Re: End-to-end encrypting US GSM phones?
Quoting Trei, Peter <ptrei@rsasecurity.com>:
I don't quite understand why getting encrypted VoIP working on a Palm Pilot (16-33 MHz dragonball) is the 'Holy Grail'. I would have said that the HG is an affordable, working, portable and pocketable encrypted VoIP system with free software, regardless of platform. The Ipaq is 2-300$ more than a Palm device, but is that the whole issue? Ipaqs have much better cpu power and memory, and decent Linux ports. Go for the low-hanging fruit first.
"Everyone" has palm pilots already. WinCE-based PocketPCs haven't made much of a dent in the marketplace. There is also a very large developer community for palm apps, and they're widely deployed in corporations. If you're assuming users will buy a dedicated device *and* put linux on it, that's reasonable (or sell pre-packaged systems). Otherwise, you also need to develop for WinCE on the PocketPC. I still would like to see from someone who is good at integer voice encoding whether a 33MHz dragonball can do tolerable-quality voice at 14.4kbps; I'm sure it's good enough to distinguish "ki" from "eight-ball", but I'd like something good enough to distinguish "weight, y'all" from "eight-ball" Since it's very unlikely any pda will have full duplex audio anytime soon, I think the bluetooth headset is the only viable solution; I'm going to get one for my t39m soon enough, and then I'll experiment with linux bluetooth support and try to get bluetooth audio integrated into OSS on my laptop, as a proof of concept. That done, it's just a question of running linux on an ipaq for a simple demo, with libgsm compiled as integer; then maybe something more efficient. I'm pretty sure someone has made the bluetooth headset work with linux already, so it's just a question of kernel patches. The security of this would depend somewhat on the bluetooth encryption, although I'd be happier trusting unencrypted but low-propagation bluetooth vs. a PSTN/cell link. -- Ryan Lackey [RL7618 RL5931-RIPE] ryan@havenco.com CTO and Co-founder, HavenCo Ltd. +44 7970 633 277 the free world just milliseconds away http://www.havenco.com/ OpenPGP 4096: B8B8 3D95 F940 9760 C64B DE90 07AD BE07 D2E0 301F
Ryan wrote:
"Everyone" has palm pilots already. WinCE-based PocketPCs haven't made much of a dent in the marketplace. There is also a very large developer community for palm apps, and they're widely deployed in corporations.
I am not sure that the existance of a large developer community has much bearing on the suitability of Palm as an encrypting phone platform. As for the hardware, it simply is underpowered. You can spend man-months trying to fight the current underpowered Palm platform or you can use the sufficiently-resourced PowerPC platform. Since I don't believe that there is a requirement for the feature to operate on a device already in the user's possession, I know what my choice would be. YMMV.
If you're assuming users will buy a dedicated device *and* put linux on it, that's reasonable (or sell pre-packaged systems). Otherwise, you also need to develop for WinCE on the PocketPC.
The OS is really of secondary or tertiary concern here. The more important question is which (if any) handheld hardware supports full-duplex audio. Do we know for fact that the lack of full duplex audio support on the VoIP handheld demo is due to lack of support in the HW or could it be a lack of support in the WinCE OS? --Lucky
participants (2)
-
Lucky Green -
Ryan Lackey