Environment AIX server v2.1.5.13, HP client v3.1.0.3

Recently, Unix script to generate ADSM archive command to archive Oracle
archived redo logs whenever there is a log generated, sample command:
dsmc archive "/all_arch/dba/oracle/mmsprd/arch/archlog93.arc"
-desc="mmsprd--archlog93.arc" -virtualn=xxxx -password=xxx
When I do a dsmc q archive "/all_arch/*" , there are entry for each level
When I do a dsmc q archive "/all_arch/*" , there are entry for each level
of subdirectory of each archived entry. Eg
q archive "/all_arch/*" -subdir=yes -desc="mmsprd--archlog93.arc"


             Size  Archive Date - Time    File - Expires on - Description

             ----  -------------------    -------------------------------

            1,024  09/17/1998 18:08:43    /all_arch/dba 12/19/1998
mmsprd--archlog93.arc

            1,024  09/17/1998 18:08:43    /all_arch/dba/oracle 12/19/1998
mmsprd--archlog93.arc

               96  09/17/1998 18:08:43    /all_arch/dba/oracle/mmsprd
12/19/1998 mmsprd--archlog93.arc

               96  09/17/1998 18:08:43    /all_arch/dba/oracle/mmsprd/arch
12/19/1998 mmsprd--archlog93.arc

      334,496,768  09/17/1998 18:12:20
/all_arch/dba/oracle/mmsprd/arch/archlog93.arc 10/18/1998
mmsprd--archlog93.arc

I can remember that for archive in v2 client, there will be only one entry.
To size ADSM db, 400 to 600 bytes for each entry; so should I size the db
by 5 entries or 1 entry for each archived file for v2 client & v3 client?

Today I have implemented this 'auto adsm archive' for one client with
several db in it for 1 month, and there are ~4000 archived files; so should
I use 4000 or 20000 for sizing? This piece of info is important to me as
this "auto adsm archive" will be rolled out for more Unix clients (some v2,
some v3).

Your help is greatly appreciated.

Regards,
Eric Tang / Hong Kong / IBM  @ IBMHK
Technical Services, Strategic Outsourcing,
IBM Greater China Group
Tel: (852)-2105-2119,   Fax: (852)-2861-2493
Notes: Eric KF Tang/Hong Kong/IBM@IBMHK
Internet : tangeric@ hk1.ibm.com