R0CKY
TSI Rocky
Premium,VIP
join:2005-05-19
Chatham, ON
1 edit | FYI: Testing before the next Phase of changes!
Hi Gang,
We're just in the middle of testing the new network that we're about to jump over to, and it seems, so long as the the new network is ready for volume, that things are about to get much better for the TekSavvy people! The new MTU default will be 1492 once we roll over and everything will be Ethernet/IP based, so we'll be riding over newer technology head-to-toe!
More to come soon!
Rocky
--
Rocky/TSIGUY - TekSavvy Solutions Inc. |
|
The Flash
You don't win friends with salad
Premium
join:2002-10-17
Toronto, ON | That's great cause it seems like your user database is booming, everywhere I hear it's join Teksavvy..Teksavvy this and that  |
|
jfmezei
Premium
join:2007-01-03
Beaconsfield, QC
 ·ELECTRONICBOX
| reply to R0CKY
> riding over newer technology head-to-toe!
Can we expect a ping difference between us and piero ?
What is the expected downtime for this cutover ? 10 minute ? 4 hours ?
Will our pings to piero change at all or will it remain roughly the same ? |
|
R0CKY
TSI Rocky
Premium,VIP
join:2005-05-19
Chatham, ON
| said by jfmezei  :> riding over newer technology head-to-toe! Can we expect a ping difference between us and piero ? What is the expected downtime for this cutover ? 10 minute ? 4 hours ? Will our pings to piero change at all or will it remain roughly the same ? No sure yet.... When an account hits the BAS it will go directly to the OE cloud instead of the ATM cloud.... Not sure just how much it will be as a result.
I suspect that we're slowly going to swing some tunnels over to the new network. This portion might not see any downtime.
There are a few unknowns to me at the moment as we're in the middle of increasing space at the same time in the Bell space, so the downtime time might be more a matter of unplugging and replugging some hardware rather than an actual service connection thing. So if all is going well you might not see any downtime and/or you might see a matter of seconds downtime. Part of the next phase includes adding a few more nodes that would allow us to bypass the main router and hand-off DSL sessions to a could more places as fail-over and load balancing. This portion will likely be the main downtime, which won't be for another month solid.
More to come soon!
Rocky
--
Rocky/TSIGUY - TekSavvy Solutions Inc. |
|
R0CKY
TSI Rocky
Premium,VIP
join:2005-05-19
Chatham, ON
| reply to The Flash
said by The Flash :That's great cause it seems like your user database is booming, everywhere I hear it's join Teksavvy..Teksavvy this and that Yeah.... this will be nice as the ATM cloud is overly saturated and, well, for the time-being at least, we're going to be one of the only ones over the Ethernet-only cloud so we'll have tons of unaffected space I'm told. The ATM side isn't being invested into anymore as it's a dying network that Bell no longer wants to use... as a result the more people they take off of it the better it will be.
Growth: yeah... We're kicking butt these days, and we're expecting to pick up more momentum in the near future as well! 
Regards,
Rocky
--
Rocky/TSIGUY - TekSavvy Solutions Inc. |
|
TSI Gabe
Premium,VIP
join:2007-01-03
Chatham, ON | There shouldn't be any down time. Basically the old and the new system will become live and all the users will have to do to be "transfered" over to the new ethernet system is to log off and back in. |
|
avernar
join:2002-05-23
Mississauga, ON
clubs:
| reply to R0CKY
said by R0CKY :The new MTU default will be 1492 Is it currently less than 1492? I have my Linux router limiting forwarded connections to 1492. |
|
R0CKY
TSI Rocky
Premium,VIP
join:2005-05-19
Chatham, ON
| said by avernar :said by R0CKY :The new MTU default will be 1492 Is it currently less than 1492? I have my Linux router limiting forwarded connections to 1492. Yes... it's 1452.
--
Rocky/TSIGUY - TekSavvy Solutions Inc. |
|
PlatooN
@teksavvy.com | thats interestnig. my router as well is set for 1492 and i've had NO problems with phantom pages or things not loading properly.
*shrug*
YOU GUYS ROCK! |
|
TSI Marc
Premium,VIP
join:2006-06-23
Chatham, ON
| yes it would work fine either way. from your computer to our equipment there is a tunnel(L2TP) that everything goes through. our equpiment has a setting that says to reassemble all packets arriving fragmented.. the difference now is that that will no longer need to happen. its a very slight performance increase.
--
Marc |
|
hidea
join:2007-02-12
North York, ON | Please, ENGLISH. Not Moonspeak. :P
What does it mean for a Graphic Designer? Will I be able to rape the Internet faster (faster load up time, faster page load, etc. etc.)? |
|
jfmezei
Premium
join:2007-01-03
Beaconsfield, QC
·ELECTRONICBOX
| reply to PlatooN
Just because your router as a default MTU of 1492 doesn't mean that your connections are forced to that value.
If I set my router to 1492, and go to the speedguide.net tests, it shows MTU of 1452. Similarly, my VMS hosts are set to have MTU of 1500, but when talking to the internet, the connections drop down to 1452. |
|
avernar
join:2002-05-23
Mississauga, ON
clubs:
| The reason that you see 1452 when you're using a 1492 MTU is because Teksavvy's L2TP tunnel is properly breaking the packet into two pieces and then reassembling them again.
What PlatooN was referring to is that most routers and hosts on the internet do not reassemble fragmented packets properly. Also, some routers on a small MTU pipe don't return a Fragmentation Required ICMP reply and silently drop packets.
If I don't do a MSS rewrite rule on my Linux router/firewall I'll get failures with bulk TCP transfers (ftp/http/etc). The problem is probably worse now that I'm running 9000 MTU on my LAN.
So basically there's two ways a higher MTU will work:
1) Something on one end of the narrow pipe fragments the packets and something on the other end properly reassembles them.
2) Something anywhere on the pipe rewrites the TCP SYN packets and drops the MSS value which effectively limits the MTU of that TCP connection. |
|
JayMan
Whoot
Premium
join:2002-06-05
Earth
·TekSavvy Solutions..
| reply to R0CKY
My firewall is showing.
Feb 13 21:00:49 mpd: [pppoe] setting interface ng0 MTU to 1492 bytes
Is there anything else I should look for with these new changes to know if they have happened in my area.
--
------
»listserver.weatherserver.net
Weather Alerts, Traffic Alerts, Toronto Fire CAD Alerts - Emailed 24/7/365 |
|
TSI Gabe
Premium,VIP
join:2007-01-03
Chatham, ON | If your MTU changed to 1492 then it is most likely you are on the new network. In which area are you located? |
|
JayMan
Whoot
Premium
join:2002-06-05
Earth | Pembroke, about 100km from Ottawa and I connect to bas11-ottawa23 |
|
Mark Rejhon
join:2004-02-02
Ottawa, ON | I'm in Ottawa. I've noticed my pings get gradually higher to piero -- it's about 58-60ms now where it used to be 35ms.
Since I connect to piero, I assume I'm still on the old ATM network...? |
|
R0CKY
TSI Rocky
Premium,VIP
join:2005-05-19
Chatham, ON
1 edit | Hmm.... Not sure which areas are now switched but the fact that you are seeing degradation would suggest there's something. I'll have someone investigate.
Rocky
--
Rocky/TSIGUY - TekSavvy Solutions Inc. |
|
TSI Marc
Premium,VIP
join:2006-06-23
Chatham, ON
| reply to Mark Rejhon
said by Mark Rejhon :I'm in Ottawa. I've noticed my pings get gradually higher to piero -- it's about 58-60ms now where it used to be 35ms. Since I connect to piero, I assume I'm still on the old ATM network...? Your pings to piero should be something like 8ms in Ottawa. Sounds like you're probably on interleave. Does your connection have any problems?
If you've disconnected and reconnected within the last week and you're still seeing piero then yes you're still on the original setup. Ottawa has been cut over though so i suspect you haven't reconnected in a while?
...restart all your equipment and then see what you get to the first hop. You should expect roughly 8ms if you're on fast path and if you're on interleave then add 20-25ms.
--
Marc |
|
Mark Rejhon
join:2004-02-02
Ottawa, ON
·Magma Communications
| Aha - I just disconnected and reconnected. I'm no longer on piero, and my pings are now slightly lower, down to 55ms. Then yes, I'm on a 1184/384 profile and on interleave.
Are you sure Ottawa users should only get 8ms (fastpath) through 30-40ms (interleave)? I think a lower ping on interleave might only apply to 3Mbps/5Mbps users...
C:\>tracert www.teksavvy.com
Tracing route to tsi.teksavvy.com [206.248.154.253]
over a maximum of 30 hops:
1 1 ms 1 ms 1 ms 192.168.69.1
2 1 ms 1 ms 1 ms speedtouch.lan [10.0.0.138]
3 56 ms 55 ms 58 ms 206.248.154.102
4 64 ms 63 ms 63 ms 206.248.154.66
5 66 ms 66 ms 66 ms www.teksavvy.com [206.248.154.253]
________
Just FYI -- I used to be on Speedtouch Home Ethernet DSL, and had a really bad connection over Xmas (not the modem's fault, but there was a lot of flux on my line). It stabilized sometime in January, then I upgraded my modem to a Speedtouch 516 (v5 - which I obtained refurbished for a good price). My ST516 improved my noise margin to 20db, but my attenuation is still a sad-looking 59.5db. But I have had 0 errors in the last 24 hours according to DMT tool, so that's good. Previously, my STH told me I could attain up to about 3500 Kbps or thereabouts, and it says I was only utilizing 33% ... However, my attenuation is really terrible. Because of this I may later request a profile change to about 1.7 or 2.0 Mbps, and use my DMT tweaking to keep errors within control... But no hurry yet, because of potential line "flux" (variability) due to my Xmas experiences. |
|
