Jan 4 ’11

Mainframe Modernization:  Modernize Mainframe Applications With Middleware

by Editor in z/Journal

Most data still lives on the mainframe, and Web front-ends have made some of that mainframe data more accessible. For example, when you use an online banking application or place an online order from a brick-and-mortar store, you’re accessing mainframe data through a Web interface.

To meet rising demand, IT organizations will continue to make even more mainframe data accessible through the Web. While Web-enabling legacy data has many benefits, it also delivers a different set of challenges: extreme availability and security requirements, the need to find and fix problems on a variety of platforms, and the ability to get the mainframe and distributed systems applications to communicate with each other.

The zEnterprise provides mainframe and distributed systems hardware in a single box. You may have initiatives to exploit this technology and make hybrid computing a reality. What’s the easiest way to enable non-mainframe access to legacy data?

New Applications, Legacy Data

The mainframe is an excellent platform for data warehousing and data management because it provides unparalleled reliability, availability, and serviceability. But when mainframe data is shared with and updated from distributed systems applications, it isn’t feasible for all users to have native access to mainframe applications. Also, most distributed systems application developers aren’t familiar with mainframe protocols.

When you need to make mainframe data available to a non-mainframe audience, you have a few options:

Middleware such as WebSphere MQ and WebSphere Message Broker makes it easy to share information between the mainframe and distributed systems because everything works through a standard protocol using messages. Using messages eliminates the need to write new application programs. Whether a transaction begins on or off the mainframe, the protocol is the same: read the message, process the message, and then write a new message.

Because communication is handled through messaging, distributed application developers don’t need to learn COBOL or even know the mainframe protocols. Mainframe programmers can work with the programming languages they know without worrying about Java and other non-mainframe languages and protocols. Programmers on both sides of the fence send messages that query and update data and receive replies through the same interface. The middleware does all the data transformation and it doesn’t care what language the request is in.

Middleware enables you to use the data and applications you already have in place. Non-mainframe developers can access just the mainframe information they need without directly accessing the mainframe. This approach limits access to actual mainframe data and reduces the risk you would have if you opened up the legacy data floodgates. 

Problem Resolution

When applications span multiple platforms, databases and systems, it can be difficult to find the root cause of any slowdowns or problems. Each layer added to an application brings another potential point of failure. When you make mainframe data available to the Web, you need fast response times and near-continuous availability, so any slowdowns or outages are unacceptable. You need to monitor the traffic flowing through your business applications at each technology tier to ensure timely delivery from both a business and an IT perspective.

To measure key performance indicators, you need a framework that transforms data and delivers synchronized information to stakeholders who have varied requirements. This type of framework provides cause-and-effect information much faster than typical business intelligence or independent IT monitoring systems.

When you’re ready to modernize your mainframe and exploit hybrid computing, invest in middleware; it’s the easiest, fastest, and safest path to follow. And to exploit your middleware investment, choose a reliable tool for managing, monitoring, and ensuring optimal performance.