IBMophiles treated themselves to a fix recently with a UK Computer Measurement Group-IBM System User sponsored debate on the AS/400 versus the RS/6000. Rather surprisingly, of the 70 or so delegates, only a handful had experience of both systems, and only a few more admitted to planning an open systems strategy. Perhaps the speakers cut their cloth accordingly since most seemed to favour IBM’s proprietary offering, while showing nothing but respect for the Unix box. Events kicked off with The Best of Both Worlds, a partisan view by Peter Idoine, IBM’s AIX marketing manager for government and public services. The two systems have a number of common traits, not least of which is support for commercial computing; communications and distributive capabilities; personal computer support; and both are said to comply with standards, Unix or Systems Application Ar-chitecture. SAA is a standard?

Skills

Anyway, differences include operating systems, with OS/400 described as high function and proprietary; the RS/6000 is both a single-user technical workstation and a multi-user commercial system, while the AS/400 is only for multi-user commercial computing and it is an evolution from System 38; functions are packaged into OS/400, but the RS/6000 gets functionality from optional products, and the AS/400 comes complete with relational database while RS/6000 users have to choose from third parties – at present; and finally, the proprietary offering requires little technical support while the Unix box requires more specific technical skills. Idione claims that open systems is where the world is heading, and IBM is committed to that movement, both in Unix and in SAA. How so? He says that recent statements from IBM have recognised the need for SAA to be more open, but he denies that SAA is becoming any less significant in IBM’s strategy. Idione suggests that users lead with the AS/400 if they are working within a System 36 SAA-based environment and they have minimal technical staff. If open systems, technical and graphics capabilities with multi-vendor communications is required, then he falls into the RS/6000 camp. Given the nature of IBM’s proprietary products, and by definition, its installed base, those guidelines seem to offer the Unix fans poor pickings. The next speaker, Tony Barsham of Price Waterhouse, IBM’s auditors, believes that the RS/6000 represents an ideal opportunity for IBM users to test the waters of open systems while staying in an IBM environment. Which begs a question that Barsham chose not to address. If they test the waters and like the temperature, will they want to stay within the IBM world with all the subtle price overheads that IBM imposes on its open systems users? He believes that the open systems decision drivers are reduced cost, breakout of proprietary lock-in and the ability to use new technology.

By Janice McGinn

However, there are inhibiting factors, and these include the immaturity of open systems and the lack of good commercial applications for Unix. Lack of packages is a major issue, according to Barsham, since many organisations base their computing strategy on packages, and they have been accustomed to a plethora of the things. Also, migration is an issue since organisations have made significant investments in proprietary technology, not only in hardware and software, but in staff and skills. Migration cannot be done overnight and it requires extensive retraining, which doesn’t come cheap. Carolyn Nimmy, technical services director at Hoskyns Insight, installed the first System 38 in the UK, and is an unashamed apologist for the AS/400. However, Ms Nimmy acknowledges that the RS/6000 has its own strengths, not least of which is a highly scalable operating system, and good price performance. She compared the two systems on the basis of architecture; database management; security and systems management; communications; user interface; development environment and applications. Ms Nimmy believes that everyone, including IBM, underestimated the success of the RS/6000 and the impact it would hav

e on the AS/400. It has affected sales and delayed the selection process, but she also suggests that the impact has been worse in the UK than the rest of Europe, and says that France has had its Unix revolution and returned to the AS/400 fold. Comparing the two systems on price performance is difficult since the AS/400’s transaction throughput is measured in RAMP-C – Requirements Approach to Measuring Performance in Cobol – and the RS/6000 is measured in terms of MIPS or MegaFLOPS. Ms Nimmy claims that in the commercial arena, the AS/400 sizing process has developed and become reliable, but the RS/6000 is still a dark horse and until we have some public benchmarks running the same commercial applications, comparisons will continue to be a matter of debate. As regards database management, Ms Nimmy has a deal of praise for the AS/400. She suggests that not having a choice of database can be an advantage for non-technical sites although it can be restrictive, but SAA-compliance and integration with all AS/400 facilities is a bonus. While the RS/6000 offers choice, it is disadvantaged in that yet another piece of jigsaw needs to be managed with regard to version control. Also, the database management system must be standard throughout a company to reduce management requirements and training costs. On security and systems management, Ms Nimmy believes that the AS/400 is easy to administer and can be very secure. It is provided at the object level, user profile level, and workstation level, and can be implemented at file, record and field level. She maintains that security is not a problem in the Unix arena, but it will improve as more and larger multi-user systems are installed. Likewise, systems management is more mature in the AS/400 arena since large multi-user Unix sites are not common, but that will change with time. Both systems score high marks on communications, although the AIX system has additional strengths with its SNA offerings. The AS/400, says Ms Nimmy, has an integrated development environment, but the RS/6000 is no less impressive with lots of choice.

Commercial

Commercial applications are plenty in the AS/400 world, with over 8,000 currently available. The number of applications and the large agent community in partnership with IBM has provided one-stop solutions, and this, claims Ms Nimmy, has contributed a great deal to the success of the mid-range system. The RS/6000 has a large number of applications, but less in the multi-user commercial environment, although that situation is changing rapidly. Ms Nimmy summarised the strengths and weaknesses of the AS/400 as follows. It is well established in the commercial and development worlds; it has SAA conformance; a scalable hardware architecture; a wealth of applications; and its easy to manage. Its drawbacks? Most noticeable is the priceperformance compared with open systems. Other weaknesses include no choice of a relational database management system, and the proprietary nature of the AS/400. The RS/6000s strengths lie in its scalable operating system; good price-performance; multiple database management systems; its openness; and governments’ open procurement strategies. It is weakened because it isn’t part of SAA – but that is a marketing decision by IBM, and may change it requires more systems integration than the AS/400; and she claims it its is expensive for personal computer-type applications. Consequently, when moving to a new environment, Nimmy recommends that users consider training, application and hardware costs. No less important is maintenance and the extent of value added reseller support available.