While You Were Sleeping…

Granted, your company’s batch processing may be running smoothly while you sleep. Or the interruptions may be at an acceptable level not to keep you up at night. But there may be some situations that warrant a review of your batch processing – like a company merger where the systems of the amalgamating companies have to ‘talk to each other’ OR like a major purchase of a COTS that has to be integrated with your existing legacy batch process. In either case, fusing two batch processes can have a compounding effect on the Batch Window which may adversely affect your Batch SLA.

3 Improvement Areas

In one of the companies I was employed at, I found and worked on resolving 3 issues that achieved significant improvement to the batch window. Looking at these 3 areas may also assist you in your batch tuning efforts.

1. Obsolete jobs that are still running

2. Incorrect job dependencies resulting in job streams (JS) running later than they should thus affecting downstream jobs (see sketch below)

3. File transfer job parameters are incorrect resulting in slow file transfer times and delay in subsequent jobs

My Sketch of a Simple Job Dependency Scenario

The Big, Black Box of Batch

Solving these 3 issues sounds simple enough. The challenge is, when you start to put a magnifying glass to the big, black box called ‘batch’, it is like opening Pandora’s box – you usually have thousands of jobs to look at, incomplete or even non-existent documentation and you must be very cautious when implementing the solutions (especially when you do not have sophisticated monitoring tools) as even a small mistake can cause the process to go haywire. This is especially probable of issue # 1 & 2 above. Some of the things to take note of:

After creating a list of batch jobs, you need to identify their logical classifications (e.g. function, frequency, criticality, etc.) but for some jobs, you cannot determine the function or criticality or even necessity and there is no documentation to reference plus the person who can already left the company years ago. If in doubt, it is safer not to alter the job as changing it could have unexpected results.

You do not have a testing environment that closely mirrors the production environment so the changes you made in the testing environment may behave differently in the production environment. Again, caution is key. If you are lucky enough to have sophisticated tools, then you have more confidence putting the changes into effect.

Post-Op

After your batch tuning exercise, it is also important to communicate the actions taken and what Software Engineers or whoever creates and supports the batch jobs, need to do to ensure the inefficiencies do not reoccur. The ideal scenario of course if your resource allocation allows, is to have a batch management team or staff in place to proactively monitor batch performance, at least maintain a repository of information on the different batch jobs and give advice on the best insertion point for a new job. The other way is to conduct regular batch ‘spring cleaning’ exercises.

Your batch support staff will thank you for a good night’s sleep.

To read more about Batch Performance, visit http://www.cmg.org/ for white papers.

Related Posts:

The Synthesizing Mind – Batch optimization can be an exercise in synthesis and patience! Read this post to be introduced to a book that can encourage you. You are not alone.

——————–

Posted by rochelleolviga

Advertisements
  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: