[04:07] mercutio, i suspect thats what it is [04:07] less loss / jitter [04:08] have a clean, straight route to ARP [04:08] I used to care about that stuff a lot more. [04:08] It gets OCD'ish though. [04:08] Now I've widened my tolerance. [04:08] it only niggled me as i wanted to play some old irc buddies on tekken [04:08] grody: do you ssh remotely? [04:08] yea [04:08] like a long distance [04:09] try out my ssh in-band latency test :) [04:09] https://weallsee.net/cl [04:09] hmm [04:09] oh and cloudflare's performance is starting to erk me :) [04:09] or cl.c for source [04:09] that's 64 bit linux binary [04:09] longest SSH is me to here, here to Aus [04:09] thats painful [04:09] do it on a medium one :) [04:09] Yeah, UK from NZ can be painful [04:10] And Australia is slightly further away than NZ from UK [04:10] like being on ssh on 28k [04:10] and im torrenting too [04:10] 28.8k i lower latency. [04:10] is [04:10] do you have a close box to ssh to [04:10] you can do dbclient user@host1,user@host2 [04:10] helps a lot [04:10] so if you have loss between host1 and you it can resend over closer hop [04:10] or loss between host1/host2 [04:10] oh wow [04:10] well worth trying to AU [04:11] it does a tcp port bounce [04:11] dynamic ssh pipes [04:11] so it's secure [04:11] ie you don't have to ssh from host1, and expose your key [04:11] i think ssh -A is insecure (slightly) [04:11] i'm not sure [04:11] i want to figure that out [04:12] i wonder if in 50 years, lag will be a thing of the past [04:12] nope [04:13] speed of light and all that [04:13] especially if quantum entanglement shows up promise [04:13] UK <-> NZ is always going to suck [04:13] but [04:13] there are ways to make lag not so bad. [04:13] aye, but we could send data directly through earth, instead of up, down, up down, across [04:13] which is what i'm interested in in a way [04:13] that's very very complicated [04:13] atm [04:14] i don't think it's likely to happen anytime soon [04:14] there's also microwave to reduce latency too [04:14] which is likely to start happening in parallel more i suspect [04:14] i'd like inner city latency to decrease myself [04:14] err i mean i wnat pings of < 1 msec instead of < 2 msec to neighbours etc. [04:15] UKs inter-links to each other are most of the time pretty good [04:15] i don't think local peering is good [04:15] it's a complicated problem though [04:15] usually connections are terminated far away [04:15] and even on gpon connections latency is often 1 msec [04:15] so 1 msec + 1 msec for neighbour, plus if they're on different isp [04:15] then the link between the two [04:15] within the same city you may get 2.5 msec [04:16] well.. i had a home @ office link on the same FTTC cab, however, they both had to go uptoward london and back before seeing each other [04:16] BT do allow a GEA access, where peering can be done on the same cab [04:16] yeah that's what i mean [04:16] it doesn't necessarily peer locally [04:16] so if my office and home were on the sae cab, they could peer directly via the cab and not over the internet [04:16] people are stil;l used to aggregating circuits remotely etc [04:17] but again latency mitigation makes it matter much less [04:17] mm [04:17] cifs sucks with latency [04:17] like even 20 msec latency can hurt quite bad [04:17] and 20 msec is what 1500 miles? [04:17] maybe 3000 miles [04:17] i dunno [04:17] anyway [04:18] atm if you can keep latency under 20 msec things are pretty good [04:18] but really 60 msec should be able to work good too [04:18] my friend down the road who i lan party with sometimes, we get 19-25 ms to each other [04:19] down the road and 19-25msec? [04:19] but with 16ms auto added from PPP - thats pretty impressive [04:19] see that's not ideal [04:19] yeah it's more like 10 msec here [04:19] 5 msec latency with vdsl [04:19] 8 for me :( [04:19] yeah [04:19] but i usually get 9/10 to 8.8.8.8 etc [04:20] but the US latency hurts more [04:20] i get 4.6 msec to dns [04:20] but google is in diff country [04:20] so it's like 30 msec [04:20] well, you have more people and more space to cover [04:20] 29.3 to 8.8.8.8 [04:20] ouch [04:20] it's not that bad really [04:20] that's 2155 km [04:20] or someting [04:21] this is why i hate 3G - latency is almost always 60ms+ [04:21] (auckland to sydney) [04:21] HSPA/4G/LTE, i've seen them as low as 15 mind [04:21] 8.8.8.8 in sydney gives terrible performance though [04:21] DNS lookups on latency really kills a browsing experience [04:21] yeah 3g latency sucks [04:21] yeh i used proxy when i did it regularly [04:21] and did initcwnd [04:21] with the 10 packets [04:21] helped a lot [04:21] but 3g sucks for other reasons too [04:22] liek if you're downloading your latency will go sky high [04:22] and the connection becomes useless for anything else [04:22] hah yea [04:22] but as long as you don't download it's ok [04:23] think i use an L2TP from my HSPA access point [04:48] http://imgur.com/fV972aK thats a mixed graph for an example - first chunk is good signal, but transferring (downlading arch via torrent), middle is idle whilst walking around town and latter is idle with strong signal [04:50] been using my ISPs wonderful graphing to prove to my mobile carrier their data network is shot to sh1t [04:50] (those graphs are excellent examples of how it /should/ be) [04:51] they probably don't care :) [04:51] i dunno why arch encourages torrents [04:52] i reckon http with close mirrors is good [04:53] lol they don't - but i've had 3 months free service and had them taken to their ombudsman, got another month free (by their request) and £35 compo [04:53] i like torrents *shrug* [04:53] wow [04:53] least if the download breaks, it automatically resumes when network is back [04:53] that so wouldn't happen here [04:54] same with wget -c [04:54] :P [04:54] UK is very strict on how companies must treat their customers and have to live up to standards [04:54] fail, they have to pay [04:55] i was honestly hoping for a recinded contract - but will wait for the free months to end and drop the priceplan to bare minimum, they buy out [04:55] go back to a better network [07:08] *** nesta_ has joined #arpnetworks [07:10] *** neish_ has joined #arpnetworks [07:12] *** KDE_Perr1 has joined #arpnetworks [07:13] *** nesta has quit IRC (Ping timeout: 240 seconds) [07:13] *** jbergstroem has quit IRC (Ping timeout: 240 seconds) [07:13] *** KDE_Perry has quit IRC (Ping timeout: 240 seconds) [07:13] *** neish has quit IRC (Ping timeout: 240 seconds) [07:13] *** nesta_ is now known as nesta [07:16] *** jbergstroem has joined #arpnetworks [08:27] *** awyeah has quit IRC (Quit: ZNC - http://znc.in) [08:29] *** awyeah has joined #arpnetworks [13:12] awwyeah [15:45] *** dj_goku_ has quit IRC (Remote host closed the connection) [15:52] *** dj_goku has joined #arpnetworks [15:52] *** dj_goku has quit IRC (Changing host) [15:52] *** dj_goku has joined #arpnetworks [16:42] *** jbum has joined #arpnetworks [18:15] http://arstechnica.com/tech-policy/2015/08/jeb-bush-crypto-makes-it-harder-for-the-american-government-to-do-its-job/ [18:15] Ars Technica: "Jeb Bush wants “a new arrangement with Silicon Valley” to ease crypto" [18:16] I really can't see that working out well :) [18:18] *** jbum has quit IRC (Quit: jbum)