Re: IDS license count per user or per session?

The answer, as with so many other things, is: "It depends." Generally, unless you push, claw, scratch, and battle, the answer winds up being "whichever is greater." For MOST applications, the number of concurrent sessions is greater than the number of actual simultaneous users, so that's what Informix goes with. However, in our application, we have multiplexed connections, so we have a lot more actual users than database sessions. The licensing nightmare was enough of a pain, that we switched (or are switching) to CPU pricing. The reason you can't get a straight answer on this is because there isn't one. HTH, - TJG "Michael Krzepkowski" <NOmichaelkSPAM@sqlcanada.com> wrote in message news:3EF28DE0.7010908@sqlcanada.com... > John Keefer wrote: > > >The question has been asked before here on c.d.i., but it's elicited > >conflicting answers. Some have said the license count is per session, > >unless there's a Web server, transaction monitor, or similar gadget that > >multplexes user connections; in which case the count is per user. > > > >Others have said the license count is per user. I read several posts circa > >1995 stating that the count is per user, with up to five sessions allowed > >per user. > > > >Our reseller has always said it's per user. One of my coworkers called IBM > >the other day and was told the count is per session. But if that's the > >case, we could theoretically recode our apps for a Web interface and run > >Workgroup Edition; and that can't be right. Anyone know where to find an > >authoritative answer? > > > >Thanks in advance, > > > >John Keefer > > > > > > > > > > > > > Licence is per concurrent session, not per "named" user. > IBM document CT0ZUNA > Quote: > > Concurrent Session - A Concurrent Session, at a give point in time, is a > logical > connection, existing at that time, between the Program and an end user > interface > device. Each such existing connection is counted as a Concurrent Session > whether > or not it is actually in use. If an individual end user establishes multiple > connections, each one is conted as a separate Concurrent Session, even > if they > share a common physical pathway. The number of connections is not reduced by > any program or machine, such as a front-end server or multiplexer, which > may be > used to concentrate the connections. Each batch process being executed > at the > same time is also considered to be a concurrent user. > > End quote. > > HTH > > Michael > SQL Systems Inc. > (remove capitals from the address to e-mail me directly) >

> For MOST applications, the number of concurrent sessions is greater than the > number of actual simultaneous users, so that's what Informix goes with. The quote from the IBM document CT0ZUNA, posted by Michael Krzepkowski above, appears to support the "whichever is greater" interpretation. > The licensing nightmare was enough of a pain, that we switched (or are > switching) to CPU pricing. If the license count is per-session rather than per-user, then we too would be better off switching to CPU licensing. In our case, CPU pricing would be $100K to $120K more than what we currently have. Just now we're preparing budget recommendations for fiscal year '04, so I'm trying to determine whether we should budget for this. > The reason you can't get a straight answer on this is because there isn't > one. That certainly seems to have been true of the old Informix, but IBM historically has been clear about software license terms, and not shy in insisting on compliance. There may have been no straight answer pre-acquisition, but that may no longer be the case.