Last months’ arrival of IBM’s CICS/MVS Version 2.1, combined with DB2 Version 2.0 availability murmurings earlier this year, have served to refocus attention on two of the more abstruse – if strategic – areas of the IBM world. Watching with particular interest from the education and consultancy sidelines is On-Line Software International, which recently took […]
Last months’ arrival of IBM’s CICS/MVS Version 2.1, combined with DB2 Version 2.0 availability murmurings earlier this year, have served to refocus attention on two of the more abstruse – if strategic – areas of the IBM world. Watching with particular interest from the education and consultancy sidelines is On-Line Software International, which recently took time out at its Hanover Square, London base, to conduct a half-day light-shedding forum, aimed at evaluating the latest – and projected – product offerings. For CICS, neatly defined as an environmental solution to data sharing, an attempt to place the product in corporate historical perspective proved a particularly pertinent exercise, with many of its user- – and IBM- – perceived defects stemming from its relative antiquity. First released in 1969 as an operating system product, CICS architecture still suffers to some degree from its event-driven architecture. Protracted attempts by IBM to kill it off, culminating in the 1975 decision to transfer US product development and maintenance to Hursley Park in the UK, were, however, foiled by the grudging realisation that, in functionality terms, it far outstripped its much-pushed at-the-time rival, IMS. Recoverability The subsequent introduction of the High-Level Program Interface, HLPI, helped to transform CICS into a more user-friendly, functional interface, with later enhancements including the addition of the DL/1 more English-like language interface. Today, the number of CICS licences has grown to over 29,000, with the latest XRF version clearly designed to appeal, in full-circle fashion, to the security and recoverability-seeking financial institutions that have traditionally opted for IMS. XRF was announced in February 1987 and promised for delivery this quarter in the UK, on a managed basis (CI No 624). Essentially, the CICS/MVS XRF concept is based upon two systems acting as one, to ensure low-overhead, near continuous operation. While an active system A does the work, an alternate system B, running in parallel, performs a continuous series of two-second tracking and surveillance cycles,communicating with A via a common set of Availability Manager data files. In the event of CICS, MVS, VTAM or system failure, system B is designed to take over operations with minimal interruption to the user. Additional refinements include automatic session switching on VTAM networks, automatic when pre-set – takeover, and operator controlled takeover – a planned takeover by the operator, firmly paving the way for the 24-hour operation era. XRF does not, however, offer an IMS/VS style hot start; as a software emergency restart facility, it often forces users to reposition themselves after a service interruption, in contrast to hardware fault-tolerant solutions, which are based on the duplication of all operations running in parallel. Another major On-Line-diagnosed weak recovery link is the sharing of database disks, which bears the additional disadvantage of confining operations to one physical site. Subsystem incompatibility also figured on the minuses list, with CICS/MVS XRF currently offering compatibility only with the very latest version of IMS, and – without the provision of the necessary XRF Logic – awkward for users of DB2. Long-term, however, the lot of the CICS/MVS XRF user looks more encouraging: the software is promised to conform to Systems Application Architecture standards, already goes along with the AS/400, and will probably be one of the first products to support the newest SNA definitions. Additionally, the user-friendly and cost-effective nature of CICS/VM will, according to On-Line, help to boost the floundering sales of the 9370 Series. DB2 Version 2 On-Line was also keen to cast a speculative eye over DB2, particularly the announced but elusive Version 2. IBM introduced it in April with first deliveries promised for this October or November (CI No 912), and it has been somewhat dangerously billed as the answer to existing DB2 user’s prayers – or complaints. To date, these can be divided into two spe
cific areas: performance; and the implementation – or not – by the existing Version 1.0 of the E F Codd-devised full relational theory. According to On-Line’s Phil Rowe, the former is the direct result of hardware lagging behind software in the development stakes – a situation which he believes will change with the arrival of Enterprise Systems Architecture, ESA. The latter is, however, a more complex issue, which centres upon two key tenets of database gospel – primary key and referential integrity. Primary key is the method prescribed by database academics to ensure that every row in a database table contains unique information but simultaneously conforms to standard methods of representation; referential integrity presupposes the implementation of the primary key theory by dictating that the value of a foreign key in one table must be equal to the primary key in another, or, in other words, that values from table to table must match. At a practical and more accessible level, both measures are intended to prevent data duplication, and ensure true data integrity. Referential integrity In contrast to its precursor, DB2 Version 2.0 offers support options for referential integrity; currently, application developers seeking to implement the theories are forced to introduce them manually, which costs both time and processing power. The trade-off, naturally, is performance; IBM maintains, however that Version 2.0 under ESA will offer a 50% improvement on Version 1.0, and – crucially – that the boost will extend to both the high-speed transaction processing and the decision support ends of the performance spectrum. On paper, Version 2.0 also appears to offer a series of operational enhancements, notably an audit facility to check for violation, a resource limit control, and a number of security modifications. Vital was the term coined in conclusion, extended to cover not merely the products, but equally the role that education has to play in ensuring that maximum efficiency is offered by products and exploited by users, well into the 1990s – that, inter alia, is what On-Line Software is in business to provide.