Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
11/15/06 07:59
Read: times


 
#128007 - Right way to design/develop a protocol
Hi,
i post my question here because the topic is not directy 8051 related.First of all sorry for the long post but it's necessary to correctly explain correctly the problem!

I'm developing a project using a Silicon labs C8051F340 chip, the device will be a sort of datalogger and is supposed to do the following things:
- Collect radio (868 hz) data pockets from many remote sensors.
- Store them in a flash memory in form of records. The buffer have a fixed lenght and after is full, the new record will overwrite the oldest.
- Store a alarm-log in the flash memory.
- handle a link with GPRS modem and UDP protocol (the UDP stack is implemented by the modem)to a remote server for downloading all the data.
- handle e local link with a PC for setup diagnostics etc.

The main question is how to correct design the communication protocol between device and PC/remote server. In the past i done small project and developed simple protocols.

My idea is that the device must be "slave" and the responsability for sending the commands must be of the PC program.
For example the sequence should be :
- Pc "call" the device.
- They autenticate.
- PC call for any alarm present, if present download alarm log.
- PC call for new record. If new records download them.
- PC close the session.

In few words there are many small and simple protocol command for (for example) reading record index, locate the record pointer at a particular location, read the record pointed etc.
The PC program must use them and create a sort of "script" doing the sequence it needs. In this way if you want to change sequence you "only" change the PC program and not the firmware in the device.

But my collegue (he's doing the PC program using delphi) don't agree because using event-driven controls it says it's difficult to realize a multiple step sequence.

Can you suggest something maybe some link/documentation to read?

Regards.

massimo



List of 59 messages in thread
TopicAuthorDate
Right way to design/develop a protocol            01/01/70 00:00      
   Many ways....            01/01/70 00:00      
      SOAP            01/01/70 00:00      
         SOAP            01/01/70 00:00      
      State Machine.            01/01/70 00:00      
      SOAP - since you mention it...            01/01/70 00:00      
   I think            01/01/70 00:00      
      more SOAP...            01/01/70 00:00      
         two things to look at            01/01/70 00:00      
            Turning the bus            01/01/70 00:00      
               IIC turning the bus            01/01/70 00:00      
                  Turning the bus            01/01/70 00:00      
   A very important caveat            01/01/70 00:00      
      gprs and IP number            01/01/70 00:00      
   RFC953            01/01/70 00:00      
      Protocols            01/01/70 00:00      
         PC State Machines            01/01/70 00:00      
            why not switch()            01/01/70 00:00      
               Style            01/01/70 00:00      
                  Huh?            01/01/70 00:00      
                     Function Pointers!            01/01/70 00:00      
                        Function pointers troublesome? How?            01/01/70 00:00      
                           Overlay            01/01/70 00:00      
                              Table of function pointers            01/01/70 00:00      
                                 Experience helps            01/01/70 00:00      
                                    Up the Tree            01/01/70 00:00      
                                    No question at all!            01/01/70 00:00      
                                       Last word            01/01/70 00:00      
                                          Good luck!            01/01/70 00:00      
                                             Oh, he can            01/01/70 00:00      
                                                Living with the boss            01/01/70 00:00      
                                                   well, sometimes it does            01/01/70 00:00      
                                                      deja vu            01/01/70 00:00      
                                                         Take a hint from the Web            01/01/70 00:00      
                                                         Don't go into this menu            01/01/70 00:00      
                                          get another job            01/01/70 00:00      
                                             change job            01/01/70 00:00      
                                             All the news that's fit to print            01/01/70 00:00      
                                                I agree            01/01/70 00:00      
                              #pragma nooverlay in SDCC            01/01/70 00:00      
                              Update on function pointers            01/01/70 00:00      
                           are you using the awful LARGE model?            01/01/70 00:00      
                              Memory-model independant            01/01/70 00:00      
                                 not really            01/01/70 00:00      
                  No one size fits all            01/01/70 00:00      
   FTP?            01/01/70 00:00      
   relevatnt to your post, not thye subject            01/01/70 00:00      
      flash memory            01/01/70 00:00      
   Here's one.            01/01/70 00:00      
      Wrong thread sorry ^ ;)            01/01/70 00:00      
         Finite state machines            01/01/70 00:00      
            state machine            01/01/70 00:00      
   State Machines - AsyncPro?            01/01/70 00:00      
      Async Pro            01/01/70 00:00      
         Sounds like rubbish to me!            01/01/70 00:00      
         Need to hire some experience            01/01/70 00:00      
            Experience            01/01/70 00:00      
               Oh well            01/01/70 00:00      
            a very common mistake            01/01/70 00:00      

Back to Subject List