RE: well written sql article -- MIDRANGE-L

Mike:

Not being sure what products you have loaded nor how each was licensed to you, 
I can only give a general response.

Use WRKLICINF to list all licenses on your system. Lock at details for each to 
see how many are *CONCURRENT or *REGISTERED to get an idea of how it's already 
being done (in OS/400).

Beyond that, in general, the simple presence or absence of a "license" (or the 
properties of a license) isn't the real controller. What's in control is the 
product program that calls the licensing API and selects how to continue based 
on what's returned from that API.

The programmer might keep count of the number of license requests from a user 
and store them by IP address for example. As long as all requests come from the 
same address, the program could simply take a code path that indicated a 
license was previously requested and granted, so skip calling the API. A 
difference in user/address could trigger a check to see if the same user was 
requesting from a different address. (Could; not necessarily would.)

Further, the OS/400 licensing stuff keeps track of various elements. For 
example, *CONCURRENT can keep track of how many licenses are currently granted. 
The license request API is called upon entering a licensed part of a product, 
and a license release API can be called when leaving. The system keeps track of 
how many are active. If it's limited by *NOMAX, it doesn't really matter. But 
if a limit is specified and exceeded, then aspects such as 'grace period' 
become active.

And *REGISTERED can be more particular. This allows assignment of license 
requests to specific userids (though they may also be released) by the 
licensing functions themselves.

Practically speaking, it's almost certainly possible to skip around various 
user-based licenses. Of course, it ~is~ stealing. Also, it's possible to 
program the product to detect it within some basic limits. Such programming 
might change with any PTF. Imagine a business that built its database access 
around cheating with a single license and after a PTF -- database access locks 
up except for the first connection. Who ya' gonna call? This is completely 
separate from explaining to auditors why all transactions are logged to a 
single userid.

In short, user-based licensing is common already and works pretty good. (Note 
-- this isn't intended as an authoritative description of how IBM LPPs check 
licenses. Just general thoughts around how I've seen some programming done over 
those APIs.)

Tom Liotta

midrange-l-request@xxxxxxxxxxxx wrote:

>   2. Re: well written sql article (Mike Eovino)
>
>Yes, but if I wanted to, I could make all of my connections with the
>same user profile.  I can't believe they'd let me get away with a
>single client license.
>
>On 4/12/06, qsrvbas@xxxxxxxxxxxx <qsrvbas@xxxxxxxxxxxx> wrote:
>> >
>> >How do they do user-based pricing for DB2 if I'm using JDBC/ODBC against it?
>> >
>>
>> Do you do that without logging into the database as a user?