ESP8266 Developer Zone

Re: new features possible ?

2015-01-05T23:32:41+08:00

I am a newbie to this group. This is my first posting. I really wanted to start a new topic but didn't know how.

We have an application where we want to use the ESP8266 as a local service provider. This is actually an interface between a Smartphone app and a system running on a CAN bus. In common with most (all?) local services, from domestic appliances, local printers, door openers, model control etc. the service provider needs to implement some sort of service discovery for the mobile phone app. The apps we use, on both iOS and Android devices, implement Bonjour / Zeroconf to connect to the service, in this case our interface. My ESP8266 already does DHCP and connects to the phone OK. However, I haven't been able to go to the next stage.

The local service provider needs to listen on 224.0.0.251, port 5353. There is a standard and simple message exchange to set up the client / server link (in TCP/IP). It needs the ESP8266 to know the name of the service (_xxxx.local). As this service discovery (SD) is such a common situation and will apply to virtually any IoT capable device, I can't envisage the ESP8266 being designed without it but have seen no reference to it so far anywhere.

Is the present ESP8266 capable of providing SD (Bonjour / Zeroconf) ? Is so, how is it implemented? If the ESP8266 doesn't have this capability, then why not, as it is such an obvious requirement? If the ESP8266 is able to listen on the IP for DHCP, can it also be set to listen to another multicast IP address and port - as well as working on its own TCP/IP server address. In which case the host processor could implement the SD.

Surely others have had this situation. Any help or suggestions welcome.

Mike Bolton

Statistics: Posted by MikeBolton — Mon Jan 05, 2015 11:32 pm

Re: new features possible ?

2015-01-03T06:27:37+08:00

Hey，
1.Thanks for your advice , we have added a data buffer, and we are trying to fix some bugs and adjust the logic struct now.
After that , the AT project would be far more powerful and stable.
2.Good idea for the request FIFO. Do you mean successive tcp request to the same port? We should think about the data buffer size.

1/that would be good, as for now, it seem not really usable / reliable as real web server.
the best i do is to serve an HTML page, a CSS and 4 images, for a total of about 60KB, in about 1 minute !!
if i decrease some long delay, i've got a lot of troubles as "busy" or others unsolicited messages !
there is not enough information to know how manage the different unsolicited messages.. please help !

2/i mean successive TCP requests when a web browser request an HTML page (HTML, CSS, script.. images.. )
the host CPU receive the "CONNECT" and '+IPD" messages during the HTML page is send.
i manage it, but it's not the best way, not really stable !
that may be simplify the host CPU job, if all request are FIFOed in the 8266...
each time the host CPU closed a connection, the next request is send to the host CPU .. until end of requests ..

thanks, regards
Phil

Statistics: Posted by phil31 — Sat Jan 03, 2015 6:27 am

Re: new features possible ?

2015-01-02T23:53:29+08:00

phil31 wrote:
Please here it is 2 suggestions / question

1/ I would like to know if the data buffer size of the AT+CPISEND command, actually 2048bytes, can be increased ?
if yes, why not do it !?

2/ in case of TCP server mode with multiple connection, is it possible to store the successive HTTP requests in a FIFO queue then send them to the host, one after one when last is ended ?
this may be singularly simplify the host job !

thanks for any comments !
regards

Hey，
1.Thanks for your advice , we have added a data buffer, and we are trying to fix some bugs and adjust the logic struct now.
After that , the AT project would be far more powerful and stable.
2.Good idea for the request FIFO. Do you mean successive tcp request to the same port? We should think about the data buffer size.

Statistics: Posted by costaud — Fri Jan 02, 2015 11:53 pm

new features possible ?

2014-12-30T00:07:28+08:00

Please here it is 2 suggestions / question

1/ I would like to know if the data buffer size of the AT+CPISEND command, actually 2048bytes, can be increased ?
if yes, why not do it !?

2/ in case of TCP server mode with multiple connection, is it possible to store the successive HTTP requests in a FIFO queue then send them to the host, one after one when last is ended ?
this may be singularly simplify the host job !

thanks for any comments !
regards

Statistics: Posted by phil31 — Tue Dec 30, 2014 12:07 am