Other Line Options: Difference between revisions
m (→Rate) |
m (→Rate) |
||
Line 15: | Line 15: | ||
[[File:Control-pages-line-options.png|border]] |
[[File:Control-pages-line-options.png|border]] |
||
Example: |
|||
[[File:Cqm-ratedrop.png|none|frame|Example of changing the rate from 100%-95% reduced average latency when filling the link when downloading]] |
|||
==Stability Options== |
==Stability Options== |
Revision as of 22:27, 24 March 2015
Line Rate
This shows the line speed:
Rate
This sets the download (internet to you) speed on the AA routers.
At 100%, the LNS throttles you to exactly your line's capacity. At 110%, it will let through 110% of your line capacity, and allow the wholesaler's systems to buffer/drop the excess. At the 95%, it lets through 95% of the line's capacity, so it will unaffected by the wholesaler's buffering. Setting to less than 100 is advised for lines running VoIP and other realtime services as it means the downlink won't be filled. Lowering the rate from 100% (e.g. to 95%) will mean that there should be no network buffering within the wholesalers network - and may well reduce latency when downloading at the full rate of the line.
Example:
Stability Options
- Available On BT 21CN ADSL and FTTCLines
These stability options refer to how aggressiveness BTs DLM is when monitoring and managing the line.
Usually we'd suggest keeping it on Normal (the default). Using increased stability options increases DLM sensitivity, so can have the side effect of making your line sync at lower speeds.
FastTimeout
Uses a PPP LCP Echo timeout of 10 seconds rather than 60 before entering the LostCarrier state.
More info: LCP echoes usually stop responding if the line has gone down. Our LCP monitoring, which produces the graphs will drop the line if there are no replies after 60 seconds. When bonding or used in a fall-back setup, having a faster timeout is useful in order to fall-back quicker.
MTU
Forces an MTU of regardless of what is negotiated by your router.
More info: MTU is negotiated upon connection. We usually like 1500 MTU, however, there can be situations where by forcing 1492 is required. For example, if your router negotiated 1500 initially but you make a connection to a server that want's to re-negigotiate a lower MTU it will try to di this using ICMP, but if your router has a miss-confuigured firewall that blocks ICMP this can't be re-negoiated and the connection to the server will have problems. This is seen when connecting to certain web sites, bank web sites and https sites. Forcing 1492 from the start means that all your traffic is in packets of 1492 bytes and so won't have this problem.
TCPFix
Modifies the MSS in TCP packets so that it does not exceed the MRU
MRUFix
When sending IP traffic, ignore the MRU we receive during PPP negotiation. This can save a LCP renegotiation (which has been known to help OBSD based PPP's)
LCPFix
Re-negotiates PPP LCP after acquiring the connection from BT (who may provide a wrong high MTU of 1500 on a 1492 link)
More info: During the ppp connection when your router initially syncs up and logs in to us, the ppp connection is passed via BT. Sometimes BT can change the MTU. With this option we will accept the PPP connection from BT, but will then re-negotiate the PPP connection with your router allowing the MTU to be reduced. On TalkTalk connections the MTU is always negotiated as 1492, you need to select LCPFix and an MTU 1500 to fix this (otherwise the LNS will use 1492 when re-negotiating).
Statistics and History
Sync History
Displays a list of your recent resyncs and the resulting sync speed and BRAS rate. This can be useful to view either if your line is regularly resyncing (bad!) or whether there is a deteriorating (or improving) trend in the sync speeds achieved.
Login/Logout
View other lines on your account that logged in or logged out at this same time (not very useful if you only have a single line though!)
Stats
Shows the recent usage / quality stats for the line, with columns Time,Period,Polls Sent,Fail,Latency Min,Ave,Max,Traffic (bit/s) Rx,Tx,Score
The steps are normally 100s. Latency is in ns. The output is in CSV format for ease of import into a spreadsheet.
XML
The same as "Stats" above, but presented in XML to provide other processing options.