[VoiceOps] Bad/not unique SIP Call-ID GUIDs
abalashov at evaristesys.com
Tue Aug 3 08:45:31 EDT 2010
On 08/02/2010 12:56 PM, Scott Berkman wrote:
> Not something I'd ever come across, but the best solution I would think
> would be to create an auto-increment key that is unique (based on
> incrementing vales) within the database.
Well, yes, unique primary keys are great. :-)
That's what I'd ordinarily do with a Call-ID in many situations, and
that's what's causing the problem; when there is a primary key
collision, the entire transaction is rolled back.
The problem, which I mentioned but perhaps underemphasised, is that I
need a value that is not only unique but occurs straightforwardly within
the SIP message body itself, so that it is possible to easily associate
messages at a higher level of abstraction. That's the purpose that a
Call-ID is intended to serve.
> You can also look into what system it is that is producing the non-unique
> call IDs and try to work with that Vendor. In theory, non-unique call ID's
> could really mess up the state of a UA. Tags are normally meant to help
> combat that, but this kind of thing is where SIP implementations (and the
> underlying interpretations of the RFC's) really vary.
Tags are dialog-bound identifiers. The beauty of the Call-ID is that it
unites a superset of messages, including many sequential interactions
which do not form a dialog.
Alex Balashov - Principal
Evariste Systems LLC
1170 Peachtree Street
12th Floor, Suite 1200
Atlanta, GA 30309
More information about the VoiceOps