I agree, WiFi direct is mostly an abomination. But, there is a useful thing it offers - discovery. Even if you don’t use it for actual data communication, you can use it to discover nearby devices.
> On Mar 20, 2015, at 1:54 PM, Pieter Hintjens <[email protected]> wrote: > > WiFi direct is an abomination yet might be usable. I've no dogma here. > What works works. > > On Fri, Mar 20, 2015 at 5:53 PM, Utsav Drolia <[email protected]> wrote: >> Apart from other discovery methods, have you thought of other communication >> technologies? >> For example, wifi-direct, Wifi 802.11z (TDLS), or bluetooth (if the data >> consists of small messages). >> >> I know adhoc-wifi is pretty much shot on most devices - but wifi-direct is >> available (at least on Android). >> Given the way wifi-direct works, zyre should work on it the exact same way >> it works over wifi access points. >> The only issue with wifi-direct on Android is the annoying pop-up to accept >> the connection every time a new connection takes place. >> >> >>> On Mar 20, 2015, at 9:41 AM, Pieter Hintjens <[email protected]> wrote: >>> >>> Could you provide me a spec for the multicast? Then I'll add it to the RFC. >>> >>> The other discovery methods I'd like to eventually make are bluetooth >>> and manual (i.e. provide the Zyre API with node endpoints taken from >>> arbitrary mechanisms). >>> >>> On Fri, Mar 20, 2015 at 10:34 AM, Arnaud Loonstra <[email protected]> >>> wrote: >>>> Zbeacon in Pyre supports multicast which can traverse network segments. >>>> As long as your network supports it. >>>> >>>> Discovery mechanisms are kind of difficult. We have 3 methods now in >>>> different versions of Zyre/Pyre/... >>>> - broadcast beacon >>>> - multicast beacon >>>> - gossip >>>> >>>> I'm hoping to get some more methods but I won't be working on that >>>> until the end of the year. If anybody has good pointers, ideas about >>>> useful mechanisms please post it. >>>> >>>> Rg, >>>> >>>> Arnaud >>>> >>>> On 2015-03-20 00:18, Joss Gray wrote: >>>>> zbeacon broadcasts udp packets broadcast for discovery so they wont >>>>> be >>>>> able to cross subnets. >>>>> >>>>> You will have to change the broadcast domain like you mentioned. >>>>> >>>>> On Thu, Mar 19, 2015 at 11:08 PM, Andrew Hume <[email protected] >>>>> [3]> >>>>> wrote: >>>>> >>>>>> we’re trying to run tyre (actually, the gyre implementation) >>>>>> across different subnets >>>>>> (that is, one node is in 10.20.59.xx and the other in >>>>>> 10.120.45.xx). they don’t see each other. >>>>>> what would be the best way to have them see each other? >>>>>> >>>>>> we could change the broadcast domain to 255.255.0.0 but that seems >>>>>> vaguely dodgy. >>>>>> any other hints? >>>>>> >>>>>> andrew >>>>>> >>>>>> _______________________________________________ >>>>>> zeromq-dev mailing list >>>>>> [email protected] [1] >>>>>> http://lists.zeromq.org/mailman/listinfo/zeromq-dev [2] >>>>> >>>>> >>>>> >>>>> Links: >>>>> ------ >>>>> [1] mailto:[email protected] >>>>> [2] http://lists.zeromq.org/mailman/listinfo/zeromq-dev >>>>> [3] mailto:[email protected] >>>> >>>> _______________________________________________ >>>> zeromq-dev mailing list >>>> [email protected] >>>> http://lists.zeromq.org/mailman/listinfo/zeromq-dev >>> _______________________________________________ >>> zeromq-dev mailing list >>> [email protected] >>> http://lists.zeromq.org/mailman/listinfo/zeromq-dev >> >> _______________________________________________ >> zeromq-dev mailing list >> [email protected] >> http://lists.zeromq.org/mailman/listinfo/zeromq-dev > _______________________________________________ > zeromq-dev mailing list > [email protected] > http://lists.zeromq.org/mailman/listinfo/zeromq-dev _______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
