One year after leading the mainframe disk business into the world of RAID 5, IBM said it would replace its first generation of Ramac disk arrays with new and markedly improved models. The 11.35Gb Ramac 2 arrays are set for general availability in September. They use 4Gb IBM Ultrastar XP disk drives, giving each four-disk […]
One year after leading the mainframe disk business into the world of RAID 5, IBM said it would replace its first generation of Ramac disk arrays with new and markedly improved models. The 11.35Gb Ramac 2 arrays are set for general availability in September. They use 4Gb IBM Ultrastar XP disk drives, giving each four-disk drawer twice the capacity of the first Ramacs; a single rack will now provide 180Gb of fault-tolerant data storage. The new Ramacs have significantly faster data access times than the first round. Sketchy reports from independent disk traders indicated they will initially be sold for approximately $1.80 to $2.20 per Megabyte. That is half the price (per unit of capacity) IBM asked for the Ramac arrays when they were first launched and somewhat less than the current street price. In addition, IBM is offering very economical lease terms that make the new Ramacs even more attractive. Yet the Ramac 2’s debut was greeted with considerable scepticism. Despite impressive sales of more than 4,500 rac ks of Ramacs to date and the likelihood that the installed population will grow beyond 6,000 during the next year, IBM’s current product line is seen not so much as the future of enterprise storage but rather as a transitional offering. The IBM Ramac array, say most observers, is not the mainframe disk IBM or its large mainframe customers want, and neither are the 3990 or 9394 controllers that give the disks much of their functionality. Customers may disagree with the consultants and migrate en masse from their 3390 or 3380 disks to new or used Ramac arrays, surprising the pundits. If they do, the Ramac 2 could be a stunning market success and the Ramac 1 could have an unexpectedly long life in the field. But IBM’s attitude seems to be more akin to that of i ndustry analysts than that of enthusiastic customers. While it is doing its best to sell both the Ramac 1 drives it has now and the Ramac 2s it will have in just two months, IBM is also developing a dramatically different disk controller for very large commercial systems, a device called Seascape. Nor are IBM’s efforts limited to the controller. IBM is also working on yet another series of disk arrays, which it may call Seastar but which will very likely be launched as Ramac 3. When these new products are unveiled, an event that could conceivably occur before the middle of next year, the company’s current mainframe storage subsystems will be made obsolescent.
Taken out of production
And when the Seascape and Seastar can be delivered in commercial volumes, perhaps at the end of next year’s third quarter, the 3990 controller and associated Ramac disks will taken out of production. The advent of Seascape will redefine IBM’s approach to enterprise storage: Seascape will be an array of controllers sharing a pool of cache memory. Users will be able to add (or remove) controller engines to better match storage subsystems with processing requirements. In addition to Escon channels, Seascape controllers will have other types of ports to accommodate links to non-mainframe processors. In principle, Seascape could become an integrated storage subsystem, managing a mix of disks, tapes and optical storage devices, although at first it may not be that versatile. Seascape will be designed to work with future types of disks, not any prior ones, although IBM has not completely foreclosed support for Ramac 2. From a marketing perspective, IBM views Seascape as an evolving frame that users will upgrade rather than replace as technology moves forward. From a user’s perspective, Seascape is a signal that IBM wishes to provide its largest customers with leading rather than trailing technology, thereby enabling mainframe systems to be used as platforms for new applications and not merely containers for legacy applications. If Seascape is IBM’s future, 3390/Ramac subsystems are not. And therein lies a problem for IBM… an a continuing opportunity for IBM’s competitors in the large systems storage market, EMC, StorageTek, Hitachi and Amdahl. During the next year – or
for a longer time if Seascape, like so many other IBM disk products, is late – IBM will remain vulnerable to the predations of its rivals. In the short run, things may get much worse for IBM. In the long run, anything could happen. At one extreme, IBM’s future disks could ruin the plans of Big Blue’s rivals. At the other, IBM could go from mishap to mishap until its large systems disk market share falls to third or fourth place in the advanced countries. From Infoperspectives International, July 1995, published by Technology News Ltd, 110 Gloucester Avenue, London NW1 8JA, phone 0171 483 2681, fax 0171 483 4541. (C) 1995 Technology News Ltd. All rights reserved.