Feb-003 A BIASED PERSPECTIVE by J. W. BENNETT Evaluating mainframe production control automation systems. This article was written by J. W. Bennett, President of Bennett Software, Incorporated. As the title of the article implies, Mr. Bennett presents a biased look at the automated job scheduling market in terms of the technology, support and costs. His company markets an automated job scheduling package, JOBTRAC and the related checkpoint/restart software RUNTRAC, along with several other products. A brief biography of Mr. Bennett is best told by himself: "I spent most of my career as an MIS consultant. You know, those things with an uncanny ability to survive in hostile environments, kinda like ALIENS. I always tried to conduct myself as a true consultant. Of course, I've met many folks that call themselves 'consultants', but the only evidence was on their business card. I define a true consultant as one that: no one influences, everyone argues with, and everything he suggests is implemented. Well, two out of three isn't bad". ----------------- In the world of systems software, one area seems to be taken for granted when it comes to productivity advancements. The IBM mainframe operating systems such as VM, MVS, and MVS/XA have seen quantum leaps in providing hardware and software performance boosts in recent years. Yet the automation of manual operator tasks is still represented by 10 and 15 year old non-IBM vendor technology. Advanced techniques in the automation of change control, batch scheduling, operator response and reply, and output archival are lacking. There are still many labor intensive, error-prone areas that need operational automation. This is not to say that advancements are nonexistent. The MIS software or operations manager must be very aware of the market to identify promising new products. With the "acquisition wars" well underway, many alternatives are being surgically removed from the market by the large software houses. Evolution in Standards IBM has charted a course for the mainframe 370 architecture. It involves a departmental-through-central-mainframe strategy that will, by the reckoning of many, be successful. It is not impossible to envision the development of a 370 standard in mainframe computing just as IBM dictated the PC-DOS operational standard. An integral component of the IBM mainframe strategy is the on-line personal performance standard: ISPF/PDF. As a matter of fact, the ISPF standard is so well defined that under MVS/XA, TSO and ISPF are required to take full advantage of the Interactive Problem Control System (IPCS) for MVS dump analysis. The System Maintenance Program (SMP) also includes a complete interactive ISPF dialog manager to enhance the usability of the product, and improve the productivity of the systems programmer. Just as it is impossible to operate MVS without JES, it is nearly impossible to operate MVS/XA without TSO and ISPF/PDF. TSO and ISPF are, virtually, no longer optional products. Yet many operational productivity systems ignore this continuing standards evolution. Many MIS operations departments are using ISPF for output functions, but use another terminal (non-ISPF) for job scheduling, another for system performance monitors, another for MVS operations... The automation of operational tasks seems, indeed, hard to identify. Automation? After 18 years of specializing in operational productivity, I have developed several slightly cynical points of view when it comes to "productivity" and "automation" software for the IBM mainframe environment. As a matter of fact, my revolution has progressed to the formation of a company with customer support and attention to operational productivity as primary corporate objectives. Possibly the most flagrant abuse of the term "automation" is found in the leading job scheduling and production control systems. Without exception these control systems require huge expenditures for software, JCL conversion, added terminals for workstations, and excessive implementation periods. Many installations report the need to actually make additions to the staff to support the new "automated", $70,000.00, production control system. I don't understand how some customers can continue to implement systems that do no interface with clearly defined standards. Complicated conversions are performed that place the customer at the mercy of the vendor and the vendor's product. In the automated scheduling systems arena, some systems require the JCL to be moved to nonstandard databases, overrides are placed in another, documentation resides in another, and schedule definitions in yet another. Does the customer consider how they will support, or integrate, future products that are based on standard access methods? The only real "next generation" alternatives they have are dictated by the current vendor. What the vendor doesn't support the customer can't have! Add to all this expenditure a 20% per year software maintenance fee, and it is not uncommon to see expenditures in excess of $120,000.00 for implementation and another $12,000.00 per year for product maintenance. Without regard for the new product-experienced production control personnel that may be hired (at the salary of a veteran programmer), the expressed purpose of automating the tasks of one or two operators, at this high price, seems hard to justify. When evaluating automated job schedulers, one of the important things to look for in the product is that it does automate job scheduling. This may sound strange, but just because the vendor calls the product automated, and it is on-line and interactive, doesn't mean it meets the definition of automation, or productivity. Product Support Support for software products varies from vendor to vendor, some are very good, others are quite poor. The poor support usually takes on the following scenario. A new maintenance tape has arrived for the $12,000.00 per year fee. Well, not always. You see, some vendors won't send you a new maintenance tape if you didn't return the last tape they sent. That's right, a five dollar tape can cause the forfeiture of all machine readable maintenance, forever. Well, back to the maintenance. The tape was returned, so the new maintenance tape arrives and everything is going great. After installing the tape it is time to IPL because the maintenance requires it. Now, for those of you not familiar with an IPL, this means that a complete operational computer shutdown and subsequent restart must occur. And, if an error is encountered in the maintenance, another IPL will be needed to fall-back to the original system. Let's assume, as Murphy would have it, an error is encountered and the above IPLs are performed. Next, the vendor support group is contacted. You are told that everyone is busy, but someone will call you right back. You are called back immediately, within the fiscal year, that is. Now one of the big selling points for the package was that there was a local office nearby. This is a powerful marketing point for mainframe software. What the vendor doesn't tell you is that the local office is a marketing office. If you want to buy one of the products, they'll be there in 10 minutes. If you want technical support, that is done by telephone only...out of Mexico City, or somewhere. Then, of course, there's not a technician alive that remembers how the 15 year old product works, anyway. The New Technology Enter the new technology. No conversions are necessary. No new terminals or workstations need be installed. Installation of the product can be accomplished in only 4 to 8 hours. The scheduler can be implemented in a few days. Everything is completely on-line and ISPF interactive. Well written training and usage manuals that can make any operator a master-scheduler, no need to import more personnel. And it works. To extrapolate on the definition of automation, an automated job scheduler should reduce the manual workload, increase productivity, reduce the potential for errors, and most importantly, reduce overall costs. What To Look For There are many products marketed in the automated job scheduling arena today. All aspects of the various products should be investigated. Naturally you'll want one that performs the functions suited to your shop, and most of them will undoubtedly meet that goal. Vendor support is usually the next subject that most customer's investigate, and it is quite important. But don't stop your evaluation at this point. Determine if it really does automate the task. If it requires additional personnel and/or equipment then it may not be as automated as it looks. And don't forget to take a look under the hood. The technology of the software itself is just as important as the functionality. Does it use the latest operating system components and access methods? If not, what will happen when the operating system changes even more? Careful and detailed evaluation and analysis of software products may be time consuming, but is well worth the effort in the long run. /* 1442