Path: utzoo!attcan!uunet!snorkelwacker!tut.cis.ohio-state.edu!ucbvax!proteon.com!sdb From: sdb@proteon.com (Scott D. Baker) Newsgroups: comp.sys.proteon Subject: Re: "Classic ProNET-80 Problems" Message-ID: <9003201958.AA18440@monk.proteon.com> Date: 20 Mar 90 19:58:51 GMT Sender: daemon@ucbvax.BERKELEY.EDU Organization: The Internet Lines: 61 Mike, I'd like to address your questions on what you term as the "classic" ProNET-80 problem. What's termed as the "data sensitivity" or "classic" problem is evident in ProNET-80 networks when one or more phase lock loop (PLL) circuits is mis- aligned. These PLL's are present on the CTL card and the fiber optic modems. 0xEB is a ProNET-80 near-worst-case data pattern. A nibble of 0xE is 001110 on the wire, a nibble of 0xB is 011100 on the wire. This concatenates to 001110011100001110011100, which has very poor transition density. As (differential) square waves on the wire you will have a near perfect sub-rate square wave: +--+--+--+ +--+--+--+ +--+--+--+ +--+--+--+ | | | | | | | | +--+--+ +--+--+ +--+--+--+--+ +--+--+ +--+--+ 0 0 1 1 1 0 0 1 1 1 0 0 0 0 1 1 1 0 0 1 1 1 0 0 <------0xE------> <------0xB------> <------0xE------> <------0xB------> There are two primary reasons for PLL's to be out of alignment: initial mis-alignment, or extreme variances in operating temperatures. What I have found to be the most effective way to troubleshoot these problems is to basically follow what Cliff had outlined: o "clear" the interface statistics on all devices attached to the 80 Mbps ring, o Perform some large file transfers over the 80 Mbps ring and note the interface statistics for the ProNET-80 interface. Specifically note the number of Input Parity Errors (IPE) and the number of Input Packets. The occurrence of IPE should not be greater than 1 IPE per 1000 Input Packets during optimal operation. o The node experiencing more than 1 IPE per 1000 Input Packets is indicating that the link is degraded between that unit and its upstream neighbor. Several types of links may be involved, including copper, fiber, wire centers, fiber optic modems, etc. o From that point we can apply what Cliff had suggested: isolate the components by swapping with other units and measure the results. In response to your question, "Is it true that Proteon, concentrating on FDDI replacing ProNET-80, does not give full effort in resolving the problem?". As a Technical Support Engineer, I can confidently respond with a definitive "Absolutely NOT !", and I certainly hope that our users feel the same way! Please feel free to call customer service at (508) 898-3100 if you would like to talk directly to a Technical Support Engineer. Scott Baker Technical Support Engineer Proteon, Inc.