[HacktionLab] Open Source / federated VOIP?

Wed Oct 14 17:42:54 UTC 2015

Hey, thanks for all the comments!

Ring.cx does look good...will try and have a go soon and see if it
actually does as it promises soon!

And unfortunately being on debian so far only have firefox (iceweasel)
38 which isn't supported by jitsi meet, so i'll have to wait to update
before trying that...nice one on getting a test server up and running John!

Cheers,

Nick

On 10/10/15 12:48, johnc wrote:
> Hi,
>
> Firefox uses the telecoms company Telefonica to proxy calls. The calls
> are encrypted but not end to end. Interception would be a concern at the
> proxy. A similar situation exist as regards Tokbox.
>
> There are still major problems with webrtc, the main ones being the lack
> of end to end encryption and the reliance on X509 certs issues by the
> certificate authorities. This blogs post was written by Daniel Pocock
> about a year ago explaining the issues but not much has changed since
> then.
>
> http://danielpocock.com/is-webrtc-private
>
> On a slightly more positive note, if you run your own trusted server and
> are prepared to carefully check SSL certs have a look at jitsi meet.
>
> https://meet.jit.si/
> https://github.com/jitsi/jitsi-meet
>
> It's a browser based multi-party video conferencing solution. This
> rocks! I set it up for work recently as we have sections in a couple of
> other European countries. You can share presentations and even your
> desktop, It's better than a lot of propitiatory solutions.
>
> If you configure this companion piece of software: jigasi
>
> https://github.com/jitsi/jigasi
>
> You can plumb in a regular (insecure) telephone number into a web
> conference which is quite useful also.
>
>
> I will build and or demo one of these systems if anyone is interested at
> the next meet up.
>
> Cheers,
> John
>
> On 06/10/15 23:28, nmd wrote:
> > Hi,
>
> > Thought I'd just resurrect this thread to ask what y'all think of some
> > of the new webrtc offerings out there?
>
> > Firefox Hello being one of the obvious ones, opentokrtc.com also allows
> > multiple people to video call (without an account), and then there's
> > others like appear.in which use webrtc but appear not to be fully open
> > source (well it says nothing about it, so i don't know if that just
> > means that the server-side is proprietary).
>
> > Any thoughts on how secure these kind of things are? I haven't seen any
> > support for OTR etc. that desktop voip programs have the potential for..
>
>
>
> > Cheers,
>
> > Nick
>
>
>
>
>
> > On Friday 16 Jan 2015 14:19:46 you wrote:
>
> >> That's really useful clarification - thanks!
>
> >>
>
> >> I agree that not having a track record is a negative cf a coder who
> does.
>
> >> Being able to identify them in meatspace is not the main issue, having
>
> >> trust is.
>
> >>
>
> >> Many thanks for this interesting and important discussion
>
> >>
>
> >> Keep on hacking!
>
> >>
>
> >> love
>
> >>
>
> >> g
>
> >>
>
> >> On 16 January 2015 at 14:08, johnc <johnc at aktivix.org> wrote:
>
> > Hash: SHA1
>
>
>
> > Hi,
>
>
>
> > The tox developer who goes by the name of irungentoo has not published
>
> > any other work (at least under that name) that I can find. I'm more
>
> > bothered about the lack of a proven track record than personally
>
> > identifying the individual to be honest.
>
>
>
> > By contrast take ZRTP voice encryption. This was written by Phil
>
> > Zimmermann who also wrote PGP (which GPG is based on). Zimmerman is a
>
> > respected programmer and cryptographic expert with a proven track
>
> > record. Blackphone also comes from the same stable.
>
>
>
> > Cheers,
>
> > John
>
>
>
> > On 14/01/15 16:12, Gareth Coleman wrote:
>
> >> Hiya John
>
>
>
> >> Just a thought - would you always put an anonymous author as a
> >> negative?
>
>
>
> >> loads of love
>
>
>
> >> g
>
>
>
> >> On 14 January 2015 at 16:07, johnc <johnc at aktivix.org
>
> >> <mailto:johnc at aktivix.org>> wrote:
>
>
>
> >> I haven't used this yet but have done a little research on the
> >> program:
>
>
>
> >> +'s:
>
> >> - it easy to use
>
> >> - Available on a range of OS's + devices.
>
> >> - reasonably decentralised - but I believe it still has some known
>
> >> "good/online" user IP's hard coded for bootstrapping like bitcoin has.
>
>
>
> >> -'s:
>
> >> - Currently no independent security review has been performed.
>
> >> - It's DHT implementation is vulnerable to Cybil attack=> not that
>
> >> hard
>
> >> to DOS.
>
> >> - No group voice chat.
>
> >> - Main developer (irungentoo) is anonymous.
>
>
>
> >> On 01/12/15 22:17, Jim McTwanky wrote:
>
> >>> .....educated /opinion/......that is.
>
> >>>
>
> >>> On 11/01/15 20:44, nmd wrote:
>
> >>>> Hi, Thanks for everyone's suggestions and thoughts - I tried
>
>
>
> >> jitsi and was
>
>
>
> >>>> semi-successful (got sound and video working in one direction
> >> and the
>
> >>>> problems
>
> >>>> may have been specific to one of the laptops). Might try that
>
>
>
> >> again in the
>
>
>
> >>>> future but will also hope that these things get a bit simpler
>
>
>
> >> with time!
>
>
>
> >>>> Cheers,
>
> >>>> Nick
>
> >>>>
>
> >>>> On 02/01/15 11:26, Tim Dobson wrote:
>
> >>>>> On 31/12/14 16:24, johnc wrote:
>
> >>>>>> Some Problems: -Mobile phone specific: -- mobile phones vary
>
> >>>>>> greatly in their ability to run sip clients using crypto. I've
> >> seen
>
> >>>>>> sip clients use 100%CPU with awful audio quality on a few phones
>
> >>>>>> including high end samsung models. -- The latency on 3G is
>
> >>>>>> typically around 1 second. Expect horrible lag etc. Using WiFi is
>
> >>>>>> the only way to go unless you are lucky enough to be on 4G.
>
> >>>>>> Non mobile phone specific: - ostel's only server is in the US,
>
> >>>>>> latency is about 120ms. Not so good if you are in Europe. We could
>
> >>>>>> build our own :-). - If you are going to build an ostel system I
>
> >>>>>> suggest you include the topology hiding setup from my wiki or
>
> >>>>>> elsewhere in your Kamailio config. SIP leaks IP/location
>
> >>>>>> information unless you make an effort to obfuscate it.
>
> >>>>>
>
> >>>>> One solution I quite like, which works *if* you:
>
> >>>>> a) trust the clients to a degree
>
> >>>>> b) are happy with non-federated, centralised phone system, with the
>
> >>>>> PBX as a single point of failure
>
> >>>>>
>
> >>>>> is:
>
> >>>>>
>
> >>>>> Your favourite SIP-based PBX system over OpenVPN.
>
> >>>>>
>
> >>>>> So, your phone connects to OpenVPN, and then the sip clients
>
>
>
> >> connects
>
>
>
> >>>>> to the PBX via SIP, over a VPN.
>
> >>>>>
>
> >>>>> Pros:
>
> >>>>> a) as secure as your deployment of OpenVPN
>
> >>>>> b) removes NAT issues - there aren't any - the SIP/RTP goes via
>
>
>
> >> OpenVPN
>
>
>
> >>>>> c) It mostly 'just works' (tested with .bg client connected to .uk
>
> >>>>> server with no issues)
>
> >>>>> d) possible on mobile [android], desktop and in modern Snom
>
>
>
> >> firmwares
>
>
>
> >>>>> Cons:
>
> >>>>> a) nontrival to setup
>
> >>>>> b) centralised [not federated, and not designed to be]
>
> >>>>> c) requires the giving out of VPN certificates to each client in
>
>
>
> >> advance
>
>
>
> >>>>> d) SPOF [or compromise] on PBX system
>
> >>>>> e) not really possible to 'just leave on' on mobile without
> >> emptying
>
> >>>>> your battery
>
> >>>>> f) only known to be *super reliable* on Snom desk phones,
>
>
>
> >> connected to
>
>
>
> >>>>> an uncongested network
>
> >>>>> g) certainly not without points of weakness
>
> >>>>>
>
> >>>>> ---
>
> >>>>>
>
> >>>>> It's not foolproof. It's not bombproof. But it is a nice
>
>
>
> >> architecture
>
>
>
> >>>>> that works for some scenarios. :)
>
> >>>>>
>
> >>>>> -Tim
>
> >>>>>
>
> >>>>> _______________________________________________
>
> >>>>> HacktionLab mailing list
>
> >>>>> HacktionLab at lists.aktivix.org
> >> <mailto:HacktionLab at lists.aktivix.org>
>
> >>>>> https://lists.aktivix.org/mailman/listinfo/hacktionlab
>
> >>>>
>
> >>>> _______________________________________________
>
> >>>> HacktionLab mailing list
>
> >>>> HacktionLab at lists.aktivix.org <mailto:HacktionLab at lists.aktivix.org>
>
> >>>> https://lists.aktivix.org/mailman/listinfo/hacktionlab
>
> >>>
>
> >>> _______________________________________________
>
> >>> HacktionLab mailing list
>
> >>> HacktionLab at lists.aktivix.org <mailto:HacktionLab at lists.aktivix.org>
>
> >>> https://lists.aktivix.org/mailman/listinfo/hacktionlab
>
> >>>
>
> >> _______________________________________________
>
> >> HacktionLab mailing list
>
> >> HacktionLab at lists.aktivix.org <mailto:HacktionLab at lists.aktivix.org>
>
> >> https://lists.aktivix.org/mailman/listinfo/hacktionlab
>
>
>
> >> --
>
> >> ------------------------
>
> >> Gareth Coleman
>
> >> layer zero labs
>
> >> l0l.org.uk <http://l0l.org.uk>
>
>
>
>
> >>>
>
> >>> _______________________________________________
>
> >>> HacktionLab mailing list
>
> >>> HacktionLab at lists.aktivix.org
>
> >>> https://lists.aktivix.org/mailman/listinfo/hacktionlab
>
>
>
> > _______________________________________________
> > HacktionLab mailing list
> > HacktionLab at lists.aktivix.org
> > https://lists.aktivix.org/mailman/listinfo/hacktionlab
>
>
>
> _______________________________________________
> HacktionLab mailing list
> HacktionLab at lists.aktivix.org
> https://lists.aktivix.org/mailman/listinfo/hacktionlab

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.aktivix.org/pipermail/hacktionlab/attachments/20151014/f31ee650/attachment.html>