[Devel] Re: [Users] Call-specific AVP DBS storage for OpenSER

[Bogdan-Andrei Iancu]

Hi,

JF wrote:

>>>But if an AVP is stored in DB, it lives beyond the transaction
>>>lifetime and it never gets deleted unless avp_db_delete() is called
>>>for it from the script, right?
>>>
>>>
>>>      
>>>
>>I see the DB as an external storage media. It's a special type of
>>location - a bulk storage.
>>
>>    
>>
>>>Isn't this a kind of global AVP?
>>>
>>>
>>>      
>>>
>>Indeed, you can load or delete from DB whenever you want, but sounds
>>like global C variables via DB :D. It's not something that should be done,
>>
>>    
>>
>
>On the other hand, you would be able to manipulate global proxy
>behaviour without having to do a restart. But I guess there are better
>ways to achieve this, e.g. FIFO or unix sockets.
>
>  
>
indeed - is what I'm thinking of - define (via script) or read (from DB) 
the global AVPs at startup into memory, use them all the time from mem 
and if you want to changes them use FIFO or unix socket.

>>>Also, if dialog IDs are used for AVP identification (UUID), these AVPs
>>>can be considered "Call-specific".
>>>
>>>
>>>      
>>>
>>yes, but the main and most important idea is to keep the AVP into memory
>>(as variables) and avoid DB operations.
>>
>>    
>>
>>>A dialog tracking module would be able to provide this identification
>>>and also guarantee that AVPs are deleted from DB when dialogs are
>>>terminated (provided record routing is active). What do you think?
>>>
>>>
>>>      
>>>
>>if you have a dialog tracking module, why should you keep the AVP s into
>>memory attached to the dialog structure? it;s is more simpler and efficient?
>>    
>>
>
>Maybe you meant "shouldn't"... 
>
yes, right :)

>In that case, I agree it would be more
>efficient to store AVPs in memory along with the dlg structure, of
>course. But if you add some other requirements, such as failover
>handling, maybe using DB for this purpose starts to make sense, IMHO.
>  
>
in a failover scenario, you will also need to replicate the dialogs, so 
you can also replicate the AVPs linked to them via same mechanism.

regards,
bogdan

>>regards,
>>Bogdan
>>
>>    
>>
>>>JF
>>>
>>>On 1/18/06, Bogdan-Andrei Iancu <bogdan at voice-system.ro> wrote:
>>>
>>>
>>>      
>>>
>>>>Hi Joachim,
>>>>
>>>>Here are some thoughts regarding this issue:
>>>>
>>>>First we have to make a clear distinction between whom the AVP belong to
>>>>(an identification in order to find it 0 this is more an abstract
>>>>aspect) and where the AVP is kept:
>>>>
>>>>   - identification: there are currently two ways of identifying an avp
>>>>- either as belonging to a user (user and domain), either via an uid
>>>>(unique ID). What is the meaning of the uid, it's a matter of
>>>>interpretation for the user - it may be a domain ID, a dialog ID
>>>>...whatever... I thing this identification mechanism gives full liberty.
>>>>Actually identifying ans AVP based on user at domain is a special case of a
>>>>UID, where the UID is the a sip uri (this just an simplification for one
>>>>of the most used case).
>>>>
>>>>   - where the AVP id kept - currently they are kept into transactions.
>>>>So they are accessible only their transaction is currently is progress
>>>>and their lifetime is the same as the transaction's one.
>>>>
>>>>Now...about improvements.....For identifications....I would say it
>>>>covers all cases..no idea what can be added - any suggestions are welcome.
>>>>
>>>>regarding the AVP location..this is a different story - I was thinkings
>>>>about adding new support: AVP per script (the AVP belong to the script -
>>>>kind of script variables and may be used for temporary ops) and global
>>>>AVP ( present in all the time, in all places - no restrictions; may be
>>>>used to control the overall behaviour of the proxy). And maybe
>>>>(again...maybe) in the future AVP per dialog (they belong to a dialog).
>>>>
>>>>Hope I manage to bring some light ;)
>>>>
>>>>regards,
>>>>bogdan
>>>>
>>>>
>>>>Joachim Fabini wrote:
>>>>
>>>>
>>>>
>>>>        
>>>>
>>>>>Hi all,
>>>>>
>>>>>Unfortunately the previous thread on this topic ended
>>>>>without a definitive statement or conclusion, probably
>>>>>because of the New Year holiday (btw, Happy New Year to
>>>>>all of you!). So here's another try:
>>>>>
>>>>>OpenSER is currently session stateful but not call
>>>>>stateful. This is valid also/especially for the AVP
>>>>>module: You can store AVPs on a per-contact base but
>>>>>(imho) _not_ specific to a call or dialog. The reason
>>>>>for this is that the AVP module does not allow/
>>>>>implement the use of call-dependent variables as AVP
>>>>>names.
>>>>>
>>>>>Adding this feature to OpenSER doesn't require a change
>>>>>in interfaces (implementation effort is another topic ;).
>>>>>It's just about enabling the use of AVP or header values
>>>>>(like, e.g. call-id) as the _name_ of another AVP for
>>>>>DBS storage.
>>>>>I attach a specific example below.
>>>>>
>>>>>Can some of the developers and/or maintainers please
>>>>>comment and/or give some hints on the feasibility of
>>>>>this approach?
>>>>>Is it possible to add this feature to OpenSER? What
>>>>>effort is required to add this enhancement to the AVP
>>>>>module? Please come back if some details are unclear,
>>>>>I might have missed something important.
>>>>>
>>>>>Many thanks in advance for your reply and help,
>>>>>Best regards
>>>>>--Joachim
>>>>>
>>>>>PS: Example: The task is to process any SIP message that
>>>>>passes through our proxy. We must compute a unique ID
>>>>>specific to the call on the first message belonging to
>>>>>this call and store it in DBS. Whenever another message
>>>>>that is part of this call/dialog arrives, we retrieve the
>>>>>unique ID from the DBS and, e.g., append it as a specific
>>>>>header field value.
>>>>>The only way how this can be done is to use some call-
>>>>>specific AVP name (e.g. call-id), as we do not have any
>>>>>influence on the structure of the AVP value (so regexp-
>>>>>search based on AVP value does not help).
>>>>>
>>>>># This is the AVP name that is required to retrieve
>>>>># our specific, call-dependent AVP value from the
>>>>># database whenever a message that belongs to this
>>>>># call passes our OpenSER proxy.
>>>>>avp_printf("s:myTmpHeaderName", "$from/username$hdr(call-id)");
>>>>>
>>>>># This is the value we'd like to store. Could be also
>>>>># only "$Ts", i.e., we can _not_ rely on regular
>>>>># expression search for later retrieving the AVP
>>>>># from the DBS based on its value.
>>>>>avp_printf("s:myTmpHeaderValue", "$hdr(call-id)-$Ts");
>>>>>
>>>>># This storing of the call-dependent info in DBS does
>>>>># currently not work. OpenSER complains at start-up
>>>>># about the variable AVP name/key on both of the
>>>>># following lines. The former variant seems much more
>>>>># flexible to me than the latter one.
>>>>>avp_db_store("s:myTmpHeaderName","s:myTmpHeaderValue");
>>>>>avp_db_store("$hdr(call-id)","s:myTmpHeaderValue");
>>>>>
>>>>># This load obviously does also not work.
>>>>>avp_db_load("$hdr(call-id)","s:myTmpHeaderValue");
>>>>>
>>>>>
>>>>>
>>>>>_______________________________________________
>>>>>Users mailing list
>>>>>Users at openser.org
>>>>>http://openser.org/cgi-bin/mailman/listinfo/users
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>          
>>>>>
>>>>_______________________________________________
>>>>Devel mailing list
>>>>Devel at openser.org
>>>>http://openser.org/cgi-bin/mailman/listinfo/devel
>>>>
>>>>
>>>>
>>>>        
>>>>
>>>
>>>      
>>>
>>    
>>