it ups to the handler you call with exec, the system resources you posses ...
in case you dont experience high loads exec is acceptable as short term solution
otherwise, use sqlbox, or, what about XML POST ?:)
i do parallel processing of MO traffic with invoking php script from kernel
inotify.h
in short when xml file is being moved to a directory, i spawn php script like
this one
/usr/bin/nohup ${php} ${loopScript} ${key} PAYMENT ${filename} &
so far i have no issues, but for this particular service the max load is like
40 MO messages in same time, which is nothing.
i measure the load level before the each new process start ...
in case you have high load peaks,
1. make sure you know what will happen when the execution of the php scrip
fails.
2. make sure you control the level of the box load ... do not let kannel crash
the box :)
with processes the box cant handle.
last thing - do not let these people change your way of work ...
well its your call anyway :)
cheers
David wrote:
> Thanks!
> Does anybody know if EXEC can handle large loads if we spawn new
> processes (with the & at the end of the command?) this seems to be our
> short term solution
>
> thanks
>
>
> --- On *Thu, 2/12/09, Falko Ziemann /<[email protected]>/* wrote:
>
> From: Falko Ziemann <[email protected]>
> Subject: Re: HELP! Patent claim! Can we use an alternate protocol?
> To: [email protected]
> Cc: "seikath" <[email protected]>, "[email protected] User"
> <[email protected]>
> Date: Thursday, February 12, 2009, 5:02 PM
>
> Uhh, found a even better one here:
>
> http://www.cellitmarketing.com/blog/the-tcs-patent-dispute-cellits-viewpoint/
>
> "As many of you know, TeleCommunication Systems, Inc. is currently
> pursuing legal actions against many players in the mobile space."
> (bla bla)
> "It is Cellit’s belief this patent claim is without merit.
>
> Freely available software, namely Kannel (available at Kannel.org),
> enables this exact type of interaction (conversion of MOs into web
> requests) and has been available since 1999. While I am not a
> lawyer, it is my belief the existence of Kannel in 1999 constitutes
> “prior art” and thus nullifies TCS’s 2000 and 2005 patent claims."
>
>
> Am 12.02.2009 um 23:56 schrieb Falko Ziemann:
>
>> http://findarticles.com/p/articles/mi_pwwi/is_200804/ai_n25346966
>>
>> "The patent describes a means for triggering an Internet
>> informational query or search using a simple text message
>> originated from a cell phone or mobile device, and is widely used
>> today for two-way premium messaging services."
>>
>> Well, I didn't read the whole patent, but this really sounds like
>> they should sue Google mobile (I would really love to see this),
>> not you...
>>
>> Regards
>> Falko
>>
>> Am 12.02.2009 um 23:31 schrieb seikath:
>>
>>> Hi David,
>>>
>>> Falko gave the solution - use sqlbox for internal communication.
>>>
>>> Anyway, I'd love to know more of this like patent ID etc.
>>> In short, I have my doubts that they will claim violation.
>>> In short you do use http as internal communication INSIDE your boxes.
>>> HTTP is NOT used to receive MO traffic from mobile devices.
>>> The SMPP/OtherProtocols are used to communicate with mobile
>>> operators SMSC.
>>> the SMPP protocol DOES not send short text messages at all ..:)
>>>
>>> These people claims sound not serious to me at all.
>>> So relax a bit :)
>>> hint: the exec module has its issues with escaping non usual
>>> characters...
>>>
>>> I could be wrong, but they have to prove something non even
>>> existing ...
>>>
>>> cheers
>>>
>>> David wrote:
>>>> Hello,
>>>>
>>>> Us users of SMS Shortcodes are getting hit in the US for violating a
>>>> patent by TCS Inc. The patent claims to cover:
>>>> "A gateway, comprising: a first communication path to accept a short
>>>> message from a mobile device; a translation module to insert
>>>> said short
>>>> message into an Hypertext Transfer Protocol (HTTP) message; a second
>>>> communication path to push said HTTP message to at least one
>>>> Universal
>>>> Resource Locator (URL); and a return communication path to receive a
>>>> return message relating to said HTTP message."
>>>>
>>>> The patent goes on to mention a whole bunch of stuff, ie, using HTTP
>>>> POST to send the http message, etc.
>>>>
>>>> I just got out of a meeting with my lawyer. He says, the only
>>>> way to
>>>> get around this is to NOT use HTTP.
>>>>
>>>> Over the weekend, I tried to use the sms-service using EXEC
>>>> instead of
>>>> GET-URL, which worked fine... and then the server crashed... i'm
>>>> guessing too many spawned processes?
>>>>
>>>> My question to you all is: how can we NOT use HTTP but have the
>>>> same
>>>> end result? ideally, it would use a command-line exec, but push
>>>> it into
>>>> a waiting server, instead of spawning a new thread. Does
>>>> anybody have
>>>> any ideas on how to do this?
>>>>
>>>> Thank you very much,
>>>>
>>>> David
>>>>
>>>>
>>>
>>
>
>
