|
The focus of development today is on business outcomes. The ability to assure the achievement of these (often, while satisfying regulators or corporate legislation) comes from providing transparency through to the underlying technology across a whole business system. In the midst of all the excitement over web applications and new agile scripting languages, it is often overlooked that this underlying technology, in many large organisations, is the mainframe. Even the most agile new web application developments often rely on robust, resilient, trusted mainframe applications, now made even more valuable because a Service Oriented Architecture (SOA) allows them to extend their delivery over the web.
As companies implement SOA and extend their mainframe applications as web services, their IT environment becomes more complicated because of the extra technology layers involved. At the same time, in order to remain competitive and responsive to its customers’ needs, the business demands new applications and changes to existing ones. Of course, the legacy mainframe application underlying these new web services still have to do the jobs they’ve always done, maintaining the “books of record” of the organisation. They must be maintained and enhanced appropriately, to satisfy all the necessary service levels.
It’s easy to see how the cost of maintaining application quality and performance, in terms of resource utilisation, consumption and additional quality assurance processes, could escalate rapidly. The business must take control of the application quality and testing process with the help of effective process automation. This includes providing tools that minimise the time and effort needed by programmers and testers to build quality into applications, thus encouraging them to build quality in from the start and reducing the chance of expensive production failures. Then, if something does go wrong in production, these tools must allow problems to be fixed quickly and reliably, minimising the cost of any production outage. Providing anything less than this is a false economy.
What this means is that Application Quality Management and Testing Tools (AQM&TT) for mainframe applications are just as important as they ever were, but they now have to integrate with similar tools on other platforms to deliver a coherent application quality and risk management message (in business terms) to management. They must take advantage of modern presentation technologies (mainframe tools can’t be a poor relation of similar tools on distributed platforms) and, where appropriate, they must present information at an aggregated business service level. Nevertheless, at the same time they must continue to provide the technical feedback that is of immediate use to mainframe developers.
However, although the mainframe has always been noted for its robustness and resilience, in the past its tools have been purely technician and programmer focussed and there is a real need for innovation in the mainframe toolsets, to build on their existing strengths. This involves rather more than just providing a business-like overlay on an existing tool.
This innovation is particularly important today because application performance and quality management are fundamental contributors to delivering energy efficient automated systems. The “green revolution” in IT is not simply an altruistic contribution to the welfare of the world; it is a vital factor in removing barriers to enterprise growth. Already the very largest enterprises such as eBay and Amazon are finding that the electricity demands their data centres make on the local power grid are the limiting factor in data centre growth.
This white paper examines CA’s family of application quality assurance and testing tools (its AQM&TT product set) in the above context.
It looks at CA’s ground-level developer-oriented toolset—CA InterTest, CA SymDump, and CA File Master Plus—and how these tools, together with its application performance management products, provide a complete solution for ensuring application quality. It looks at how new technology delivered in IBM CICS Transaction Server 3.1 and the recently released IBM CICS Transaction Server 3.2 can be leveraged to improve application quality generally and to extend mainframe applications within a SOA environment.
It examines the automation of mainframe application performance management with TRILOGexpert TriTune1 and TRILOGexpert APC for TriTune with Performance Desktop and the opportunities for proactively identifying problems before they affect production. It emphasises the importance of root cause analysis to identify the issues resulting in excessive demands for production resources. It describes the benefits of monitoring the symptoms of potential problems 24x7 and reducing a bewildering mass of “quota exceeded” “timeout” “threshold exceeded” error and warning messages to a few simple causes that can be addressed in a straightforward manner.
An important focus is on the integration of APC for TriTune with Performance Desktop with CA Wily Introscope in order to provide a coherent picture across both z/OS application performance management and also web application performance analysis.
Finally, it wraps up with a description of how automated application quality is fundamental to the delivery of the service levels and managed risks that are essential to delivering business outcomes you can depend on.Service orientation is a radically new approach to business automation that is becoming mainstream today. It is a business-led innovation and it is important that it isn’t seen merely as a technology issue. True, you may think of presenting IT applications as “services” but it is important that these IT services are described in terms of the business goals they support; and that associated service level agreements and incident reports, etc. are expressed in business terms.
The business services that, taken together, allow an organisation to function can be automated or manual; increasingly they are automated, but the business doesn’t care about the technology involved, it cares about the business outcomes delivered and the corresponding service level agreements that guarantee these outcomes.
In this environment, “rip and replace” of existing technology makes no economic sense. It is best to make use of existing technology and re-purpose it for service delivery using web services—if the business is profoundly uninterested in the underlying technology but interested only in what it delivers, there is little benefit available from rebuilding working systems from scratch, in more fashionable technologies.
However, the IT environment is increasing in complexity. Applications have to support web services across the whole organisation (not just a few key corporate applications for a well-defined set of users) at the same time as increasing emphasis is being placed on cost control. This has implications for the support of existing technology:
You must be able to change your applications as business services evolve—which means that you must be able to find, manage and regression test your existing code.
As business these days is global and 24x7, your tools must be able to support 24x7 operation.
CA Application Quality Management and Testing Tools
As technology will be expected “to do more with less”, the tools must manage resource utilisation effectively.
As service levels depend on delivering adequate performance as well as functionally correct operation, you must be able to QA the underlying technology from both points of view.
As business depends on holistic business services, your tools must be able to support services as a whole. You can’t afford to fail to understand the root cause of an issue just because it’s on a part of the platform your tools don’t know about.
The reality is that a lot of the underlying technology supporting these new business-oriented services will be mainframe technology, at least in large organisations. Mainframe technology still scales better than other technologies and it uses resources efficiently. Also, it has been virtualising the underlying hardware for some 30–40 years—if you want to provide on-demand services from service-oriented applications running on virtual Linux machines, the mainframe is the obvious platform to host them on. Luckily, as we shall see, mainframe technology has the tools needed to support innovative service delivery.
|
