Simon W
Regular Contributor
Back in the days of fiddling with all-in-one routers and their integrated modems connected to crappy ADSL lines, frequent reboots (eg. as might be necessary when making unrelated configuration changes, or people unwittingly unplugging them!) often had a detrimental impact on link speed - your ISP's cabinet would see the connection go down/up, sometimes frequently, and assume there was some problem so negotiate a more conservative rate (=slower), which would either persist for a number of hours/days before then renegotiating back higher.. or in worst cases never go back up again without a call to your ISP to convince them to reset your line. One of the (many) reasons I prefer a separate modem - to keep to link leaving the premises as stable as possible.
Winding forwards to present day and up until now I've been running a double-NAT'd solution - g.fast modem connected to ISP provided router, in turn connected to my Asus AX88U and onwards (business level contract offering 4G cellar resilience should the line go down, along with automatically logging a line fault call and a decent SLA to getting it fixed - all works well but requires their box is used.. which otherwise sucks!.. so this was my workaround) - yet I'm now looking to remove this additional hop. From the AX88U WAN configuration perspective simple means reconfiguring from a Static IP to PPPoE connection type and popping in my credentials.
All works fine (either way) and in testing I get the expected benefits from simplifying: slightly lower latency (~2ms less) and ever-so-slightly higher up/down ceiling.
But one thing I've noticed, recognising the AX88U WAN interface is currently going up and down quite a lot as I play with various configurations, is that while my previous setup produced extremely consistent ping response to a known external test point (10-11ms), in my new configuration this seems to vary between WAN connections (7-9ms) though remains consistent while the interface is up. No other known forces are at play here, wound things right back and this is testing from either the AX88U itself or a directly connected ethernet client with no other network load. Reboot, let settle, do lots of tests = all the same. Reboot again, settle again, do lots more test, equally as consistent but slightly different from the previous tests.
Question: is there some variability at play when a PPPoE connection is established, ie. connecting to a better/worse entry point (forgive my terminology as how this exactly works is beyond my knowledge, hence the question)? As opposed to my previous setup where the PPPoE was (re-)established very infrequently by the ISP's box and therefore remained consistent? Or complete rubbish.. and clearly something else must be at play (I'm hoping for the latter!)?
Thanks.
Winding forwards to present day and up until now I've been running a double-NAT'd solution - g.fast modem connected to ISP provided router, in turn connected to my Asus AX88U and onwards (business level contract offering 4G cellar resilience should the line go down, along with automatically logging a line fault call and a decent SLA to getting it fixed - all works well but requires their box is used.. which otherwise sucks!.. so this was my workaround) - yet I'm now looking to remove this additional hop. From the AX88U WAN configuration perspective simple means reconfiguring from a Static IP to PPPoE connection type and popping in my credentials.
All works fine (either way) and in testing I get the expected benefits from simplifying: slightly lower latency (~2ms less) and ever-so-slightly higher up/down ceiling.
But one thing I've noticed, recognising the AX88U WAN interface is currently going up and down quite a lot as I play with various configurations, is that while my previous setup produced extremely consistent ping response to a known external test point (10-11ms), in my new configuration this seems to vary between WAN connections (7-9ms) though remains consistent while the interface is up. No other known forces are at play here, wound things right back and this is testing from either the AX88U itself or a directly connected ethernet client with no other network load. Reboot, let settle, do lots of tests = all the same. Reboot again, settle again, do lots more test, equally as consistent but slightly different from the previous tests.
Question: is there some variability at play when a PPPoE connection is established, ie. connecting to a better/worse entry point (forgive my terminology as how this exactly works is beyond my knowledge, hence the question)? As opposed to my previous setup where the PPPoE was (re-)established very infrequently by the ISP's box and therefore remained consistent? Or complete rubbish.. and clearly something else must be at play (I'm hoping for the latter!)?
Thanks.
Last edited: