CPU was struggling, although why doesn't that happen to all SHs?

Because wireless is a notoriously variable medium and not all SHs exist in the same wireless environment?

---------- Post added at 04:51 ---------- Previous post was at 04:48 ----------

Quote:

Originally Posted by thenry

I can't vouch for Ravens experience. it happened here with or without wifi enabled. being in combo mode caused problems. modem mode was steady.

Others have seen the same phenomenon with wired devices, so it seems certain types or patterns of traffic cause this behaviour on the SH, regardless of where it's coming from.

Qtx · 03-04-2013, 05:15

Got a feeling the firmware will get the blame again

Now if only you could use a web interface to upload saved superhub settings that had been modified to set usually hidden options, that allowed the shub to download a better firmware from a LAN tftp server. That would be ace

qasdfdsaq · 03-04-2013, 06:12

Quote:

Originally Posted by Qtx

I would have thought that ping replies were so basic and low level that cpu load wouldn't affect them (or the silly low amount of ram) but im far from an expert in this area and have no idea how it actually works.

And what would you think responds to the pings other than the CPU?

Quote:

Just did a quick google to see what chipset and if a single chip is used for all the functions but didn't find anything enlightening, except for a possibly too simple to be true way of bypassing the shubs security.

The chipset and configuration of the SH has been discussed here previously.

The cable modem components are integrated in the main CPU. The LAN ports are handled by a dedicated switch chip.

Quote:

If ping replies are not dealt with by the chips own routines and is affected by firmware, that could be another avenue.

Ping replies are dealt with by the OS kernel.

---------- Post added at 06:09 ---------- Previous post was at 06:08 ----------

Quote:

Originally Posted by GuestUK

As for the traffic, it's really odd. Either a packet gets through perfectly fine, or it doesn't get through at all, there is no inbetween!

Well there is... There's packets that get through partly and are corrected by ECC and there's packets that do not get through on first attempt and are retried.

---------- Post added at 06:12 ---------- Previous post was at 06:09 ----------

Quote:

Originally Posted by Qtx

that allowed the shub to download a better firmware from a LAN tftp server. That would be ace

Don't even go there. You don't want to try it.

Qtx · 03-04-2013, 06:46

Quote:

Originally Posted by qasdfdsaq

And what would you think responds to the pings other than the CPU?

The chipset and configuration of the SH has been discussed here previously.

The cable modem components are integrated in the main CPU. The LAN ports are handled by a dedicated switch chip.

Ping replies are dealt with by the OS kerne

Thanks for the info. Still not sure of its an atheros/broadcom/rtl type chipset though.

I did wonder if the actual IP stack was handled on a separate chip in the same way as the LAN ports being handled with a different chip. Specialised dedicated chips taking work from the cpu is common in other areas but I wasn't sure in router technology.

So the common factors with the ramping are the cpu and kernel/firmware.

---------- Post added at 06:46 ---------- Previous post was at 06:39 ----------

Quote:

Originally Posted by qasdfdsaq

Don't even go there. You don't want to try it.

Wouldn't let them give me a superhub, so never had one, never want one and will never get to try it

When trying to fix their obvious utilisation problem, they stalled with a repull and palming me off with a vmdg280. Obviously none of that fixed their overutilisation problem but means I had beern downgraded from a modem to the 280. Didn't want to downgrade it any futher lol.

qasdfdsaq · 03-04-2013, 07:01

Quote:

Originally Posted by Qtx

Thanks for the info. Still not sure of its an atheros/broadcom/rtl type chipset though.

http://www.cableforum.co.uk/board/35272167-post580.html
http://www.cableforum.co.uk/board/35...-post1097.html

Quote:

I did wonder if the actual IP stack was handled on a separate chip in the same way as the LAN ports being handled with a different chip.

Many routers, including the Superhub, can use hardware accelerated NAT which does this, but only for traffic going through the Superhub and not to/from it.

Quote:

Wouldn't let them give me a superhub, so never had one, never want one and will never get to try it

I never let them give me a Superhub either but I still ended up with two after I left VM...

Qtx · 03-04-2013, 10:12

Quote:

Originally Posted by qasdfdsaq

http://www.cableforum.co.uk/board/35272167-post580.html
http://www.cableforum.co.uk/board/35...-post1097.html

Many routers, including the Superhub, can use hardware accelerated NAT which does this, but only for traffic going through the Superhub and not to/from it.

I never let them give me a Superhub either but I still ended up with two after I left VM...

Thanks for digging out the links for me. Exactly the info I was looking for. I was looking for info on the CG3101D last night which is how I came across the alternative firmwares. No one really wants to play with that stuff due to the lowly 8mb of memory.

Not gonna ask where you got your shubs. My vmdg280 left here in a blue body bag

horseman · 03-04-2013, 14:07

Quote:

Originally Posted by qasdfdsaq

….
Well there is... There's packets that get through partly and are corrected by ECC and there's packets that do not get through on first attempt and are retried…...

You can almost hear the "cogs turning" as an inspired thought enlightens the Stygian darkness…..
Only a relatively short time before someone mentions "Solar flares", "alpha particles" and "Double parity errors"……

qasdfdsaq · 03-04-2013, 19:49

Quote:

Originally Posted by Qtx

Thanks for digging out the links for me. Exactly the info I was looking for. I was looking for info on the CG3101D last night which is how I came across the alternative firmwares. No one really wants to play with that stuff due to the lowly 8mb of memory.

It has 64MB memory, which is above average and typical for high-end routers a couple years ago

---------- Post added at 19:49 ---------- Previous post was at 19:48 ----------

Quote:

Originally Posted by horseman

Only a relatively short time before someone mentions "Solar flares", "alpha particles" and "Double parity errors"……

I would have thought beta particles would be more relevant

partyrock · 04-04-2013, 21:10

Ex NTL
Upstream 36.5 dBmV
DS 1 1.0 dBmV SNR 43.2 dB
DS 2 1.1 dBmV SNR Same
DS 3 1.1 dBmV SNR Same
DS 4 0.5 dBmV SNR 43.1 dB

Today is worser and i never seen it so bad, Is the green street box overloaded because of the school holidays?

Last weeks 26/03/2013 the usual Xbox 360 black ops 2 and no one eles was on the network (much smoother graph!
http://www.thinkbroadband.com/ping/s...6-03-2013.html

CMTS

RealDiamond · 05-04-2013, 13:20

Any one else lose the network today.

default gate way was fine but Ping and packet loss have started spiking a bit even on the gateway..

Jumping · 05-04-2013, 13:29

Mine is all fine and normal (Uddingston UBR)

Martin_D · 05-04-2013, 13:55

Mine is fine too (Uddingston UBR)

RealDiamond · 05-04-2013, 15:53

Quote:

Originally Posted by Martin Dee 11

Mine is fine too (Uddingston UBR)

Ta I was hoping it was VM turning off the unannounced new Traffic shaping that had effected patch downloads on Sony systems but if it was not VM wide then it was not likely that... must be just local down time o well..

GazCBG · 07-04-2013, 15:09

Here is mine, I am on 120Mb / 12Mb
Superhub in modem mode with a Asus RT-N66U Dark Knight Gigabit, Wireless N900 Router

pcoventry · 08-04-2013, 18:58

Hre's my graph after I disconnected

http://www.thinkbroadband.com/ping/s...e91e9c97a.html

Oodly someone else is now using the IP address I had but the IP Address TBB was tracking was the one given out by my AP and not the one VM gave me.

It replys to pings too. I find that confusing. As I said it's not the IP address VM gave me. It was one my router assigned to me. Not sure how or why