rich@eddie.MIT.EDU (Richard Caloggero) (11/04/87)
We recently acquired an Aliant FX8. We wish to fit the Alicnt into our Apollo ring. As we see it, we have a choice among the following options: 1). We could configure all of our Apollo nodes (DN3000's) to use Ethernet as the main network (replace the ring controller with an Ethernet board and get the new rev. of the boot prom). This seems to imply to me that the Aliant will then look like another node, and if we run NFS, the file system will then be homogeneous throughout the environment. To what extent is this true? 2). We could configure one DN3000 as a gateway between the ring and the ethernet. This would allow us to interact with the Aliant via the normal telnet/rlogin etc. utilities, and allow file transfers, however, seems to me that the file systems cannot be made homogeneous, even if NFS runs on both the ringnet and the ethernet. Is this true? 3). Does anyone out there have any other suggestions? Thank you for your time; I will summarize any responses any post the results. -- Rich
hays@apollo.UUCP (11/06/87)
In article <7346@eddie.MIT.EDU> rich@eddie.MIT.EDU (Richard Caloggero) writes: > > > We recently acquired an Aliant FX8. We wish to fit the Alicnt >into our Apollo ring. As we see it, we have a choice among the >following options: > > 1). We could configure all of our Apollo nodes (DN3000's) to use > Ethernet as the main network (replace the ring controller with an >Ethernet board and get the new rev. of the boot prom). This seems to >imply to me that the Aliant will then look like another node, and if we >run NFS, the file system will then be homogeneous throughout the >environment. To what extent is this true? 1> When running NFS to Apollo you may mount the entire Apollo Filesystem (Everything from '//' on down) with a single mount. The Apollo configuration need only mount the Alliant filesystem on one Apollo to be accessable to all (provided the NFS type manager is installed on all nodes). To this extent you will have a single filesystem within the limitations of NFS. 2> Your Alliant <-> Apollo connection will be more efficient if the Apollos are directly on the ethernet. (1 vs. 2 hops) > > 2). We could configure one DN3000 as a gateway between the ring >and the ethernet. This would allow us to interact with the Aliant via >the normal telnet/rlogin etc. utilities, and allow file transfers, >however, seems to me that the file systems cannot be made homogeneous, >even if NFS runs on both the ringnet and the ethernet. Is this true? 1> Same as above. 2> This configuration is logically no different than the direct to ethernet configuration. However all of your Workstations will now be going through the gateway and contending for that resource. --John -- John D. Hays, Consultant, Corporate Systems Engineering | My opinions are my Apollo Computer Inc. - Chelmsford, MA - USA | own -- If anyone ARPA: apollo!hays@EDDIE.MIT.EDU // UUCP: apollo!hays | really cares! PACKET RADIO: KD7UW @ K1UGM // CIS: 72725,424 GEnie: KD7UW // PO Box 21, 01824