× The internal search function is temporarily non-functional. The current search engine is no longer viable and we are researching alternatives.
As a stop gap measure, we are using Google's custom search engine service.
If you know of an easy to use, open source, search engine ... please contact support@midrange.com.



Of course, Coglin Mill has an ETL product, RODIN. Good product, I believe.

On 6/12/2012 9:36 AM, AJordan@xxxxxxxxxxxxxx wrote:
My two pence:

Triggers are intrusive on the database and are a performance hit on the
production system. Also, "manipulation and joining to other files" means
more than just the simple processing a trigger program should be limited
to - again a definite performance issue. As an example, consider an order
header record. An order gets created in the system, and it's quite
possible that the record will get updated 5, 10 even 20 times before the
order processing is complete. That's up to 20 calls to the trigger program
with the additional I/Os to capture each of those changes (potentially
dozens more if you do any join processing in the trigger). Contrast that
to an end of day ETL approach - you only need just one read of the record
(remember there is no real time requirement here). Zero hit on the
production system during the day.

Triggers also do not satisfy the other requirement for a full refresh as
well as changed data, so you need to develop 2 different solutions if you
go the trigger route. Double the development effort, at least.

HA products are only a partial fit. Yes they can capture the changes etc,
but they don't have strong (if any) data manipulation, joining, error
management capabilities etc.

It seems to me that the decision to use an ETL tool is well grounded in
this instance. ETL satisfies all of your requirements whereas none of the
other suggestions come close.

Regarding the choice for a tool or roll your own... how much time and how
many resources do you have? Do you want it to be documented, easy to
maintain and to have extensive error management, auditing and metadata
capabilities. You will almost certainly not include any of that if you
grow your own. And depending on the complexity, it will still take you
anything from 3-10 times as long to develop.

And once you have an ETL tool it can be used for other future projects
besides this one.

Kind Regards
Alan Jordan

RODIN by Coglin Mill | 507.282.4151 x 103 | 507.261.4495 Mobile |
Rochester, Minnesota USA | www.thinkrodin.com




From: rob@xxxxxxxxx
To: Midrange Systems Technical Discussion<midrange-l@xxxxxxxxxxxx>,
Date: 06/12/2012 07:16 AM
Subject: Re: data Extract, Transform, Load (ETL) on Iseries
Sent by: midrange-l-bounces@xxxxxxxxxxxx



Reasons for ETL may include:
- Slows down the .NET developers to have to learn how to connect to a real

database.
- Faster response time for non volatile data.
- Access to data regardless of availability of machine due to backups,
month end, PTF apply, etc.
All able to be worked around but still not totally negligible concerns.
But that decision has been made.

Mimix does have an HA solution, true, but them and/or their parent
company, Vision Solutions, also has products for this kind of processing
also. There's been a few other companies also peddling products. I don't

use any.

Then there's roll your own. It's a Make vs Buy decision. Common enough
occurrence in IT. Won't be the first or, God willing, last time you'll
ever hit it.


Rob Berendt

As an Amazon Associate we earn from qualifying purchases.

This thread ...

Replies:

Follow On AppleNews
Return to Archive home page | Return to MIDRANGE.COM home page

This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].

Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.