Chris Allen schrieb:
> hmm, well I'm not sure I understand the question, but... In the
> project I did for MGH it had nothing to do with Red5, in fact Red5 was
> in its very early stages at that point.
>
>   
you answhered the question later in here ;)

> I see what you are trying to do. You are considering making Red5 a
> type of proxy between Flash and the Jabber server. It should work just
> fine. You can have Red5 act as a Jabber client in that case, or there
> is also a protocol for server to server communication in XMPP. Either
> way the Red5 server becomes a sort of client to the XMPP/Jabber
> server.
>   
thats right and jabber because it offers many options also a secur 
protocal and has a wide range of users
and interface´s to allmost any messaging system.

> Yeah, that is correct. We were looking for a way for people to use the
> XMLSocket class as an alternative to RTMP, XMl-RPC or AMF. There could
> be all kinds of other possibilites as well
>
>   

^^ here you answhered what my first question was, thanks thats good to 
know:D.
> It should keep the socket open. you may at some point be concerned
> that you may need to store partial XMPP Stanzas between packets
> received. There is a specific decoder class to handle this case:
> http://directory.apache.org/subprojects/mina/apidocs/org/apache/mina/filter/codec/CumulativeProtocolDecoder.html
>
>   
thanks again i will have a look at that, but if it holds the socket open 
does it also check the client status anyway, in the non MINA basic 
bufferedReader socketserver
i made i had just one problem , recognizing when a client is dead 
instead of idleing or busy, a busy client wouldent respond aswell, a 
ping to the client machine isent usefull,
the most ppl has routers or firewalls that doesent pong. there then i 
tried to read not the whole line just each bit, worked fine until i 
sended a 20 kb xml document,
guess what the server died on just parsing the 20 kb string(bit by bit), 
that was the reason for the dedicated sourcecode question if the MINA 
would try to read
the message needing a %0A as termination delimter it will stuck on 
reading or never finsih parsing ;).


but i have also two more questions, one regarding the xml-pull-parser, 
in the lib of the red 5 server there is also the XERCES
project xml-parser, why should i use the XML-Pull-Parser instead, it 
looks much easier to handle, is that the reason?

and the second question would be about MINA again, if you would like to 
cluster the red5 server,
do you see a big difference in the possible bandwith if you
1) register it as a client to a webapplication and make a protocol 
handshake to veryfe its a server
2) build that from the red5 core using a total different port to bind 
both servers

>> anyway, it would be nice if u could tell me more about the plans of red5
>> with this xml socket and ACEGI.
>>
>>     
>
> We are still in the process of defining this stuff, so any feedback
> about what you would like to see is very helpful.
>   
uhm well in case of i will build the applications from the red5 source 
there wouldent be something that i would wish,
if i would like to use the scripting language that would be build in 
into red5, i would have several wishes what will make
a flash only developer the work much easier.

1) one security zone in ACEGI for each webapplication where the zone iss 
full configureable by the buildin scripting language
or even thru an adapter in java
the reason for that its just much easier to use a serversided 
usermanagment instead of accessing a external JSP or PHP based.

2) serversideded presitent secure data storage option ((XML)database or 
hibernate, just for maybe profile datas, small xml documents etc.)
may sharedObejcts will work for that but there should be permissions 
that them could only be accessble through the web
application not the client itselfs

those two thing would be very nice gimmics as basic adapters to build 
web applications on it that wheter needs a different php/jsp backend 
and/or database
and would be suitable for the most small webapplications,it would also 
lowers the amount of configuration work on installing and later on 
developing with red 5
such applications and also the later adminsitration, that all lowers the 
costs very much from the installtion untill the adminstration in 
productive phase, and gives
one more argument to use red5,

but i think you should first focus on the maingoals of red 5, its still 
a long way to version 1.0, you are all doing a great job go on that way
later maybe there is time to dream, and build in the gimmics;).

thanks again for the reply

greetz Sascha

_______________________________________________
Red5 mailing list
[email protected]
http://osflash.org/mailman/listinfo/red5_osflash.org

Reply via email to