From holger at freyther.de Sun Oct 14 19:39:52 2012 From: holger at freyther.de (Holger Hans Peter Freyther) Date: Sun, 14 Oct 2012 21:39:52 +0200 Subject: Kick-off for the 29C3 Message-ID: <20121014193952.GA20812@xiaoyu.lan> Hi, another year, another congress, another OpenBSC installation. There are some changes though. The most noticable one is that we are at a new venue and this has various implications. * A lot more space to cover and challenges to radio planning. How many BTS do we need? Would we use a dual TRX setup? Only in some rooms? Can we cheat and use an amplifier? What kind of patch-work equipment do we get? * I wanto to have increased stability at this congress. We are mostly using the nitb for the congress setup and there are various bugs that I would like to address before. I believe it is time we offer GSM as a reliable service. This does not mean there is no room for experiments but they need to be planned and scheduled. E.g. experiments with tsunami warnings, GPRS at night.. etc. * Some increased logistic issues. We really need to plan what we need and how we get this transported to Hamburg and the CCH. All the small things like boxes of patch cables, routers, screws, locks, belts... I have started this wiki page[1] to track our requirements. The basic idea is we have a task and someone that is responsible for executing it or if he starts to drop it.. officially return the task. cheers holger [1] http://openbsc.osmocom.org/trac/wiki/FieldTests/29c3 From holger at freyther.de Sun Oct 14 19:50:10 2012 From: holger at freyther.de (Holger Hans Peter Freyther) Date: Sun, 14 Oct 2012 21:50:10 +0200 Subject: 29C3 osmo-nitb sprint and issues Message-ID: <20121014195010.GB20812@xiaoyu.lan> Hi, as indicated in the previous mail there are some things to be done before we get to the venue. I would like to experiment with sprints to solve some of the stability issues we had with osmo-nitb. The idea is that we assemble (physically or online) and work on some specific issues of osmo-nitb that impact the quality of service for the upcoming event. I can provide food and snacks during these sprints and guidance on OpenBSC code. The first sprint is scheduled for the 9th, 10th (and maybe 11th) of November. There is no venue for it yet, maybe the CCCB. Please indicate if you are interested to join. The general list of osmo-nitb issues are: * Channel release process. There is a branch I created at the last congress but I have never tested it. This includes making a SAPI=3 local release, using the proper timeouts for the abnormal release making sure the SACCH is deactivated and some more issues. This requires quite a bit of testing. * Allow to block/unblock failing lchans. We have had this issue with Nokia BTS but it can happen on other BTS as well. E.g. a given timeslot might not be allocatable, or we get LCHAN ACT NACK. In that case OpenBSC should block the lchan. * smsqueue/auth/paths. In the past years we have seen that the smsqueue and other parts can get stuck because of the 'subscr_get_con' not being properly called in all error conditions. We need to review the code and check how things get stuck. * Handover testing and patching out the 'timestamp' adjusting or fixing it. Right now it can be used to add a lot of delay by switching cells. cheers holger From peter at stuge.se Sun Oct 14 19:59:33 2012 From: peter at stuge.se (Peter Stuge) Date: Sun, 14 Oct 2012 21:59:33 +0200 Subject: 29C3 osmo-nitb sprint and issues In-Reply-To: <20121014195010.GB20812@xiaoyu.lan> References: <20121014195010.GB20812@xiaoyu.lan> Message-ID: <20121014195933.9439.qmail@stuge.se> Holger Hans Peter Freyther wrote: > as indicated in the previous mail there are some things to be done > before we get to the venue. I would like to experiment with sprints > to solve some of the stability issues we had with osmo-nitb. Please post this on the openbsc@ list to reach the right people? I think this orga list is more for practical stuff like what's on the wiki page. Thanks for getting this year's wiki page started! My photos are btw. available at http://stuge.se/photos/CCH_29C3/ //Peter From holger at freyther.de Mon Oct 15 07:41:54 2012 From: holger at freyther.de (Holger Hans Peter Freyther) Date: Mon, 15 Oct 2012 09:41:54 +0200 Subject: 29C3 osmo-nitb sprint and issues In-Reply-To: <20121014195933.9439.qmail@stuge.se> References: <20121014195010.GB20812@xiaoyu.lan> <20121014195933.9439.qmail@stuge.se> Message-ID: <20121015074154.GE20812@xiaoyu.lan> On Sun, Oct 14, 2012 at 09:59:33PM +0200, Peter Stuge wrote: > Holger Hans Peter Freyther wrote: > > as indicated in the previous mail there are some things to be done > > before we get to the venue. I would like to experiment with sprints > > to solve some of the stability issues we had with osmo-nitb. > > Please post this on the openbsc@ list to reach the right people? I > think this orga list is more for practical stuff like what's on the > wiki page. Thanks for getting this year's wiki page started! good point From 246tnt at gmail.com Mon Oct 15 14:45:51 2012 From: 246tnt at gmail.com (Sylvain Munaut) Date: Mon, 15 Oct 2012 16:45:51 +0200 Subject: OpenBSC / OpenBTS integration Message-ID: Hi, I was just wondering what kind of inter-operation was going to be used ? If both system are to be deployed and be seamless for the user, I guess they're going to broadcast the same MCC/MNC as we always do. Now two cases are possibles: * We use distinct LAC: This means we need to somehow keep track on what phone is where and have someway to route calls / SMS appropriately * We use the same LAC: We just send each requests (call/sms) to both systems, they each initiate paging and whoever respond to the SIP call is the winner I guess. In any case there is probably some work to be done there as well since both system will have to operate from the same HLR (things like the cipher key sequence and last auth tupe would need to be in sync all the time). For voice call both system can talk SIP so although the multi destination call would need to be tested, that "should" work. For SMSes however there is no external layer in osmo-nitb AFAIK ... Cheers, Sylvain From holger at freyther.de Mon Oct 15 17:22:39 2012 From: holger at freyther.de (Holger Hans Peter Freyther) Date: Mon, 15 Oct 2012 19:22:39 +0200 Subject: OpenBSC / OpenBTS integration In-Reply-To: References: Message-ID: <20121015172239.GF17104@xiaoyu.lan> On Mon, Oct 15, 2012 at 04:45:51PM +0200, Sylvain Munaut wrote: > Hi, Hi, > > I was just wondering what kind of inter-operation was going to be used ? to be honest I think we have enough work with stability, new venue, not being located in Berlin, new team already. What I can imagine is that during certain hours we switch off the OpenBSC network and let OpenBTS handle it. We would only need to share the HLR/VLR state and then copy undelivered SMS. holger From alexander.chemeris at gmail.com Mon Oct 15 18:55:45 2012 From: alexander.chemeris at gmail.com (Alexander Chemeris) Date: Mon, 15 Oct 2012 22:55:45 +0400 Subject: OpenBSC / OpenBTS integration In-Reply-To: <20121015172239.GF17104@xiaoyu.lan> References: <20121015172239.GF17104@xiaoyu.lan> Message-ID: Hi Holger, On Mon, Oct 15, 2012 at 9:22 PM, Holger Hans Peter Freyther wrote: > On Mon, Oct 15, 2012 at 04:45:51PM +0200, Sylvain Munaut wrote: >> I was just wondering what kind of inter-operation was going to be used ? > > to be honest I think we have enough work with stability, new venue, > not being located in Berlin, new team already. These all are valid concerns. We would love to test OpenBTS/OpenBSC integration, but this is a second goal for us (Fairwaves/OpenBTS team). I.e. we'll plan to do that, but only if we have enough time/resources. > What I can imagine is that during certain hours we switch off the > OpenBSC network and let OpenBTS handle it. We would only need to share > the HLR/VLR state and then copy undelivered SMS. Sounds good. Do you plan to cover the whole area? To switch smoothly we have to have a similar coverage. May be even share installation locations. If we achieve OpenBTS/OpenBSC interconnection, we could run a convergent network by using only some of the installed hardware. -- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ??? ??????? http://fairwaves.ru From holger at freyther.de Tue Oct 16 08:53:05 2012 From: holger at freyther.de (Holger Hans Peter Freyther) Date: Tue, 16 Oct 2012 10:53:05 +0200 Subject: OpenBSC / OpenBTS integration In-Reply-To: References: <20121015172239.GF17104@xiaoyu.lan> Message-ID: <20121016085305.GA17154@xiaoyu.lan> On Mon, Oct 15, 2012 at 10:55:45PM +0400, Alexander Chemeris wrote: > These all are valid concerns. > We would love to test OpenBTS/OpenBSC integration, but this is a > second goal for us (Fairwaves/OpenBTS team). I.e. we'll plan to do > that, but only if we have enough time/resources. well, the easiest from a GSM point of view would be to have OpenBTS speak OML/RSL. But from the questions tnt asked it appears to be a more high level integration (non-GSM nitb<->OpenBTS interop). > > Sounds good. Do you plan to cover the whole area? To switch smoothly > we have to have a similar coverage. May be even share installation > locations. I think we would experiment in the early morning or late night and having a reduced coverage (hackcenter, foyer) would probably be acceptable. cheers holger From alexander.chemeris at gmail.com Tue Oct 16 19:19:09 2012 From: alexander.chemeris at gmail.com (Alexander Chemeris) Date: Tue, 16 Oct 2012 23:19:09 +0400 Subject: OpenBSC / OpenBTS integration In-Reply-To: <20121016085305.GA17154@xiaoyu.lan> References: <20121015172239.GF17104@xiaoyu.lan> <20121016085305.GA17154@xiaoyu.lan> Message-ID: Hi Holger, On Tue, Oct 16, 2012 at 12:53 PM, Holger Hans Peter Freyther wrote: > On Mon, Oct 15, 2012 at 10:55:45PM +0400, Alexander Chemeris wrote: >> These all are valid concerns. >> We would love to test OpenBTS/OpenBSC integration, but this is a >> second goal for us (Fairwaves/OpenBTS team). I.e. we'll plan to do >> that, but only if we have enough time/resources. > > well, the easiest from a GSM point of view would be to have OpenBTS > speak OML/RSL. But from the questions tnt asked it appears to be a > more high level integration (non-GSM nitb<->OpenBTS interop). Our target is to test OpenBTS and it's network model with SIP core. So interconnection could be only through SIP. IIRC, we discussed using LCR with SIP interface previously, but I have to recall where those discussions led us. What kind of interconnection do you have with the POC? >> Sounds good. Do you plan to cover the whole area? To switch smoothly >> we have to have a similar coverage. May be even share installation >> locations. > > I think we would experiment in the early morning or late night and > having a reduced coverage (hackcenter, foyer) would probably be > acceptable. Do you have statistics from previous CCC about the usage patterns? We would be happy to use mornings/evenings if they provide enough load to expose bugs/issues. If the load would be close to zero, then the testing will be useless for us. -- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ??? ??????? http://fairwaves.ru From alexander.chemeris at gmail.com Tue Oct 16 19:24:25 2012 From: alexander.chemeris at gmail.com (Alexander Chemeris) Date: Tue, 16 Oct 2012 23:24:25 +0400 Subject: OpenBSC / OpenBTS integration In-Reply-To: References: Message-ID: Hi Sylvain, On Mon, Oct 15, 2012 at 6:45 PM, Sylvain Munaut <246tnt at gmail.com> wrote: > If both system are to be deployed and be seamless for the user, I > guess they're going to broadcast the same MCC/MNC as we always do. > > Now two cases are possibles: > > * We use distinct LAC: This means we need to somehow keep track on > what phone is where and have someway to route calls / SMS > appropriately > * We use the same LAC: We just send each requests (call/sms) to both > systems, they each initiate paging and whoever respond to the SIP call > is the winner I guess. OpenBTS has to have different LAC for each BTS, so only the first way is possible. OpenBTS needs LURs to update registration tables for its softswitch and smqueue - otherwise it doesn't know where to send paging and we don't support paging on multiple BTS'es at the moment. > In any case there is probably some work to be done there as well since > both system will have to operate from the same HLR (things like the > cipher key sequence and last auth tupe would need to be in sync all > the time). Well, we could just have two copies of HLR in different formats. I don't think it's changed frequently at CCC? > For SMSes however there is no external layer in osmo-nitb AFAIK ... Could we write to the OpenBSC SMS DB to send transfer messages from OpenBTS to OpenBSC? -- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ??? ??????? http://fairwaves.ru From holger at freyther.de Tue Oct 16 20:06:55 2012 From: holger at freyther.de (Holger Hans Peter Freyther) Date: Tue, 16 Oct 2012 22:06:55 +0200 Subject: OpenBSC / OpenBTS integration In-Reply-To: References: <20121015172239.GF17104@xiaoyu.lan> <20121016085305.GA17154@xiaoyu.lan> Message-ID: <20121016200655.GA25633@xiaoyu.lan> On Tue, Oct 16, 2012 at 11:19:09PM +0400, Alexander Chemeris wrote: > Hi Holger, Hi Alexander, > > What kind of interconnection do you have with the POC? E1, or last time: E1 over IP. I wouldn't want to switch from E1 to SIP at this point in time for osmo-nitb/LCR. So in your setup this would probably be OpenBTS -> LCR/SIP -> LCR/E1.. > Do you have statistics from previous CCC about the usage patterns? We > would be happy to use mornings/evenings if they provide enough load to > expose bugs/issues. If the load would be close to zero, then the > testing will be useless for us. Harald had some but I never copied the data. But there are two things. 1.) Whenever you switch on/off BTS you should see LUs from both Congress SIM users and international subscribers. 2.) Amount of voice calls was low IIRC. holger From alexander.chemeris at gmail.com Tue Oct 16 20:23:59 2012 From: alexander.chemeris at gmail.com (Alexander Chemeris) Date: Wed, 17 Oct 2012 00:23:59 +0400 Subject: OpenBSC / OpenBTS integration In-Reply-To: <20121016200655.GA25633@xiaoyu.lan> References: <20121015172239.GF17104@xiaoyu.lan> <20121016085305.GA17154@xiaoyu.lan> <20121016200655.GA25633@xiaoyu.lan> Message-ID: Holger, On Wed, Oct 17, 2012 at 12:06 AM, Holger Hans Peter Freyther wrote: > On Tue, Oct 16, 2012 at 11:19:09PM +0400, Alexander Chemeris wrote: >> What kind of interconnection do you have with the POC? > > E1, or last time: E1 over IP. I wouldn't want to switch from E1 > to SIP at this point in time for osmo-nitb/LCR. So in your setup > this would probably be OpenBTS -> LCR/SIP -> LCR/E1.. Ok. We'll look into this. Do you know how we could test this interconnection before CCC? Could we interconnect with OpenBSC in the same way? I have little knowledge of OpenBSC, so I'll let Dmitri Soloviev to handle this part, but I want to know your opinion too. >> Do you have statistics from previous CCC about the usage patterns? We >> would be happy to use mornings/evenings if they provide enough load to >> expose bugs/issues. If the load would be close to zero, then the >> testing will be useless for us. > > Harald had some but I never copied the data. But there are two things. > > 1.) Whenever you switch on/off BTS you should see LUs from both > Congress SIM users and international subscribers. I don't think LUs will be a problem. Calls and SMS are much more interesting to test. > 2.) Amount of voice calls was low IIRC. And that's what worries me - we would love to test the system under more voice/SMS load. One idea which came to my mind recently is to implement some fun (or useful) value added services on our network which would encourage people to make calls or send messages. I hope we'll have integration with SwitchCoder.com soon and we could use it for VAS development. Or may be even let participants to develop services using it. This is a bare idea right now, thus please comment, criticize and improve. -- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ??? ??????? http://fairwaves.ru From 246tnt at gmail.com Wed Oct 17 11:37:51 2012 From: 246tnt at gmail.com (Sylvain Munaut) Date: Wed, 17 Oct 2012 13:37:51 +0200 Subject: OpenBSC / OpenBTS integration In-Reply-To: References: <20121015172239.GF17104@xiaoyu.lan> <20121016085305.GA17154@xiaoyu.lan> <20121016200655.GA25633@xiaoyu.lan> Message-ID: Hi, > Ok. We'll look into this. > Do you know how we could test this interconnection before CCC? You can try just setup LCR / OpenBSC combo locally. and If you can talk to LCR via SIP, the E1 POC uplink should just work. > I don't think LUs will be a problem. > Calls and SMS are much more interesting to test. There isn't that much call at all, even during the day AFAIR. SMS is more used, especially since some people do scripts to send hundreds of them :p > One idea which came to my mind recently is to implement some fun (or > useful) value added services on our network which would encourage > people to make calls or send messages. I hope we'll have integration > with SwitchCoder.com soon and we could use it for VAS development. Or > may be even let participants to develop services using it. This is a > bare idea right now, thus please comment, criticize and improve. The POC IIRC has some voice services like alarm / next talk announce / live audio from the various rooms. Cheers, Sylvain From alexander.chemeris at gmail.com Wed Oct 17 17:46:08 2012 From: alexander.chemeris at gmail.com (Alexander Chemeris) Date: Wed, 17 Oct 2012 21:46:08 +0400 Subject: OpenBSC / OpenBTS integration In-Reply-To: References: <20121015172239.GF17104@xiaoyu.lan> <20121016085305.GA17154@xiaoyu.lan> <20121016200655.GA25633@xiaoyu.lan> Message-ID: Hi, On Wed, Oct 17, 2012 at 3:37 PM, Sylvain Munaut <246tnt at gmail.com> wrote: >> Ok. We'll look into this. >> Do you know how we could test this interconnection before CCC? > > You can try just setup LCR / OpenBSC combo locally. and If you can > talk to LCR via SIP, the E1 POC uplink should just work. Yes, we'll try to put together a test setup in our hackerspace this Sat. When we get it fully working, could we somehow test interconnection with POC before the event? E.g. over Internet. >> One idea which came to my mind recently is to implement some fun (or >> useful) value added services on our network which would encourage >> people to make calls or send messages. I hope we'll have integration >> with SwitchCoder.com soon and we could use it for VAS development. Or >> may be even let participants to develop services using it. This is a >> bare idea right now, thus please comment, criticize and improve. > > The POC IIRC has some voice services like alarm / next talk announce / > live audio from the various rooms. Yeah, I recall those too. I hope we could do something interesting/fun too. -- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ??? ??????? http://fairwaves.ru From holger at freyther.de Wed Oct 17 18:29:06 2012 From: holger at freyther.de (Holger Hans Peter Freyther) Date: Wed, 17 Oct 2012 20:29:06 +0200 Subject: OpenBSC / OpenBTS integration In-Reply-To: References: <20121015172239.GF17104@xiaoyu.lan> <20121016085305.GA17154@xiaoyu.lan> <20121016200655.GA25633@xiaoyu.lan> Message-ID: <20121017182906.GA6595@xiaoyu.lan> On Wed, Oct 17, 2012 at 09:46:08PM +0400, Alexander Chemeris wrote: > Yes, we'll try to put together a test setup in our hackerspace this > Sat. When we get it fully working, could we somehow test > interconnection with POC before the event? E.g. over Internet. not that I know off. But if your LCR connectivity works it will mostly work. Peter has the copy of the last LCR installation. The other bit we run (and it should be in the wiki) is a cron job to download the new extensions for subscribers that have registered to the network. From alexander.chemeris at gmail.com Wed Oct 17 18:42:47 2012 From: alexander.chemeris at gmail.com (Alexander Chemeris) Date: Wed, 17 Oct 2012 22:42:47 +0400 Subject: OpenBSC / OpenBTS integration In-Reply-To: References: Message-ID: Hi all, On Mon, Oct 15, 2012 at 6:45 PM, Sylvain Munaut <246tnt at gmail.com> wrote: > If both system are to be deployed and be seamless for the user, I > guess they're going to broadcast the same MCC/MNC as we always do. Btw, what if we run two different MNCs - would there be any difference from running "same MNC, separate LAC" configuration? We've been discussing this with Ivan here and it seems from the OpenBTS point of view there is no difference. The benefit of having two MNCs is a better user experience - they will understand on which network they're camped and what features/bugs to expect. OTOH i think it may increase the paperwork and payment to authorities, required to get the permission. Holger, you said most likely it will be your responsibility - what do you expect from the paperwork if we request two MNCs? -- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ??? ??????? http://fairwaves.ru From 246tnt at gmail.com Wed Oct 17 18:57:32 2012 From: 246tnt at gmail.com (Sylvain Munaut) Date: Wed, 17 Oct 2012 20:57:32 +0200 Subject: OpenBSC / OpenBTS integration In-Reply-To: References: Message-ID: Hi, >> If both system are to be deployed and be seamless for the user, I >> guess they're going to broadcast the same MCC/MNC as we always do. > > Btw, what if we run two different MNCs - would there be any difference > from running "same MNC, separate LAC" configuration? We've been > discussing this with Ivan here and it seems from the OpenBTS point of > view there is no difference. The SIM distributed (including ones at past events) have a specific MNC/MCC, they'll preferrably camp to that specific MNC/MCC. Cheers, Sylvain From alexander.chemeris at gmail.com Wed Oct 17 19:06:22 2012 From: alexander.chemeris at gmail.com (Alexander Chemeris) Date: Wed, 17 Oct 2012 23:06:22 +0400 Subject: OpenBSC / OpenBTS integration In-Reply-To: References: Message-ID: 17.10.2012 22:57 ???????????? "Sylvain Munaut" <246tnt at gmail.com> ???????: > > Hi, > > >> If both system are to be deployed and be seamless for the user, I > >> guess they're going to broadcast the same MCC/MNC as we always do. > > > > Btw, what if we run two different MNCs - would there be any difference > > from running "same MNC, separate LAC" configuration? We've been > > discussing this with Ivan here and it seems from the OpenBTS point of > > view there is no difference. > > The SIM distributed (including ones at past events) have a specific > MNC/MCC, they'll preferrably camp to that specific MNC/MCC. > Yep. But if the preferred MCC/MNC is not available, they will roam into a network which allows them to roam. And the only such network is the OpenBTS network. Here I'm assuming that OpenBSC and OpenBTS networks have different coverage. Sent from my Android device. -- Regards, Alexander Chemeris CEO, Fairwaves LLC http://fairwaves.ru -------------- next part -------------- An HTML attachment was scrubbed... URL: From Alexander.Chemeris at fairwaves.ru Thu Oct 25 19:21:30 2012 From: Alexander.Chemeris at fairwaves.ru (Alexander Chemeris) Date: Thu, 25 Oct 2012 23:21:30 +0400 Subject: LAN access at the venue In-Reply-To: References: Message-ID: Hi all, What access will we have to the LAN at the venue? 1. We'll we have a separate network for GSM equipment? 2. Will it be 1GbE or 100Mbit network? -- Regards, Alexander Chemeris CEO, Fairwaves LLC http://fairwaves.ru From peter at stuge.se Thu Oct 25 19:24:04 2012 From: peter at stuge.se (Peter Stuge) Date: Thu, 25 Oct 2012 21:24:04 +0200 Subject: LAN access at the venue In-Reply-To: References: Message-ID: <20121025192404.2249.qmail@stuge.se> Alexander Chemeris wrote: > What access will we have to the LAN at the venue? > 1. We'll we have a separate network for GSM equipment? We'll get a VLAN. > 2. Will it be 1GbE or 100Mbit network? There's not much copper in the building, mostly fiber, so expect GbE. //Peter From alexander.chemeris at gmail.com Thu Oct 25 19:49:27 2012 From: alexander.chemeris at gmail.com (Alexander Chemeris) Date: Thu, 25 Oct 2012 23:49:27 +0400 Subject: LAN access at the venue In-Reply-To: <20121025192404.2249.qmail@stuge.se> References: <20121025192404.2249.qmail@stuge.se> Message-ID: On Thu, Oct 25, 2012 at 11:24 PM, Peter Stuge wrote: > Alexander Chemeris wrote: >> What access will we have to the LAN at the venue? >> 1. We'll we have a separate network for GSM equipment? > > We'll get a VLAN. Ok, >> 2. Will it be 1GbE or 100Mbit network? > > There's not much copper in the building, mostly fiber, so expect GbE. Hope so! -- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ??? ??????? http://fairwaves.ru From peter at stuge.se Thu Oct 25 20:59:06 2012 From: peter at stuge.se (Peter Stuge) Date: Thu, 25 Oct 2012 22:59:06 +0200 Subject: LAN access at the venue In-Reply-To: References: <20121025192404.2249.qmail@stuge.se> Message-ID: <20121025205906.9702.qmail@stuge.se> Alexander Chemeris wrote: > >> 2. Will it be 1GbE or 100Mbit network? > > > > There's not much copper in the building, mostly fiber, so expect GbE. > > Hope so! Are there special concerns for your deployment? If yes, please tell, so that I can bring them up with network guys. Anything is generally possible as long as it is planned in advance, that is different from the Berlin venue where anything could also be set up ad-hoc. The new venue requires more planning ahead. //Peter From alexander.chemeris at gmail.com Fri Oct 26 13:02:33 2012 From: alexander.chemeris at gmail.com (Alexander Chemeris) Date: Fri, 26 Oct 2012 17:02:33 +0400 Subject: LAN access at the venue In-Reply-To: <20121025205906.9702.qmail@stuge.se> References: <20121025192404.2249.qmail@stuge.se> <20121025205906.9702.qmail@stuge.se> Message-ID: On Fri, Oct 26, 2012 at 12:59 AM, Peter Stuge wrote: > Alexander Chemeris wrote: > > >> 2. Will it be 1GbE or 100Mbit network? > > > > > > There's not much copper in the building, mostly fiber, so expect GbE. > > > > Hope so! > > Are there special concerns for your deployment? If yes, please tell, > so that I can bring them up with network guys. Anything is generally > possible as long as it is planned in advance, that is different from > the Berlin venue where anything could also be set up ad-hoc. The new > venue requires more planning ahead. I'm thinking whether we could operate OpenBTS in "Remote Radio Head" mode, where base CPU units are located in a server room and transceivers are streaming I/Q data to them over LAN. In this case we need a guaranteed bandwidth to avoid packet loss - from 9Mbit for single-channel TRX with 1SPS to ~72Mbit for dual-channel TRX with 4SPS. I have never used VLAN, but from what I understand it means that we share the bandwidth with other network users and thus we hardly could expect to occupy that much bandwidth. Is that correct? -- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ??? ??????? http://fairwaves.ru From 246tnt at gmail.com Fri Oct 26 13:09:32 2012 From: 246tnt at gmail.com (Sylvain Munaut) Date: Fri, 26 Oct 2012 15:09:32 +0200 Subject: LAN access at the venue In-Reply-To: References: <20121025192404.2249.qmail@stuge.se> <20121025205906.9702.qmail@stuge.se> Message-ID: > I'm thinking whether we could operate OpenBTS in "Remote Radio Head" > mode, where base CPU units are located in a server room and > transceivers are streaming I/Q data to them over LAN. In this case we > need a guaranteed bandwidth to avoid packet loss - from 9Mbit for > single-channel TRX with 1SPS to ~72Mbit for dual-channel TRX with > 4SPS. You'd also need copper gigabit, so if there are nanoBTS and UmTRX at the same location we can't just have a single gbit upstream going to a local 100 switchs for example (like the PoE switch we use for nanobts). > I have never used VLAN, but from what I understand it means that we > share the bandwidth with other network users and thus we hardly could > expect to occupy that much bandwidth. Is that correct Yes, VLAN will share bandwidth on the upstream links / core network. But if the core network is 10Gbe ... Cheers, Sylvain > > -- > Regards, > Alexander Chemeris. > CEO, Fairwaves LLC / ??? ??????? > http://fairwaves.ru > From peter at stuge.se Fri Oct 26 13:19:15 2012 From: peter at stuge.se (Peter Stuge) Date: Fri, 26 Oct 2012 15:19:15 +0200 Subject: LAN access at the venue In-Reply-To: References: <20121025192404.2249.qmail@stuge.se> <20121025205906.9702.qmail@stuge.se> Message-ID: <20121026131915.21550.qmail@stuge.se> Alexander Chemeris wrote: > I'm thinking whether we could operate OpenBTS in "Remote Radio Head" > mode, where base CPU units are located in a server room and > transceivers are streaming I/Q data to them over LAN. In this case we > need a guaranteed bandwidth to avoid packet loss - from 9Mbit for > single-channel TRX with 1SPS to ~72Mbit for dual-channel TRX with > 4SPS. Thanks for the info. I'll talk to the NOC guys about this. Did you already consider some different sites for your deployment? I don't think we have decided for OpenBSC yet. I would actually like to use a MetroSite or two. Transceiver location will matter for how easily NOC will be able to provide dedicated andwidth, if at all. > I have never used VLAN, but from what I understand it means that we > share the bandwidth with other network users and thus we hardly could > expect to occupy that much bandwidth. Is that correct? VLAN in this case means that NOC gives us a private layer 2 network. It doesn't have to mean that the layer 2 network is distributed around the building in a shared medium with other VLANs, although that is the standard practise. It may still be possible for us to have dedicated bandwidth at least in some locations. We might be able to allocate Saal 11 for our use. One server room is nearby, around the first corner (like a hook) south of Saal 11. http://stuge.se/photos/CCH_29C3/01._OG //Peter From peter at stuge.se Wed Oct 31 20:50:54 2012 From: peter at stuge.se (Peter Stuge) Date: Wed, 31 Oct 2012 21:50:54 +0100 Subject: CCH uplink situation so far Message-ID: <20121031205054.25152.qmail@stuge.se> ----- Forwarded message from NOC ----- darkfibre: - will end at IPHH(AS12731) at the ECIX Hamburg Core - currently the fibre ends in the basement of the hotel. they start building a fibre from the hotelbasement to the "Colt Raum" in the cch uplinks: - 10ge KPN (AS286) - 10ge EuroTransit (AS33926) - IPHH as transit to the ECIX - some more will come bit 20gigs should be enough ;P ----- End forwarded message ----- From alexander.chemeris at gmail.com Wed Oct 31 21:13:31 2012 From: alexander.chemeris at gmail.com (Alexander Chemeris) Date: Thu, 1 Nov 2012 01:13:31 +0400 Subject: CCH uplink situation so far In-Reply-To: <20121031205054.25152.qmail@stuge.se> References: <20121031205054.25152.qmail@stuge.se> Message-ID: Thanks, Peter. Is there an option for guaranteed QoS if we want to have a VoIP trunk to outside? On Thu, Nov 1, 2012 at 12:50 AM, Peter Stuge wrote: > ----- Forwarded message from NOC ----- > darkfibre: > - will end at IPHH(AS12731) at the ECIX Hamburg Core > - currently the fibre ends in the basement of the hotel. > they start building a fibre from the hotelbasement to the > "Colt Raum" in the cch > > uplinks: > - 10ge KPN (AS286) > - 10ge EuroTransit (AS33926) > - IPHH as transit to the ECIX > - some more will come bit 20gigs should be enough ;P > ----- End forwarded message ----- > -- Regards, Alexander Chemeris. CEO, Fairwaves LLC / ??? ??????? http://fairwaves.ru