[03:08] loaded fie for me in Chrome [04:43] mnathani_: no, i don't think i'll go with telehouse [04:45] normally i'd take that kind of class of data center, but since everything is going to be fully remote (e.g. I don't live in Germany), i need a DC with more on-site staff that can do work at a reasonable price. First-Colo (first-colo.net) looks to fit (toured them the same day as Telehouse); they can provide a somewhat more "semi-managed" solution [04:46] like, if you buy servers from them (they have a store, bunch of supermicro stuff, same models as we use), they include rack / stack / cabling into your cabinet for free and if anything needs to be RMA'd, they handle that too [04:46] and I can still get access to DE-CIX from there [04:52] Telehouse also didn't have enclosed cold-isle containment, while First-Colo did, as well as Interxion (the biggest DC in Germany, and maybe all of EU) [04:52] Demand at Interxion is too high at the moment; lead time for a cabinet is like 6 weeks [04:53] Both Telehouse and Interxion also have very elegant entrances, conference rooms, hallways, etc... etc... all things I figure at the end of the day I would be helping to pay for and not really needing ;) [07:51] up_the_irons: what about IP space, will you be using your existing allocation or will you have to get more space? [08:12] mnathani_: existing [08:37] up_the_irons: glad to hear it's gone well. sorry I can't meet up in frankfurt [08:49] *** [api] has joined #arpnetworks [09:19] Thanks up_the_irons, appreciate the update/info on DC selection. :) [09:52] up_the_irons: I can get ya that in STockholm. [09:52] localhands. [09:54] I had no idea there was another Stockholmer in here. [09:54] hehe :) [09:54] Im in gothenburg.. but I got all my hosting mainly in stockholm.. amsterdam london frankfurt and chicago [09:54] :) [09:55] Soo im already 100% remote [09:55] Heh, makes sense. [09:55] so im dependant on working setup with localhands for onduty if something breaks that I cant fix remote.. most stuff you can fix.. except hw replacement :) [09:56] Where do you host in Sthlm? [09:56] Interxion and Glesys datacenter.. very good one and they are building more rooms all the time. top class and good pricing.. [09:57] easy to get fibers installed to providers and things..and I also have 2x10g ring connected into that dc.. diffrent fiber providers(routes) hehe [09:57] Right on. I've worked with Glesys in the past for a client; they were definitely not the problem. [09:57] im pal with the owners :) [09:59] But I would like to give an offer for arp for sure. [09:59] if interested. [10:01] and I got racks that I have booked that is empty :) [10:01] hehe [10:03] and glesys dc do have 3 ix-switches connected there. one of them is the legendary netnod :) [10:13] lol mike-burns "they were definitely not the problem." [10:13] mike-burns: I go to asia for 6 months starting 28th this month soo Im really dependent on good services.. :) [10:13] (It's a sad state of affairs when that's a significant statement) [10:14] brycec: hah, yes. [10:14] mrsaint: get out while it's still somewhat light out. It's been raining for four days here in Sthlm. [10:14] mike-burns: yeah :) [10:14] Hey here too! [10:14] I go thailand :) [10:14] The sun just came out. [10:15] Heh, everyone in Sweden goes to Thailand. [10:15] mhoran: then I'm not moving to NYC! [10:15] ofcoz.. :) But I go with my thai wife and my dauther. [10:16] mike-burns: im originally from stockholm. But moved here due to consulting contract 5 years ago and never ended it here yet soo :) They keep giving me yearly contracts hehe [10:16] and now they let me go work remote off in thai :) [10:16] Nice deal. [10:16] yes :) [10:17] but I could live off my hosting but its boring.. as its just working so well never have issues and no need to work tomuch.. soo I rather do some nice consulting and make more money :) [10:18] feature of having customers knowing what they do and what they want hehe.. makes life much easier. [10:39] up_the_irons: msg me when awake. read pm aswell :) [12:08] * dne is also stockholmer, actually at walking distance from that glesys dc (haven't used it though) [12:08] I've never been to thailand though :) [12:20] dne: ok.. you should join #Isp on ircnet.. :) One of my channels I created back in 94. Most of the .se operators hang there. [12:23] Damn thats long time ago :) [13:03] hehe [14:43] up_the_irons: will certain prefixes be announced only in Germany? Also will you have a semi-dedicated link between Germany and Los Angeles [14:44] * mercutio doesn't know for sure, but think it's likely that one subnet, or part thereof will be advertised in germany. [14:45] and that a dedicated link between the two is unlikely, but you never know [14:46] and so the same IP addresses aren't likely to be advertised in both locatinos [14:46] unless there's an anycast /24 :) [14:51] mnathani: I would assume it would be completely different IPs in each location, possible different ASNs as well if you're doing it by the book [14:53] I guess then customers' vlans wouldn't exist/persist between locations (for those $few customers with hosts in both locations) [14:57] if it was san jose / los angeles i could see more chance of bridging betwen locations. [15:01] I wonder if an OpenVPN tap bridge could carry vlan tags [15:02] oh i'm sure there are ways to do such things. [15:03] As am I. I'm just curious if that particular method is such a way :P [15:03] yeah i'm just googling heh [15:03] I realize that for ARP it's a question of whether it's worth the effort. [15:03] it's not just that [15:04] it's what do you do about volume accounting, what do you do when it doesn't work in a timely fashion etc [15:04] @tableflip0 [15:04] @tableflip [15:04] (╯‵Д′)╯彡┻━┻ [15:04] ^ that's what I do :p [15:05] how much demand is there etc. [15:05] i think it'd be useful for more people to push for a shared /24 anycast segment [15:05] but even then, if you want to be able to take hosts in/out, then you probably want something like bgp per customer. [15:05] That could be a nice upsell [15:06] for $2/mo get an anycast IP between two locations, something like that [15:06] or you end up with the situation that you are doing something with either of your vm's and your anycast goes down. [15:08] i like anycast :) [15:08] but yeah not everyone can easily get a /24 for anycast. [15:08] That's what she said!! [15:11] Not to mention it can be quite a waste of a /24. If you can share the costs with other users, super [15:12] yeh [15:13] i used to think it'd be cool to have anycast in lots of locations. [15:13] now i'm much more keen on anycast in few reliable locations. [15:13] there are some caveats - like debugging where there are issues if a provider has issues etc. [15:14] a lot of people do anycast badly, in that they'll have two dns servers in the same location, leading to lower availability if there are routing issues to that location. [15:15] where you really want to mix anycast and non anycast, or have different anycast domains that go to different locations. [15:15] well esp. with large cdn's etc. [15:16] 20 or 50 msec doesn't matter as much as 50 or 200. [15:16] and if 20/50 can be more reliable than 20/20 then i'd rather have 30 msec time taken extra on dns lookups etc. [15:17] but even 50/200 isn't that bad. [15:17] even people as large as google with short ttl's aren't running dns servers close to users and you don't really notice [15:17] because all of their site stuff is fast and optimised [15:18] (and because you have really high hit chance rates for google services) [16:12] *** [api] has quit IRC (Quit: Textual IRC Client: www.textualapp.com) [20:19] Do American Nationals need permits and such to setup business in Germany?? [20:19] s/??/? [20:53] don't think that'd be considered a business in germany [21:13] *** carvite_ has joined #arpnetworks [21:16] *** sorresse1n has joined #arpnetworks [21:17] *** carvite has quit IRC (*.net *.split) [21:17] *** sorressean has quit IRC (*.net *.split) [21:17] *** carvite_ is now known as carvite [21:50] *** ben3 has joined #arpnetworks [21:50] *** ChanServ sets mode: +o ben3 [21:52] *** mjp_ has joined #arpnetworks [21:54] *** mercutio has quit IRC (Ping timeout: 250 seconds) [21:54] *** mjp__ has quit IRC (Ping timeout: 250 seconds) [21:54] *** staticsafe has quit IRC (Ping timeout: 250 seconds) [21:54] *** trobotham-cloud has quit IRC (Ping timeout: 250 seconds) [21:55] *** staticsafe has joined #arpnetworks [22:04] *** trobotham-cloud has joined #arpnetworks [22:11] *** ben3 is now known as mercutio [23:06] *** eryc has quit IRC (Ping timeout: 255 seconds) [23:06] *** mike-burns has quit IRC (Ping timeout: 255 seconds) [23:06] *** pjs has quit IRC (Ping timeout: 255 seconds) [23:08] *** eryc has joined #arpnetworks [23:08] *** eryc has quit IRC (Changing host) [23:08] *** eryc has joined #arpnetworks [23:08] *** mike-burns has joined #arpnetworks [23:08] *** ChanServ sets mode: +o mike-burns [23:08] DNSMadeEasy has a network of 600+ Servers for their DNS deployment [23:08] *** pjs has joined #arpnetworks [23:09] *** pjs is now known as Guest80896 [23:11] mnathani_: wow. [23:11] mnathani_: do they anycast with two servers for one location total? [23:11] they're probably sensible. [23:12] damn [23:12] they're not sensible. [23:12] they have 6 dns servers all in the same region over the same provider. [23:12] and better from where you are mnathani_ ? [23:13] oh there's lots of diversity from arp. [23:13] but yeh from new zealand they all terminate in sydney over megaport to the same provider asn. [23:41] they put all ns in all locations I believe [23:41] at least there were diff routes for uk and arp [23:41] instead of all going to the same location in sydney for me [23:44] they have pretty high uptime [23:44] used to be 100% for like the first 8 years of operation [23:46] they had a graph of one of the DDOS attacks targeted at them. Can't seem to find it at the moment though [23:46] can still have localised outages. [23:48] yes, but is an outage really an outage if customers are not affected? [23:48] in the case of redundant anycast provisioned servers [23:48] what [23:48] if a peering exchange has issues [23:48] and all of your routes to a dns server go over that peering exchange [23:48] then it's like all the servers are down if ther'es a peering exchange outage. [23:49] right [23:49] i haven't seen a specific outage on this peering exchange yet [23:49] but i've seen reduced connectivity to it [23:49] which was affecting other stuff to australia too [23:49] there's a backup path, but it's much higher latency etc.