 Okay, this means we finished early so we can take some more questions that we couldn't take before so we have 10 minutes for postponed questions if people can still remember their original questions Yeah, and we have a Whose mouse was that by the way it can be return. Okay. Yeah Okay, so any questions that was still postponed. Okay. Yeah, I had a question regarding the GPRS and Channels and how many Basically, if if a user is connected via GPRS, does it use up that traffic channel? And how long does it use it up and how many users can use that that traffic channel? He's sorry. Could you repeat it? Okay, I'll answer sorry so The allocation is dynamic on GPRS. It's not unlike circuit switched where basically the the Each call takes up one time slot whether it's a full rate or a half rate time slot In GPRS basically that even inside a single time slot the allocation is dynamic So basically if there is data for 10 phones, then it will Take some some messages are sent for one phone some messages and for another phone. So it's completely dynamic Like you would have in any other packet oriented shared media, let's say Wi-Fi or something like that. Okay Well, there you have collisions. So it doesn't really match that way, but okay. Anyway, it's dynamically allocated So there is no theoretical limit on how many Phones can use that channel at the same time. Of course at some point it will get infinitely slow So that's a practical limit, but theoretically you can have you know as many phones as you want And they can exchange data over concurrently even over single a time slot or over collection of time slots There is one limitation though in the number of temporary flow identifiers because each phone that currently actively transmits an IP packet has What's called this TBF the temporary block flow and I think it's 16 is the maximum But I'm not 32. Does anyone know exactly? I don't know exactly how many but can be like 16 temporary 64 even okay. Yeah, thank you So you can have 64 concurrently on a protocol level in parallel But then as soon as one temporary block flow has ended, let's say the the the website has been downloaded Like the current page your opening has been downloaded completely and the user is just reading Then this temporary block flow is closed and the identifier is freed again. So another subscriber can use it So it's very Dynamic and I wasn't asked but following on this. So this is on a protocol level, but specifically on an implementation level how does How does the osmo verse handle a fair resource allocation as far as as far as gprs and edge I Don't think we fair. I don't even think we do. I mean, I don't even think we do much effort I'm not sure whether Holger has something to say about that Do you have something to say about it? Very little okay, I mean the point with gprs is right We had very very very limited resources to implement something as complex as gprs So we are happy that we have something that works at all right and and fair allocation and other esoteric topics Which of course in reality matter, but that's sort of where you know who who has an interest in in donating resources to that I mean, we're always very happy to you have okay Sure Specifically in a way that doesn't make it seem like it deserves an apology, which it doesn't we are all happy to have what we have Or we should put our money where our mouth is But I'll be more specific say for example you create a cell that has You know one whatever the required time slot is on tea at like the first tea is zero and then one through seven are all PDCH different You know MS hardware can support different like channels and block allocations or whatever else Do you max it out every I suppose I could answer this question by looking into the code But it's more interesting in the form of a discussion Yeah, look at you code now. It's it is complicated in in many ways So there are different MS classes that have different requirements how quickly they can switch from receive to transmit and You have different amount of channel combinations available Like seven slots in a row or some booth holds and what we try to do is Spread allocations equally across the PDCH. So even if you have a mobile station That is assigned to multiple time slots There's always one control time slot where the TFI is actually being used and we try to Allocate MS across different time slots equally But depending on the channel combinations You have it already limits of how many time slots you can allocate So if you're the first time slot already has some users you try to use the second one But then depending on the MS class you can only allocate two time slots and ten instead of three So this is where fairness is already Failing but the only other part is that what Daniel mentioned earlier is that you have like an allocation algorithm And he mentioned dynamic and the dynamic part is that if the cell is really really crowded We go from like multi-slot allocations to single slot allocations On the fly depending on on the loads here. Thank you I have to continue. I'm do we have more questions Three more minutes three more questions that keys Did I okay, I had a question about the femtosell and I Believe that the that one that you have sourced on that you were using it There isn't anything particularly special about it, right? It's just in a standard mode and you can access it But is there Are there any considerations in like which models there was there was something about one that was opened, right? That was what was what was good? I was you consider consideration is for you What femtosell you can actually get Okay, that's really or how much you're willing to pay because easy market is very limited They tried to have like a huge market for femtosells Which means the company's having a business model femtosells wanted to sell them in 100,000 units so mostly the operator customers tried to buy them and there's not really an open market for 3g femtosells Which was the one that that some firmware got accidentally Many I can refer to Kevin if he wants to give her a small overview So the very first femtosell which got somehow hacked was by THC. It's the voodafone UK femtosell But There's not a lot of documentation except from from this group. The second one is the one I worked on It was the SFR femtosell which was ubiquitous then Broadcom Then pico chip or all the way around They also have a 3g a newer model now, which isn't broken. The older model is still broken Then there is the one from Huawei You can access your broken broken means the security we found the security weakness so so we can access it There is one from Huawei But which we played quite of some some time with but each model We both had complete different configurations. Sometimes you art was there sometimes JPEG was there sometimes the The firmware was completely locked. Sometimes the symbols were stripped. So we stopped with it How I'd also spoke about another one where we didn't get any provision This is the exact same model then the one from the newer one from SFR, which has no where no weakness has been found in a slow time and For 3g, I think that's that's the big overview except now the one which is used now by By Osmo comm and so for the accelerate 3g. This is the one which is easiest to use because You can still find it. You don't have to get through to an operator Because previously you had to get to voodafone UK SFR France. So you need a subscription, which is quite a hassle and Yeah, they are the tools and it is open. So I think the best bet would be try to found whatever they are using One more minute another question No, sorry, it's nothing One one of the issue of using the regulated frequencies, of course the legal aspect of it in moving from BTS to femtocell do we soften that issue or not? Do we suffer from that issue? Soften? No, not really because I mean with GSM depending on the jurisdiction you're in depending on the country it's still possible to get test licenses and More or less much more easily because it's only 270 kilohertz and it's more likely that somebody has 270 Hertz 270 kilohertz of free spectrum somewhere that you can use as opposed to 3g where you need 5 megahertz of spectrum Which is a very large chunk which is much less likely that anyone has a spare 5 megahertz somewhere that you can use At a given location. So that doesn't really improve the situation. I mean, maybe on obscure bands But then you would need hardware that supports obscure bands Which makes your choice of femtocells even much less limited unless you go for like one of those software defined small cells Which have no band filters and so on and they're completely reconfigurable, but then what phones do you find for the obscure band? So it's Doesn't really improve the situation