IBM Corp has duly decided to license its Advanced Peer-to-Peer Networking – APPN – code, but no-one has yet come forward to say that they will be using it. Cisco Systems Inc, for example, says that it is still in negotiations – the company has had a team of people working on reverse engineering the APPN code, and has yet to decide whether to abandon the effort. We recently learned that Cisco’s Phase 3 products will support local acknowledgment of traffic in order to improve wide area network connections, and will translate between SDLC and LLC traffic. And now, lo and behold, IBM’s new 6611 MultiProcessor Network Processor Program software does too. Coincidence, says Cisco. Together, the software implements a set of features that IBM calls Data Link Switching Technology and it is this that the company has licensed to Network Equipment Technologies Inc. As you would expect, the 6611 includes a Simple Network Management Protocol agent, which is useful since the new AIX NetView/6000 implementation is specifically designed to manage these devices – and these alone. In a network that incorporates SNA devices, the AIX NetView/6000 sends SNMP data back to the mainframe implementation of NetView, so it seems a bit rich to call the thing a version of NetView when it looks more like a fancy SNMP management station with a NetView gateway bolted on. IBM has been busy ensuring that other vendors’ SNMP devices will work with it, though. Chipcom Corp, Fibermux Corp, Optical Data Systems, Proteon Inc, Synoptics Communications Corp, Wellfleet Communications Corp and Xyplex Inc have all enrolled into its network systems programme and a copy of their SNMP management information bases will be included with AIX NetView/6000 when it ships in June. This March should see the RS/6000 machines get much-needed FDDI interfaces, with both dual-attached and single attached stations supported. The device driver is written to the ANSI X.3T9.5 standard except for the station management function, which is limited to agent support. Also lurking in the small print are the details of IBM’s plan to turn the RS/6000 into a redundant, fault-tolerant server: High Availability for Network File System or HANF for short. The idea behind HANF is familiar – take two RS/6000s, four network adaptor cards, and so on, and set them up so that they mirror each other.
