I really enjoyed reading Bill Carico’s recent column on Mainframezone.com titled “Sanity Check: Let Me Be Perfectly Vague.” The story he tells about one company’s misguided efforts to save money by moving workloads from the mainframe to a distributed platform (with dire consequences), and one vendor’s determination to help them do this by focusing on all the wrong issues, is all too familiar. Indeed, Bill has collected a number of similar case studies over the years, and they’re available in the Re-Boot Hill archive at www.actscorp.com. It’s remarkable how often companies have made similar mistakes.
As Bill says in his column, the competitive vendor will generally target senior execs within the business rather than the existing IT professionals, and will go in with a highly polished marketing presentation rather than a clear, comparative analysis of the technologies involved.
It’s strange to consider that, while the S/360 architecture on which today’s mainframes are based is nearly 50 years old, the platform has been under fierce attack from distributed system advocates for well over half of its life. The number of expensive attempts to move away from the mainframe that have taken place during that period is hard to assess, but large enterprises seem incapable of learning the lesson. During those years, many of the execs who were won over by glib promises of cost savings and simplified technology from competitive vendors have now moved on, and a new generation of mainframe-skeptic CxOs need to be educated in the reasons why mainframes have endured so long and still perform a unique role within the enterprise.
Much of this “education” concerns the need to see the mainframe as more than an IT platform, to understand that it’s a whole set of management disciplines and best practices that are ill-defined in other areas of IT. As such, System z potentially offers a strong foundation for today’s cloud-based computing environments, which depend so heavily on seamless scalability and capacity management issues. Yet expressing these concepts in simple, bottom-line financial terms is extremely difficult.
In the IT service management area, where I currently spend most of my time, we face similar difficulties in quantifying the return on investment that many organizations are trying to make in disciplines such as change-, problem-, and service-level management. The fact these processes were introduced and nurtured in the mainframe world long before they were adopted elsewhere is often forgotten; and, properly implemented, they can help build a far stronger relationship between IT and the business than has existed until now.
As financial pressures in the commercial world show little sign of easing, senior execs will continue to face really tough financial decisions, and will continue to be won over by what seems to be a good argument for pushing Big Iron out the door and finding a more fashionable platform for enterprise applications. One can only hope they will demand a full, detailed technical comparison of the proposed solutions before they make a decision that could seriously add to their woes.
Compuware Highlights Mainframe Savings
A recent report from Compuware titled “Mainframe Efficiency in the Real World” describes how 60 companies made savings totaling $105 million by using the vendor’s suite of tools. The diverse range of savings included a 50 percent reduction in CPU consumption, productivity gains of 55 percent, an 81 percent reduction in storage costs, and a batch processing window reduction from five hours to 30 minutes on a business-critical application.
While other ISVs might be able to provide comparable figures, what strikes me most about the report is the sheer scope for improvement among mainframe users that it reflects. When faced with the need for cost reductions, as discussed earlier, many users will take the drastic step of trying to move critical applications to another hardware platform, when the most obvious and straightforward savings can be gained by re-examining hardware and software contracts, implementing software asset management more effectively, and using tools such as those offered by Compuware and others to fine-tune performance and identify inefficiencies.
Mainframe Madness Is Back—and Has Been Extended
Once again, CA has provided an impressive virtual conference throughout the month of May, with a program covering a broad range of technical and strategic topics, all available through the Web. The event uses online chat and networking tools to make the more than 100 sessions as interactive as possible, creating one of the best virtual resources available to mainframe users worldwide.