RE: ADABAS vs IMS vs DB2 Who is faster?
- From: I_Mugzach@xxxxxxxxxxxx (Itschak Mugzach)
- Date: 9 Oct 2007 12:52:04 -0700
Boys, please relax.
I wonder why TPC-C tests are only done for the distributed platforms
databases (or their DP versions). Within its members are IBM, Oracle (MF
version) Sybase(the same) and others. Maybe, because TPC heard (too early, I
must say) that the mainframe is dead?
Itschak
-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@xxxxxxxxxxxx
Of Jones, Kelly (Indust, PTL)
Sent: Tuesday, October 09, 2007 9:29 PM
To: IBM-MAIN@xxxxxxxxxxx
Subject: Re: ADABAS vs IMS vs DB2 Who is faster?
Hello Mike,
Most of the restrictions in the classes you mentioned disappeared from
ADABAS years ago. In our shop we yo-yo the ADABAS databases weekly - though
I know many shops that yo-yo the database only when an IPL is required. We
run backups every day - this can be done on only those blocks that have
changed, only certain files, while in use ...
Production changes can be implemented on the fly - new fields, new indexes
... can be made while the file being modified is in use.
ADABAS fully supports a sysplex and can actually scale quite well. I have
practically no DB2 or IMS experience so am not able to compare.
An optional optimizing compiler can be purchased for Natural. I've seen the
code it generates (it's good) and benchmarked it against optimized COBOL -
Natural won every test, from highly computational, to highly i/o bound. I
don't know what troubles you were having but things have changed a lot.
ADABAS does not support SQL but an interface can be purchased from Software
AG or a couple other vendors.
Best Regards
Kelly Jones
Penske Logistics
-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@xxxxxxxxxxxx
Of Mike Bell
Sent: Tuesday, October 09, 2007 2:24 PM
To: IBM-MAIN@xxxxxxxxxxx
Subject: Re: ADABAS vs IMS vs DB2 Who is faster?
I supported ADABAS for 2 years about 20 years ago. What I remember was the
SVC that modified itself and still running with 24 bit addressability.
Don't get me wrong - I could write amazing programs in Natural in just a
couple of hours but the other restrictions were painfull. Went on an
interview for a DB2 job and discovered they still had ADABAS - asked how
they managed it - basically they created a new ADABAS collection for every
application area that wanted it. You have to backup the entire database -
can't select one application and create a backup point.
Started working as IMS DBA with 1.1.5 - I think - been applications, system
DBA and sysprog. Never did past path but all combinations of HDAM, HIDAM,
SHISAM, etc It is very difficult to beat the performance of a properly
designed HDAM application unless the users require a half dozen secondary
indexes. Problem was the 2G/4G dataset limit - IBM didn't release
partitioned database until V9. Yes, you could move segments to a separate
dataset (max of 10) but it created performance problems as often as solving
them. People did all kinds of unusual design work to support more than 4G of
data. One place had a dozen PCB's that were selected by application based
on key range.
first supported DB2 as sysprog on 1.2. My mistake of being in the IMS group
and saying it isn't that hard - just another database -setup 1,2,3 write a
userid exit and let the applications groups go. SQL is different. It
requires a different type of design. I think it is harder to build a good
SQL design than IMS because you have so many choices and the SQL hides the
performance impact of those choices.The ability to use big bufferpools can
also hide bad design. It can still meet response time goals but use more
resources than it should. The other part is that DB2 started with a 64G
limit and has expanded that multiple times.
ADABAS does use less memory and disk space than DB2 but it doesn't scale to
really large applications the way DB2 does and it doesn't let you keep the
database up while some portion of the database is having utilities run.
Mike
On 10/7/07, Itschak Mugzach <I_Mugzach@xxxxxxxxxxxxxxxx> wrote:
this true?
I spoke few days ago with an ADABAS specialist that claimed that
ADABAS is much faster and has low overhead compared to IMS and DB2. Is
Itschak
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions, send
email to listserv@xxxxxxxxxxx with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
--
Mike
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions, send email
to listserv@xxxxxxxxxxx with the message: GET IBM-MAIN INFO Search the
archives at http://bama.ua.edu/archives/ibm-main.html
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions, send email
to listserv@xxxxxxxxxxx with the message: GET IBM-MAIN INFO Search the
archives at http://bama.ua.edu/archives/ibm-main.html
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to listserv@xxxxxxxxxxx with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
.
- References:
- RE: ADABAS vs IMS vs DB2 Who is faster?
- From: Jones, Kelly , Indust, PTL
- RE: ADABAS vs IMS vs DB2 Who is faster?
- Prev by Date: Re: SLIP sliding away
- Next by Date: RE: SLIP sliding away
- Previous by thread: RE: ADABAS vs IMS vs DB2 Who is faster?
- Next by thread: Re: ADABAS vs IMS vs DB2 Who is faster?
- Index(es):
Relevant Pages
|
