Evangelizing Mainframe
Print Email

CICS and Cloud

Gabe Goldberg talks about sources of information in his Destination z article, “Use Career Resources to Grow.” One source of information for CICS specialists that hadn’t caught Gabe’s eye is the Virtual CICS user group. At the group’s January virtual meeting, IBM’s Matthew Webster was discussing “Five Compelling Reasons for Creating a CICS Cloud.”

Webster is an IBM Senior Technical Staff Member in the CICS Transaction Server (TS) for z/OS team, based in Hursley, U.K. He holds a Bachelor of Science in Physics with Computer Science from Southampton University, and has more than 25 years of experience as a software engineer at IBM, leading a wide range of projects from mainframe to open source. His areas of expertise include CICS, Java, Agile software development and cloud technologies. Webster is a regular blogger, podcaster and presenter with published articles and papers on CICS and software development.

Webster started his presentation by saying that cloud computing is one of the fastest-growing technologies, and more applications and services are being provided through the cloud to offer cost savings and scalability. CICS TS V5 delivers a pattern-based, policy-managed, private CICS cloud environment, improving the agility of CICS application deployments while reducing the risk of implementation errors.

Webster suggested that sites needed to create a CICS cloud because:

• Agile methodologies allow developers to increase the rate of change of applications in response to business needs
• IT operations need to respond by deploying applications into production more frequently, while reducing cost and maintaining reliability
• A CICS cloud allows you to deploy new applications, application features or bug fixes while minimizing any impact to existing users or requiring additional infrastructure

Webster’s first scenario was: “The mobile guys need a new feature but I’m concerned about the potential performance impact for the existing high-volume website. This is a pilot so I really don’t want to spin up new regions just to support a few users. I’m also still trying to move the back office users off an old version of the Application.”

A CICS application has a name, version, entry point and now resources. This goes in the platform. It’s now possible to have application V1.0.0 and V1.1.0 hosted on the same region(s), where each define PROGRAM P. Older versions of PROGRAM P can be retired when users have migrated to newer versions.

Webster’s second scenario was: “I need to apply a hot fix to an application in production but I want to use the same process that I use for my weekly updates so I get an audit trail and correct monitoring data. I really want to make sure it’s installed correctly before making it live, while the existing version is still being used. Also, if the update makes things worse, I want to roll back the change as quickly as possible.”

If application V1.0.0 has a bug, users can INSTALL and ENABLE V1.0.1, but requests still go to V1.0.0. User can then make application V1.0.1 AVAILABLE so it receives new requests. If there is a problem, users can make V1.0.1 UNAVAILABLE to rollback to V1.0.0. CICS Explorer can then be used to make the changes required.

Webster’s third scenario was “When rolling out a new version of an application I want to initially give access to just 10 percent of my users. That way I can reduce the impact of any potential problems related to either the new features or to the platform because of performance. The requests arrive over MQ, so I want to use origin data to route each one to the appropriate application version.”

Using the cloud, most users get application V1.1.0, but “early access” users can be pointed to the latest application V1.2.0.

The fourth scenario was: “I have two applications (one of which was developed by a company we acquired) that currently run on different sets of CICS regions. I’d like to take advantage of the recent scalability improvements, especially being able to increase MAXTASK. However, I know that there are PROGRAM name clashes, which prevent these applications from being hosted together.”

It’s now possible to run two different programs with the same name in the same region.

Webster’s fifth and final scenario was: “We have spent a lot of time using CICS Interdependency Analyzer to understand the call structure of one of our applications. This has enabled us to add more validation logic to avoid ABENDs when a copybook changes and someone hasn’t recompiled all the right modules. But now we need to ensure no one bypasses these new checks.”

Again, Webster was able to show how this could be done using CICS Explorer cloud.

Webster summarized by saying that sites could use a CICS cloud to deploy new applications, application features or bug fixes while minimizing any impact to existing users or requiring additional infrastructure by:

• Implementing the different scenarios using private PROGRAM and LIBRARY resources, the AVAILABLE | UNAVAILABLE application status, and INVOKE APPLICATION API
• Managing CICS cloud with Explorer

A copy of Webster’s presentation is available for download from the Virtual CICS user group website. You can see and hear the whole user group meeting by downloading the WMV file.

Trevor Eddolls is CEO at iTech-Ed Ltd, an IT consultancy. A popular speaker and blogger, he currently chairs the Virtual IMS and Virtual CICS user groups. He’s editorial director for the Arcati Mainframe Yearbook, and for many years edited Xephon’s Update publications.

Posted: 3/17/2015 3:00:15 AM by Trevor Eddolls

Print Email

Join Now!