sob@harvisr.harvard.edu (Scott Bradner) (01/15/88)
I've been testing the LattisNet product for the last 2 months. It seems to work just fine. There was some badmouthing of this product by a-large-computer-company, saying that it did not run at 10MB. I have seen no reason to feel that might be true. I have tested a sun 3/60 connected via LattisNet to Harvard's backbone & see no change in throughput or error counts when compaired to a sun direct connected to the backbone. Scott
lekash@ORVILLE.NAS.NASA.GOV (John Lekashman) (01/16/88)
We've been using Lattisnet for about fifty machines, and it seems to work pretty well. When it was first installed, there were some packet dropping problems, and they came out and installed some sort of additional buffering. Perhaps you have an older version, or else the lower cost, less buffered version. Maybe when the machine count goes to a hundred, we might have a different story to tell, (ie its time to buffer more.) However, single point to point transfers seem to run as fast between machines on the Lattisnet as those not on it. This was tried only with sun 3/260's and SGI irises, so far. john
ddp+@ANDREW.CMU.EDU (Drew Daniel Perkins) (01/16/88)
We've been using Lattisnet for quite some time now (2 years?). In fact, we were one of their initial beta testers. We aren't using the thinwire stuff yet, but we are using their system over our IBM cabling system and over fiber. We have a few hundred connections to multiple lattisnets and have never had a single hardware failure. Our campus backbone is also a lattisnet box. I can't say enough good things about the product. We are very happy with it and we are using it in almost all future ethernet connections, as well as converting many old connections to it. It's ease of use and maintenance are great since it gives us a star shaped ethernet in centralized wire closets. As far as cost goes, it is a bit more expensive. However, we feel the additional cost is small compared to what we get in maintainability. In the long run it may even be cheaper when we factor in technician time for installation and maintenance. Drew