IT experts, like real estate consultants, often suggest that you should modernize to maintain the value of your investments. While giving a makeover to the mainframe has a nice ring, it’s not clear what that actually means. IBM is investing heavily in the platform, but that doesn’t mean you can sit back and assume your job is done. This article takes a look at your options, and how ongoing investment in your property can maintain, and even increase, its value.
According to a 2005 Gartner survey, IBM’s share of mainframe shipments almost doubled from 2000 to 2004, reaching 70 percent of the worldwide total. While other mainframe manufacturers (e.g., Amdahl, Bull, Fujitsu, Hitachi, Siemens, Unisys, and others) are shifting their investments to platforms based on industry-standard hardware and software, IBM’s z9 mainframe is the result of a $1 billion investment in its flagship System z architecture. The issues for non-IBM platforms are somewhat unique, so this article will focus only on users of IBM System z and its predecessors.
Let’s also be clear about the scale of industry investment in mainframe applications. Analysts predict that by the end of this decade, the number of lines of COBOL code running core business applications could exceed 200 billion. While there are more than 80 variations of COBOL, applications written in PL/1 and RPG [Report Program Generator] add substantially to this inventory. The Aberdeen Group estimates there are “ … $2 trillion worth of mainframe applications in corporations that house approximately 70 percent of all critical business logic and data.”
As businesses struggle to compete in a global economy, the cost of maintaining and running mainframe systems is coming under close scrutiny. Further, the lack of flexibility afforded by legacy applications is hampering growth. Most IT organizations are now considering some form of mainframe modernization project to reduce costs and provide a foundation to support more innovative, agile business models. The IT services industry is waking up to this need by gearing up to offer practices around legacy modernization, but will they work?
Tools of the Modernization Trade
For several years, software vendors have been building tools to help modernize aging mainframe systems. Pressure on IT budgets since the dot.com crash has reinvigorated this market, resulting in a complex, often confusing array of available options. Vendors are often quite small, and constant consolidation adds to the confusion and technology “churn.”
While industry analysts report on the same broad spectrum of tools and techniques available to modernize your mainframe systems, there’s no accepted taxonomy or classification scheme to help guide you to the right approach in any specific circumstance. This can be confusing. What follows are simple categories to help you position different approaches on a scale of overall cost and risk. This categorization draws on the degree to which the different approaches are “invasive.” That is, to what extent does the modernization technique require changes to the original application, its data, development and operational practices, and user behavior? Any such change gives rise to additional costs, and simultaneously increases risk from potential errors and omissions.
• Data access and integration tools enable the construction of a Web service or other programmed interface to provide remote Create, Read, Update, and Delete (CRUD) operations on legacy databases. These may be VSAM or similar flat file structures, hierarchical (network) databases such as IMS or IDMS, or relational databases such as DB2. These tools are useful where the existing data is well-structured, access is infrequent, and data migration isn’t a practical option. Tools also are available to convert from legacy data formats to relational data and to integrate multiple data sources into a single “virtual” database. Existing applications may be recoded to use SQL access or left unchanged to access the data through data adapters.
• Application extension and integration using screen-scraping tools allows a new Web interface to be built upon existing mainframe green screen online applications (also called presentation integration, Web-to-host, or legacy access and extension tools). Many of these tools also can be used to build a Web service or other programmed interface, coordinating multiple online transactions from different applications as a complete business “service” (also called programmatic integration or integration servers). Because these tools require no changes to the original application or data, they’re considered non-invasive. They provide a fast way to connect legacy services into a Service-Oriented Architecture (SOA). To be effective, both the original screen layouts and new service interfaces must be subject to rigorous change control procedures. This approach benefits from the coexistence of the original user interface (the green screen) and the new interface (Web or SOA) but doesn’t provide any modernization of the underlying application or platform.
• Application migration tools and middleware make it possible to move an application, largely unchanged, from one platform to another by providing an emulation of the underlying operating environment. Also called platform re-hosting, this is an effective way to reduce hardware and software costs and to modernize the underlying platform, but requires extensive application testing and changes to infrastructure (i.e., job scheduling, database backup, systems management, etc.). This approach is endorsed by Microsoft and other members of the Mainframe Migration Alliance, but note that the Quality of Service (QoS) of most contemporary platforms still lags that provided by the IBM System z software and hardware architecture. The migrated application also may require extension or re-engineering to Webenable the application, to integrate application services with other applications through an SOA, or change the underlying software or database environment.