More stats and info
Shadow Hawkins on Thursday, 12 January 2012 00:36:03
I would like to know more stats and more technical info about the pops and their tunnels. Not because i need it, but because its cool..
Examples could be architecture info and hardware-info about the pops.
"Top" for a range of parameters for pops and their tunnels.
More info about your core network, with regards to the BGP-network an the AS structure. As well as being interesting to geeks, such a comprehensive network, offering openness would without a doubt have educational applications.
More stats and info
Jeroen Massar on Thursday, 12 January 2012 13:27:01 Examples could be architecture info and hardware-info about the pops.
See the PoP requirements page.
"Top" for a range of parameters for pops and their tunnels.
See the PoP overview page
More info about your core network, with regards to the BGP-network an the AS structure
SixXS does not have a 'core' network, as the networks are run by the various ISPS, please see their websites if they have any public documentation. For a lot of companies this is confidential information though.
For BGP, one could always peek at RIPE RIS, RIPE ATLAS or PeeringDB
offering openness would without a doubt have educational applications.
I think that all possible statistics possible which are not confidential in nature are on the website already.
And yes, we are aware that traffic monitoring needs to be fixed. But some things need to be done first.
More stats and info
Shadow Hawkins on Friday, 27 March 2015 06:19:38
Jeroen Massar wrote:
I think that all possible statistics possible which are not confidential in nature are on the website already.
And yes, we are aware that traffic monitoring needs to be fixed. But some things need to be done first.
I don't know whether replying to old posts is considered bad from ovet here, but since yoy mentioned statistics:
Tunnel statistics could be changed not to crop input side of graph in cases where input/output are close to each other. If I recall correctly, earlier it had some bandwidth scaling selectbox, but I guess that was removed in favor of automatic scaling.
Seeing uncropped version of both input & output graphs besides latency graph would help understand various AQM settings better.
More stats and info
Jeroen Massar on Friday, 27 March 2015 09:52:29 Tunnel statistics could be changed not to crop input side of graph in cases where input/output are close to each other. If I recall correctly, earlier it had some bandwidth scaling selectbox, but I guess that was removed in favor of automatic scaling.
The range selection was causing unnecessary burden on our backends, we thus removed it.
Seeing uncropped version of both input & output graphs besides latency graph would help understand various AQM settings better.
Which AQM settings?
Remind yourself that a tunnel spans over the IPv4 network. Hence a lot of hops are involved that could do a lot of things to the tunneled packets.
You won't be able to make any conclusions from RRDs that get compressed/bucketed over time.
Posting is only allowed when you are logged in. |