[01:26] *** pyvpx has joined #arpnetworks [01:29] *** zhangxiaobao has joined #arpnetworks [01:29] do we have a looking glass now? [01:33] ``yeah [01:39] so where is it? [01:39] *** zhangxiaobao has quit IRC (Remote host closed the connection) [02:44] *** LT has joined #arpnetworks [03:41] *** toddf has quit IRC (Ping timeout: 260 seconds) [03:41] *** toddf has joined #arpnetworks [03:41] *** ChanServ sets mode: +o toddf [04:04] *** zhangxiaobao has joined #arpnetworks [04:05] I'm sorry I had to leave to attend a meeting [04:06] is it public? the looking glass [04:07] zhangxiaobao: http://lg.arpnetworks.com/ [04:08] thank you [05:49] *** zhangxiaobao has quit IRC (Remote host closed the connection) [08:56] *** LT has quit IRC (Quit: Leaving) [11:02] *** mkb has joined #arpnetworks [17:03] this whole 95th percentile method for bandwidth billing really has me confused [17:03] apparently not all ISPs calculate it the same way? [17:05] ie: max-of-total(in,out) vs in+out vs max-of-each-in/out-reading [17:06] Doesn't surprise me. [17:21] * brycec wonders who Spirit Communications si/are that ARP is directly peering with them [17:42] *** reardencode has joined #arpnetworks [17:47] just another network that requested peering... about 2000 prefixes, so not too shaby [17:48] up_the_irons: was there ever a mention of appearing at AMS-IX via coresite and ARP Networks peering even further [17:55] did the /30 come from ARP space or did they provide it? [17:57] what /30? [17:58] things are different when a route server involved [17:58] ? [17:58] i only vaguely remember something about coresite at AMS-IX [17:58] mnathani: we're actually on a shared /23 [17:58] 1 IP for each member [17:58] at any2 [17:59] ya know, cuz it's many-to-many, not one-to-one (single link /30) [17:59] could you get paid transit through that setup or is it strictly peering [18:02] strictly peering [18:02] CoreSite has a new-ish service where it is similar to Any2, but *is* designed for paid transit [18:04] does insufficient bandwidth ever become an issue when you use their fabric to peer with multiple networks, assuming single gigE connection [18:09] no, never [18:09] we don't have enough traffic overall for it to matter [18:09] * brycec sets about to change that :D [18:10] I am currently reading: The 2014 Internet Peering Playbook: Connecting to the Core of the Internet [18:11] Not sure if an actual core exists any more though [18:12] brycec: lol [18:12] stop plugging the tubes [18:12] (Not really though. I did that once before locally, clogging up mirrors.arpnetworks.com with about 1.1TB of bandwidth) [18:13] (but I wouldn't intentionally degrade service for other customers) [18:13] oh yeah i remember that ;) [18:14] here is the ams-ix / coresite article: https://ams-ix.net/newsitems/2 [18:15] up_the_irons: Fun times :D http://i.imgur.com/A6YyTWy.png [18:16] mnathani: oh that's pretty cool [18:16] brycec: \m/ [18:18] up_the_irons: here is a more recent one with a Chicago DC http://www.datacenterknowledge.com/archives/2014/10/07/ams-ix-adds-internet-exchange-at-cmes-chicago-data-center/ [18:20] i'll write my sales rep about that, see what is involved in getting hooked up to ams-ix [18:35] cool [19:16] oh, nice. [19:17] about ams-ix that is. [19:17] That's what she said!! [19:17] up_the_irons: did coresite ever get back about that outage? [19:18] BryceBot: no [19:18] Oh, okay... I'm sorry. 'about ams-ix that is.' [20:09] mnathani: peering is a hugely complex issue [20:10] some locations tend to be better than others [20:11] so liek in the US, smewhere like say tampa tends to not have good peering nearby [20:11] and the edge locations tend to be overloaded more than more inner land ones. [20:12] like miami connects to south america, san jose to asia etc. [20:12] and in betwen places liek texas seem so have the best peering in general [20:13] also california seems to get a lot of cable cuts.