libera/#neo900/ Wednesday, 2018-07-04

houkimeGemalto is one of the 52 companies which Facebook had "integration partnerships". Just great(16:46
houkime*partheships" with.16:47
houkimereally wish the best of luck to OpenBTS so that there's a hope to end this modem nightmare.16:49
houkimeAlso situation with gecko and chromium (and by extension electron) kits bothers me. Both are not community-lead and both have shown illicit behavior.16:58
houkime2018 is a real privacy nightmare. But also a good time to do sth with all that mess.17:02
houkimeThe question is - is the open dev community large enough to cope wth it?17:02
Joerg-Neo900OpenBTS won't help change anything in the mobile phone industry and ecosystem19:00
Joerg-Neo900it's not like it was the lack of BTS hardware that stops the emerge of "Open" carriers19:01
Joerg-Neo900it's not even the lack of open modem chipsets. The problem is legal rules forbid operation of uncertified radios. And in fact this is the one fundamental rule that makes the whole system actually work reliably19:04
Joerg-Neo900about "integration partnerships" I think it's mainly facebook to blame, not the companies that used this opportunity to legally optimize their marketing19:09
houkimeif this system needs such rules to work it needs to be redesigned.19:11
houkimeand obviously one way to do it is to omit it entirely and replace by voice-over-ip. And for ISP use sth like next-gen freifunk19:17
houkimeyeah. of course, gemalto hardware focusing on security of their users data may have the legitimate right to abuse their Potential users data to toss their ads at them. This might be legal, but the contradiction is quite evident and by doing so they are indeed guilty for not following their own proclaimed goals.20:11
houkimeif they do so, they in fact Start your relations with them that are supposed to be safe with trickery.20:12
Joerg-Neo900sorry I can't follow20:18
houkimegemalto is a company which produces modems for Neo900. Besides being in an old scandal with a data heist, it also appeared in the list of companies Facebook was in integration partnership with.20:21
houkimeyou said that these companies are not really guilty, however in gemalto's case it isn't really so because they proclaim to be on a pro-privacy side.20:22
houkimeand it is their selling point.20:22
Joerg-Neo900you're stating that but I don't see how you come to that concusion20:23
houkimeerr... their whole website is about this basically.
Joerg-Neo900and I see no link between any (im)possible interpretation of any details related to the facebook stuff and any BTS/WWAN related considerations20:26
Joerg-Neo900what is "this"?20:26
houkimeclaims. If they haven't had claims this would be totally ok for them to be integrated.20:28
houkimeoffer of privacy and security which they break right at start if they are integrated20:28
houkimeBTS and WWAN are connected to gemalto in neo900's case because the cellular network for neo900 is working via gemalto's phs-820:31
houkimewhich we have in the sch20:31
houkimethat was the starting point of a rant about openBTS and my general will to not use closed modems and networks in the long perspective.20:34
houkimewhich is not really possible now, but may be possible in the future.20:35
Joerg-Neo900please elaborate what are those claims and how Gemalto breaks any such claims they did, by using a API provided by Feacebook to get data from facebook20:35
Joerg-Neo900and no, since the legal situation is the way it is and it's for a reason, to make the whole system stable and working and not getting messed up by anarchist transmitter operators not obeying the air interface and protocol specs, you won't see any really open and legal modems any time into the future20:39
Joerg-Neo900there's one root factoid that is bvery popular to get ignored or neglected or simply not be aware of: to operate ANY TRANSMITTER you need a approval or certificate and you need to follow the regulations that come with that, basically everywhere on this globe. If you don't, you're prone to legal prosecution and for sure any business concept for a product that deliberately ignores this fact is not very likely to succeed20:47
Joerg-Neo900this it true for *all* transmitters, no matter if they are ISM, WLAN, WWAN, radar or a microwave oven20:48
Joerg-Neo900iirc the frequency range covered starts at basically 0Hz (with a general aproval for emissions up to iirc 20kHz or somesuch)  and ends at one or two magnitudes below near infrared20:50
Joerg-Neo900no matter which band and protocol you choose for your cellular phone network, you _always_ will need to obey the applicable regulations20:52
houkimeI would rather undergo a process of certification of *my personal* settings of open modem rather then use a closed brick.20:55
houkimeit can be made simple enough.20:55
Joerg-Neo900you're free to do that. Just note please that part of the regulations the cert is based on demand that the user cannot change the way the transmitter works20:56
houkimethispoint seems like outright crime and needs to be changed politically.Besides, I will likely just do an open modem and THEN see what they do.20:57
houkimethey can't ban the design itself.20:57
Joerg-Neo900I think there actually exist a few GSM/2G modems that have fully disclosed and thus "open" firmware stack20:58
Joerg-Neo900I don't understand what you're trying to accomplish20:59
houkimeI'm trying to accomplish me controlling my own transmissions.20:59
houkimethat's all20:59
Joerg-Neo900won't happen unless you deliberately accept you're doing something illegal20:59
houkimeAny man at any second can do a crime. This doesn't mean that someone needs to be in a jail from birth, right?21:01
houkimedenying users of the right to have their settings and control is jailing from birth21:02
Joerg-Neo900sorry, I can't follow21:02
houkimelet's punish users AFTER they do sth bad rather then prevent them from having freedom, that's my point21:03
Joerg-Neo900depends on definition of "something bad"21:03
Joerg-Neo900drug *possession* is illegal in most jurisdictions21:04
Joerg-Neo900so yes, you're in that jail of not being allowed to possess drugs, from birth to death21:04
Joerg-Neo900you can build a chipset that works as modem, you can write frimware for it. All fine. But you never will get a cert for it unless you make sure nobody but you can change the firmware21:06
houkimeThat's ok. I'll just publish it.21:06
houkimeopen source21:06
Joerg-Neo900no problem21:06
Joerg-Neo900but Neo900 is meant for customers who want to operate a legal and certified product21:07
Joerg-Neo900and that implies that the customers must not be allowed to for example modify/change the IMEI of the modem21:08
Joerg-Neo900nor any other of the bazillion mandatory parameters in the air interface spec21:09
Joerg-Neo900since the whole network is a "cooperative bus infra"21:09
Joerg-Neo900which means that *all* participants need to adhere to the specs to allow the complete system to operate as designed21:10
Joerg-Neo900the specs are public and everybody free to study them21:11
houkimeHmm... Well, why not set these specs in silicon and have the chip and drivers open?21:12
houkimeso that a master branch can be certified21:12
houkimeand the operation os legal21:13
Joerg-Neo900there's a whole class of devices like network analyzers (refer CMU200 etc) that actually alow user to modify whatever they like in any of the air interface parameters. But those devices are not certified for general purpose use as a phoone in existing public networks21:14
Joerg-Neo900yes, that would be possible in theory21:14
Joerg-Neo900and is actually what WLAN chip manufs try to do now - well a very few of them21:14
houkimemy biggest concern is that chip and drivers producers squeeze extra stuff in there.21:15
houkimeso a good open chip with some KNOWN settings fixed would be quite nice.21:17
Joerg-Neo900this is a valid cncern, however it's a too short shot. Consider this: modems all have a OverTheAir frimware update feature nowadays, so you can't even *know* if the firmware your modem is running is still the same you evaluated and audited a week ago. Obviously "code in silicon" would defeat that, but it's incredibly expensive and also defeats all possibilities to fix bugs later on21:17
houkimeThen one needs to have a chip designed without this feature. And code in silicon only constant numbers that correspond to net parameters.21:20
houkimebesides that is not that hard21:21
houkimeif you use FPGA for debug21:21
Joerg-Neo900Neo900 approach is radically different from the "audit the modem 'firmaware` ": since the modem is incredibly complex and thus more complicated to actually audit comprehensively than a full linux system, and since it could change from not to next minute aven after you accomplished thius mission impossible to audit it, we chosen the approach to sandbox and monitor the complete blackbox called modem, to make sure none of the threats possibly21:23
Joerg-Neo900introduced by it can take any effect since we detect and counteract every single one of them21:23
Joerg-Neo900s/from not/from now/21:24
enycJoerg-Neo900: not to bention 'in any case phone does not control whats going on in the network' ....21:25
houkimeok, and my long-shot desire then will be to have a soft-modem-on-FPGA, because this way publishing this on the net would be enough for people to start using it.21:29
houkime'even if it is illegal21:29
Joerg-Neo900go for osmocom stuff21:30
* Joerg-Neo900 idly muses if a 'simple' tcp/ip stack is implementable on a fpga without adding a virtual MCU core incl 'ROM' with 'firmware' to the FPGA setup21:32
enychoukime: whtas' this  FP515 blind-via-requiring thing?21:32
Joerg-Neo900GSM RAT is a tad more involved than TCP/IP stack, I'd say21:32
Joerg-Neo900the OMAP3 footprint I'd guess21:36
enyc"12:16 <+DocScrutinizer05> I'd like to see us demonstrating mastering the needed skillset in proto_v2 already, *before* we go proto_v3 with FP515 that needs inpad capped blind vias no matter what"21:36
houkimeI guess it just has too much density needed.21:37
enycim ware of the classic 'feature creep' ... we want this to work and be achievable/completable21:37
Joerg-Neo900how's that related?21:38
enycgood question, for some reason cropping up in my brain in this location21:38
enycbeen very uugh today and just woke up again so not thinking so clearly21:38
Joerg-Neo900when you got a chip with a 0.4mm pitch BGA with 515 balls, you can't route traces between the balls so you need multiple layers and you need vias to connect from solder pads to those21:39
Joerg-Neo900and you can't run a via for a pin in a more 'outer' ring of balls straight through the complete PCB stack from one surface to the other, piercing traces on lower layers by doing so. Thus you need blind vias21:41
Joerg-Neo900and you need in-pad vias since the "free space" _between_ two solder pads on 0.4mm pitch is way too smal to place the vias outside the pads between them21:44
Joerg-Neo900refer "dog bone vias" which are such offset of vias to the space between pads and don't fly for 0.4mm21:45
Joerg-Neo900I.E this doesn't work for 0.4 pitch21:47
Joerg-Neo900none of it21:47
Joerg-Neo900another terse summary:
Joerg-Neo900we need copper plated in-pad blind vias21:52
Joerg-Neo900TI's paper on 0.5mm pitch:
Joerg-Neo90020.2 Via-In-Pad, Microvias, Stacked Vias, Filled Vias Introduction22:00
Joerg-Neo90021  Know Your Tools22:06
houkimebtw what are those wlan chips you mentioned? <Joerg-Neo900> and is actually what WLAN chip manufs try to do now - well a very few of them22:16
houkimei mean what models?22:16
houkimeor they don't exist yet?22:17
Joerg-Neo900sorry, I have no reference at hand. But with WLAN you have the general problem of regulations getting more tough to enforce air interface parameters even in "hackable" routers etc, to stop the messing around with WLAN that's commonly found practice noweadays. So EU and USA passed new rules that affect that topic22:17
Joerg-Neo900for now only affects 5GHz iirc22:18
Joerg-Neo900prolly a reason why lately more devices come with 2.4GHz WLAN nly, after 2.4&5 been almost defacto standard22:19
Joerg-Neo900the nasty "radar coexistence" thing on 5GHz22:20
Joerg-Neo900I heard the artifacts you occasionally can see on weather forcest rain radar maps are caused by a single messed-up 802.11a WLAN device22:21
houkimeI think there will be time when such regulations will not be longer enforcable et all. For example when one will be able to make electronics at home.22:24
houkimeso it's strategically better for them to learn how to do forecasts the other way.22:25
houkimei'm quite sad they try to regulate instead of redesigning and adaptation. Impeding the progress.22:27
Joerg-Neo900I'm quite happy about every standard that does NOT get "redesigned" after 6 months, forcing me to uy new devices *again*. Actually when everybody does what they like, there _are no_ standards anymore, and nothing will work22:33
Joerg-Neo900anarchy isn't progress22:33
Joerg-Neo900actually standards got invented when the world realized anarchy doesn't work. Was around the time the titanic sunk, for global RF regulations22:35
houkimeSo what? Ban independent electronics production? Deprive humans of freedom to make? This won't work either.22:37
houkimeI think wide spreading throughout space to limit interference is pretty much the only answer.22:38
Joerg-Neo900regulations are always enforcable as long as you can notice violations against them. And by the time silly kids print regulations-smashing transmitters on their 3D printers, there wikll be technology like an armada of RF-sensing drones to spot those transmitters and make the patents pay massife fines and teach their kids to not mess up established services like radar or GPS or aeronautics safety systems22:39
Joerg-Neo900the parents, actually22:40
houkimeI don't want to live in your future then. I will go to space and find security and freedom through obscurity.22:40
Joerg-Neo900good luck22:40
houkimeyour future is too interconnected for its own good22:41
houkimeit will bang one way or another. Probably terrorist will kill it.22:41
houkimeyou will need to chip people and alter their thought for them not to break anything.22:42
Joerg-Neo900no, a silly kid will, abusing their DNS toy system to breed a killer virus instead of glowing daisies22:42
Joerg-Neo900or another idiot kid writing a computer virus for fun, with more impact that they anticipated22:44
houkimeso in the system where free autonomous space dwellers and highly-regulated planet people initially coexist, the planet people find the way to wipe themselves out while aoutonomous and obscure space dwellers physically can't wipe the whole population.22:47
Joerg-Neo900regarding "wide spreading throughout space to limit interference" I couldn't care less about what the idiot kiddies on a future mars colony will do with their homemade transmitters. As long as they don't mess with their DNS experimenter toys, or alternatively never have a spaceship come from Mars to Earth22:48
houkimeand that't the point. If spreaded throughout asteroid belt, the mass extinction is nearly impossible.22:49
houkimeso anarchy is perfectly viable22:49
houkimebut not on the planet surface22:50
Joerg-Neo900I'm in sympathy with an anarchism that accepts the general rule no never let anything escape your own domain and reach to individuals that wasn't agreed upon before. Be that noise, interfering RF, toxic or annoying substances in air or water, high velocity lead bullets. laser beams, whatever22:54
Joerg-Neo900keep your anarchism to your own backyard, since it's not better or inferior to my very own anarchism and must niot interfere with mine as mine must not interfere with yours22:56
houkimeVastness of space and obscurity enforces this rule by simple physics. So in my vision this is as close to perfect as it can be.22:58

Generated by 2.17.0 by Marius Gedminas - find it at!