IBM Corp is using Raytheon Co’s UK software development and services division, Data Logic Ltd, as the prime development partner on its AIX CICS/6000 and Encina on AIX/6000 products announced at Unix Expo last month (CI No 2,013) and due next June and this December respectively. The Harrow, Middlesex firm has been responsible for designing, re-engineering, coding and implementing IBM’s Customer Information Control System, CICS, on top of the Open Software Foundation’s Distributed Computing Environment and Transarc Corp’s Encina transaction processing services to run on RS/6000s under AIX v3.2. Much of the work is being done at IBM’s Hursley Park, Hampshire lab. Despite its role, Data Logic hasn’t shied away from voicing some concerns over its employer’s attitude to transaction processing-on-Unix, concerns lodged not on the technology side, but at IBM’s marketing approach. Offering both a straight implementation of Encina and CICS over Encina on its RS/6000s offers users more confusion than choice, Data Logic muses – though probably even more than that, given that IBM will also be offering a different transaction processing technology, Unix System Laboratories Inc’s Tuxedo under AIX on its ES/9000 mainframes. Offering Encina as the only transaction environment on the RS/6000 would have been better, argues Data Logic, since 90% of Encina is in the CICS/6000 product in any case – CICS/6000 sits on top of the Encina toolkits – and only the monitor is different. Whether you buy Encina or CICS/6000 for an RS/6000, you buy Encina, Data Logic notes. However despite the commonality between them, CICS/6000-to-Encina Application Programming Interfaces are not available.

Straddle

Although it is possible, according to Data Logic, to build transaction processing systems that straddle the two technologies, you end up with software that may be interoperable, but won’t be portable between the two. That said, in its work to bring CICS over to AIX, Data Logic has provided the capability for CICS/6000 to be migrated to other systems, meaning that IBM can make the environment available for other vendors’ machines, as already announced with the Hewlett-Packard Co’s agreement (CI No 2,019). This is the route that Data Logic thought IBM was originally treading, a CICS/Open strategy in which the transaction processing environment, plus the wealth of CICS applications already availaable, could be brought to the open systems market, which has only a handful of fledgling transaction processing technologies to offer, and few software packages. There are no engineering obstructions to IBM putting it [CICS/6000] on other systems, Data Logic says. In addition, it believes Encina could have quite conceivably gone onto the ES/9000 mainframe in place of Tuxedo, but thinks IBM’s deal with Unix Labs for that technology is effectively a spoiler for Amdahl Corp’s Tuxedo TP system offering on its IBM-compatible Unix mainframes.

By William Fellows

Given the portability of CICS/6000, IBM could even have – or could still – offer CICS running over Tuxedo on its mainframe line. Encina already encompasses Open Software Foundation’s X/Open Co Ltd-endorsed Distributed Computing Environment technology – Unix Labs will incorporate DCE into Tuxedo over time. Despite the fact that all of the currently available transaction processing systems on offer come with proprietary Application Programming Interfaces, each of the vendors is quietly pressing the X/Open standards body to adopt its particular transaction processing methodology. Both Encina and Tuxedo could interoperate with CICS/6000 systems, though neither provide for a direct and easy migration route to it, according to Data Logic. That said, Tuxedo and Encina are aeons away from each other too, while CICS/6000 is at least an open systems migration route for standard CICS skills and applications. Without a CICS/Open product on the market, Data Logic casts the race for Unix-based transaction processing systems outside the IBM world as a Tuxedo-versus-Encina affair where the market will decide whether they

get pushed into interoperability. The embryonic Encina, it believes, with far fewer users than Tuxedo, certainly has a lot more to do to win the mindshare it needs. Meanwhile Unix Labs is setting up infrastructures that will enable Tuxedo to interoperate with CICS-based systems and Unix International has, with Data Logic’s help, published a guide for integrating Tuxedo with mainframe CICS. It’s the same kind of impetus that is needed to enable Encina and CICS/6000 to work together, argues Data Logic. CICS/Open is certainly one of the double-edged swords IBM is currently wielding. While a CICS/Open strategy would likely furnish it with a lead position in the Unix transaction processing market by dint of the number of CICS applications available for it, CICS/Open could also open the way for traditional CICS users to move not only to Unix, but away from IBM hardware systems altogether. However, Data Logic says it is more than willing to show how a CICS/Open implementation could be achieved, and is prepared to support companies willing to undertake the work. CICS/6000 is a good product, Data Logic argues, but falls down because it is now caught up in the world of IBM marketing strategy. Highlighting the client-server benefits of CICS/6000 would have made Data Logic feel better about Blue’s intentions, but then the event at Unix Expo a couple of weeks ago was owned by IBM’s Personal Systems Division. IBM will probably have two sales teams, one peddling Encina, the other CICS/6000, but the main problem, Data Logic envisages, is one of kicking over statues. CICS has traditionally been a tool for selling IBM mainframes. To help CICS/6000 succeed, supporting software and application availability will have to become selling points, the firm believes.

Management

Doing its part, Data Logic says it will sell software and education services for the environment and help IBM market it through indirect channels. Data Logic is still working on the CICS/6000 implementation, which isn’t due for another nine months, and is keeping around one third of its workforce busy. Meanwhile the next challenge for transaction processing technology is the management of distributed transaction systems, the company argues, because the emerging Distributed Computing and Distributed Management Environment-type technologies are aimed primarily at managing distributed systems of the conventional sort. The next generation of transaction processing technologies now coming out of the labs are based on reliable queuing mechanisms – messaging services that enable distributed systems to be linked together but which maintain the integrity of transactions. Transarc has a reliable queuing mechanism in the works – which like Encina, is being developed on Digital Equipment Corp kit – while Unix Labs has /q under way for Tuxedo.