I recently had the opportunity to interview Sridhar Sudarsan, Chief Architect, and IBM Batch processing strategy.  He has architectural and strategy responsibility for batch processing capabilities across IBM’s products and platforms.    

Sridhar, thank you for agreeing to talk with us about batch. Let’s dive right in--Why Batch?

Batch processing has actually been around since the beginning of electronic computing. It’s even more important now as the need to cost-effectively process large quantities of data becomes more challenging as data volumes increase exponentially and customer expectations continue to rise. In today’s world, the only way to achieve business agility and efficiency is by designing for and providing an always-on environment for batch and OLTP workloads. The notion of batch processing is evolving as you look at running “elastic batch” windows.

Where does Java fit in? Can it be used for batch or is ”Java batch” an Oxymoron?

Java is actually growing in popularity as a programming language for batch use. There are a number of factors driving this trend. For example, batch windows are shrinking and there’s a business need to reuse business and application logic, code artifacts, system administration and operations across batch and OLTP workloads. If you run new batch applications in Java on the mainframe, there’s cost savings to be gained by offloading to zAAP engines on z/OS.

In addition, the concept of ‘batch as a service’ is emerging, and Java batch front-ended with web services is gaining ground. Java batch also allows companies to leverage in-house skills across a broader set of business requirements.

What are some of the approaches to Java batch? 

There are actually a couple of broad approaches used to engineer batch processing with Java and each has its pros and cons.

One approach is custom construction. This approach involves using custom code to tie JVM launch utilities and programming frameworks together. In the short term, this may seem like a cost effective way to extend custom functionality to meet tactical needs, but over time it becomes costly to maintain a suite of custom middleware.

Another approach is an integrated managed environment. This approach includes software that integrates batch processing in to a full function Java EE execution platform. It provides a JVM environment and a development framework, as well as a managed container environment for batch applications. This provides the batch infrastructure and allows for separation of concerns from the application code.

What role does WebSphere Application Server play?

WebSphere Application Server (WAS), IBM’s Java EE runtime environment, provides application services within a managed container environment. IBM offers WAS on hardware and operating system platforms ranging from smaller x86 all the way up to the mainframe running z/OS, providing open standard support in all cases.

Can you talk more about IBM’s Java batch offerings?

IBM has two batch offerings—IBM WebSphere Compute Grid and WebSphere Application Server Feature Pack for Modern Batch.

WebSphere Compute Grid is an enterprise strength Java batch execution platform. It runs on the WebSphere Application Server, and provides functions like the batch container, batch toolkit, parallel job management, advanced operations support and integration with enterprise schedulers.  WebSphere Compute Grid is supported on distributed and the z/OS platforms.  WebSphere Compute Grid for z/OS exploits additional functions of the z/OS platform such as WLM, RRS and SMF.

The Feature Pack for Modern Batch is an optional function that can be added to an existing WAS runtime environment. It provides a way to get started to run batch applications on WebSphere servers.  It provides a subset of the Compute Grid functionality – batch container, toolkit and the job dispatcher.  The feature pack is available at no charge to customers with Software Subscription and Support.   

Do you have any other comments you wish to add?

As I mentioned earlier, there are more and more applications that need to be written in Java batch.  You want to be careful about not getting carried away and be in the business of writing middleware.  It is best to take a decision early to write applications using the batch platform like WebSphere batch, so you can focus on the business of your applications. 

You can get started with batch tooling on RAD v8.0.1 to build your applications and run them on either of the WebSphere batch solutions.

Sridhar, thank you so much for taking the time to talk with us today. We look forward to hearing more about Java batch solutions in upcoming conversations.

More information about Modern Java Batch

WebSphere Batch Solutions: Join our Lead Architects as they review IBM's product strategy for WebSphere Compute Grid and the WebSphere Batch Feature Pack -- our comprehensive, enterprise-class, Java-enabled batch processing platform.

Register for Part 1 of the 2 Part series.

Register for Part 2 of the 2 Part series.

Install WebSphere Application Server V7 Feature Pack for Modern Batch. Click here.

Read the white paper: Java-enabled batch for cost-optimized, efficient business processing. Click here.

Download the Redbook: batch Processing with WebSphere Compute Grid. Click here.

Follow us on Twitter @Kmholm #websphere, #computegrid, #smartwork