SNA’s free-fall down the network popularity chart has stopped, finds a report from Cambridge, Massachusetts-based analysts Forrester Research Inc. According to IBM’s Second Chance, 55% of Fortune 1,000 management information systems executives interviewed in the US said that the size of their SNA networks is no longer shrinking as fast as a year ago. When questioned in May 1993, 64% of interviewees said their SNA network was shrinking, a figure that had declined to 29% when they were reinterviewed last December. Of the same sample, 55% said that the size of their SNA network was now stable, against only 20% in May two years ago. The reason for the slowdown in SNA’s decline is rooted in the fact that all the easy-to-move applications have already migrated from mainframes to local networks, the managers reported. Conversely, they said that tearing down the remaining SNA strongholds would require huge application re-engineering efforts. Also, despite the strategic goals of rushing to client-server, they added that the money to replace dumb terminals with personal computers was still in short supply. According to Jay Batson, author of the report, most big companies have embraced local network internetworks as the new backbone for their firms, and have a good handle on which SNA installations can easily be converted over to the new network. This is good news for IBM, says Batson. The SNA mudslide has stopped and IBM has survived to fight another networking battle… it need not be distracted by SNA’s slide as it climbs back into the game using Asynchronous Transfer Mode technology. However, Batson is critical of IBM’s Asynchronous Mode approach. In his view, the company has a two-pronged strategy for its comeback: first it believes that, since Asynchronous Transfer Mode is new to every vendor, its research and development muscle will carry it triumphant back into the game, beginning with Asynchronous Mode wide area network switches to replace ageing wide-area SNA nets; second, it is confident of convincing users to converge future networks around its new Advanced Peer-to-Peer Network/High Performance Routing protocol and AnyNet software. This strategy flies in the face of three powerful market forces that will not change, thinks Batson. Customers will be slow to replace mountains of legacy local area network gear overnight with Asynchronous Transfer Mode.
Lightning speed
Plus, when they do, today’s router vendors will hold the upper hand in the conversion. Finally, he feels that TCP/IP is blanketing corporate desktops at lightning speed, making IBM’s protocol plans irrelevant. Despite these forces, Forrester still gives IBM a better than even chance to get back into the game – if it changes its thinking to recognise that networking is an ‘open’ game. According to the report, the company will make it to the next round of ‘networking playoffs’ if it shifts its Asynchronous Transfer Mode wide area network investments to Asynchronous Mode local networks, succeeds in riding the TCP/IP boom effectively, and re-casts its home-grown protocols in a lesser role. Finally, Forrester notes that, with Asynchronous Transfer Mode-attached 3090 mainframes running TCP/IP, the pressure to unplug mainframes would abate, giving IBM’s mainframe business a stay of execution.
