Re: [Nolug] Introducing myself - SMPP client

On 12/07/08 08:39, Petri Laihonen wrote:
> Ron Johnson wrote:
>> On 12/06/08 22:43, Petri Laihonen wrote:
>> [snip]
>>> Another reason was that the programs need to be able to multithread and
>>> be wicked fast.....
>> Threaded programming can be really hazardous, because:
>>
>> a) in Windows (and sometimes in Unix) they all run in the same address
>> space, and an errant pointer in thread A can stomp on something in
>> thread B, and
>>
>> b) you need to be very careful to avoid race conditions.
>
> There is a remote possibility that I may have misunderstood the meaning
> of multithreading, but the one below is one of the main functions of an
> SMPP client.
> - The ability to establish 2-way communications via socket.
> - There are 2 modes, BTX/BTR (uses 2 sockets, one receives the other one
> transmits, however both do acks, thus in real life must be somewhat
> 2-way. BTRX handles both directions via one socket.
> - Amount of connections via one client program is normally one to two,
> but could be n. (so far highest we had has been 10 BTX/BTR connections)
> - Communications are most often synchronous, however more and more, the
> SMSCs we would connect to, can be capable of asynchronous transactions,
> thus needed as well to increase efficiency.
> - naturally also all the activity according to SMPP 4.2, 4.3, and 5.0
> standards.
> - Some additional goodies for manipulating the messages. (Currently we
> have our own scripting language for in and out traffic.)
>
>
> Therefore, does the above require multithreading or not?
> Any suggestions of what programming language would be the most suitable one?

You might be able to accomplish that with the poll() function, or
two processes talking via IPC.

A Python network guru would now whether something like the Twisted
framework would be suitable.

-- 
Ron Johnson, Jr.
Jefferson LA  USA
How does being physically handicapped make me Differently-Abled?
What different abilities do I have?
___________________
Nolug mailing list
nolug@nolug.org