When discussing IBM’s mid-range and the RS/6000 in particular, one factor that is of crucial significance is the RS/6000’s operating system. This is because while IBM’s AIX is in existence and work is ongoing with OSF/1 there is going to be confusion in the marketplace. After all, why should the punters buy the RS/6000 now […]
When discussing IBM’s mid-range and the RS/6000 in particular, one factor that is of crucial significance is the RS/6000’s operating system. This is because while IBM’s AIX is in existence and work is ongoing with OSF/1 there is going to be confusion in the marketplace. After all, why should the punters buy the RS/6000 now when it comes with an operating system that will be obsolescent as soon as OSF/1 is adopted? John Glyde, IBM UK’s AIX manager, says that the RS/6000 will have one operating system called AIX which will be an implementation of OSF/1. IBM’s version will have bits added to it, such as SNA communications, and other participating vendors will add their own flavour to the open operating system, which will be compatible at the level of Application Environment Specifications.
Glyde says that OSF/1 will not be implemented in one leap – it will be implemented in stages. Each release of AIX will include more of OSF/1 – there will be upgrade charges and Glyde does not know how many upgrades will be required – and at some point different vendors’ implementations will merge into one operating system compatible at the applications level. Glyde argues that a big chunk of OSF/1 comes from AIX anyway, so that the two operating systems looked the same from day one. However, features will be added to AIX-OSF/1 such as IBM’s System Management Interface Tool, which, while it will not prevent the portability of applications, will, IBM hopes, keep users hooked to AIX-OSF/1. Glyde is adamant that no functional enhancements will be made that will affect application portability. However, he says that members of the Open Software Foundation have no intention of following Unix International’s lead – OSF/1 will not be the same operating system on different vendors’ machines. Instead the Foundation will define a wide set of standards to ensure application portability and interconnectedness between different vendors’ machines via Application Environment Specifications, the Distributed Computing Environment and OSI, TCP/IP standards. He says that by the end of 1991 the Foundation camp will have a fully interoperable operating system. The first implementation of AIX-OSF/1 will appear on the PS/2 sometime next year – the PS/2 because reference ports developed by the Foundation were on to an Intel machine. The implementation for the 3090 is not a high priority and, as might be expected, most IBM engineers will be working on the RS/6000 version. While it is reassuring to know that members of the Foundation, including IBM, will not keep their versions of Unix alive as a competitive offering against OSF/1, it seems likely that astute customers and software developers will await the arrival of AIX-OSF/1 before making an RS/6000 commitment.-
By Katy Ring
Another area where there is much interest in the relative roles of the RS/6000 and the AS/400 is Systems Application Architecture and AD/Cycle. Henry Douglass, IBM UK’s Application Business Systems’ marketing manager, is in no two minds that SAA offers a welcome sales pitch for the AS/400. As he says, SAA means the ability to develop co-operative processing so that applications can be put where they will run best within Advanced Program to Program Communications, APPC – this is not an opportunity that is lost on large enterprise customers. And within SAA, the AS/400 is billed as the application server and the key machine for handling transaction processing. Within AD/Cycle, the AS/400 is positioned as the machine to use for the development of transaction-based applications – it delivers application code much more quickly and less expensively than is the case in a large MVS environment. Douglass says there is no slippage in development plans for the AS/400 Repository Manager, which is still under development in Rochester, Minnesota. Douglass said he was aware that there was a lot of user interest in making the RS/6000 part of AD/Cycle because of its fast processing power, making it eminently suitable for application development. He also said that he would welcome a clos
er affinity between the RS/6000 and AD/Cycle, adding that bringing the RS/6000 into AD/Cycle is a very attractive proposition. Douglass agreed that the role of the AS/400 as the application development machine for AD/Cycle may have been overplayed in the sense that within distributed co-operative processing data and code are accessible by applications on other systems and, therefore, there is not such a need to re-write applications. However, Douglass denied that this move to downplay this particular role for the AS/400 signalled a diminished profile for the AS/400 within AD/Cycle. Similarly, Douglass denied that the fact that IBM had gone out to small UK company Synon Ltd to develop the AS/400 code generator, without bothering to take an equity stake in the company, reflected any lack of will on its part to emphasise the AS/400’s part in AD/Cycle.
No clear plans
Meanwhile, John Glyde was of the opinion that over time we would see a more intimate relationship between the RS/6000 and SAA in particular this will come about via the AIX database which IBM is developing (CI No 1,552). It will also come about through the implementation of the Distributed Computing Environment throughout IBM’s ranges. However, whereas Douglass appeared to assume that the RS/6000 would, sooner or later, be part of AD/Cycle, Glyde was more circumspect. He said that there were no clear plans yet to integrate the RS/6000 into AD/Cycle. He agreed that customers were taking an interest in using the RS/6000 for computer-aided software engineering within large IBM environments. But, he said that the RS/6000 tie-in with AD/Cycle and SAA will occur gradually through a host of little developments – in other words it is unlikely to be heralded by a sweeping statement of direction. Once the RS/6000 is tied in, and should IBM reconsider developing the RS/6000 and AIX for commercial transaction processing, then the time may have come to bury the AS/400, since there will be little logic to its existence. Don’t expect any statements of direction on the matter from IBM – after all it has no mid-range mess to clear up – just remember that when two worlds meet it usually means war.