Article View: comp.dcom.lans.fddi
Article #177Re: What's difference between Scrub & Strip
From: Paul Koning
Date: Thu, 10 Sep 1992 15:20
Date: Thu, 10 Sep 1992 15:20
62 lines
3881 bytes
3881 bytes
In article <1992Sep9.162240.25518@webo.dg.com>, mike@vulcan.webo.dg.com (Mike Fugere) writes: |>I am reasonably sure that the new FDDI chipset offering from |>Motorola can be set to automatically perform frame scrubbing. |>This feature can also be disabled. If I am wrong, perhaps |>someone can correct me on this.... It would then be up to |>the hardware and firmware guys designing the FDDI network |>adaptor to utilize this feature in their products or not. |> |>Mike Fugere |>Design Engineer/Aviion Systems Development |>Data General Corporation |>Westborough, MA |>"Of course, my employer has absolutely nothing to do with anything I say..." Unfortunately, there are a number of similar-sounding terms around to describe very different functions. You were caught by one of them. What the Motorola chipset can do is "ring purging". This function -- usually enabled in Digital products -- indeed removes fragments left by the Strip function. Its main purpose, though, is to remove "no owner" frames, which are things that can cause real trouble. You can find more information about the ring purger in the following paper: Proceedings of 16th Conf on LCN, Oct 14-17, 1991, "A Ring Purger for the FDDI Token Ring", by Henry Yang and KK Ramakrishnan, page 503-514. Fragments are harmless things, since they consume no ring bandwidth. (This may seem strange at first. You can prove it to yourself by drawing diagrams of where they are found on the ring with respect to the token. This will show that the space taken up by fragments does not take away from real traffic; instead, that space would be filled with Idles if it weren't filled with a fragment.) About the only annoyance caused by fragments is that they MAY consume some adapter, and maybe even host bus, bandwidth. Typical FDDI chipsets will internally buffer and discard "short" fragments -- those caused by stripping. Fragments can also be caused by bit errors; such fragments may be any length. If they are longer than whatever the FIFO size is between the FDDI chipsets and memory, the data from the fragment will start to move into memory. Depending on the adapter design, that may or may not matter at all. Incidentally, there is another function that will eliminate fragments: the Smoother (in the PHY) will remove the last symbol or byte from a fragment being repeated. (At least this is required in some cases, and some implementations do it even where it is not required.) This causes fragments to shrink and eventually disappear. So what is scrubbing? It's neither stripping nor purging; it is a function of the PHY, used at station insertion and removal. The purpose of scrubbing is to ensure that a configuration change doesn't leave any no-owner frames on the ring. One reason this is important is to make sure that such a frame isn't received over and over by the station(s) it is addressed to. A second, more subtle, reason is to put a bound on packet lifetime, which is important to the correct operation of higher layer protocols. There are several ways to do scrubbing; a commonly used one is for the PHY to send Idle symbols for a time period sufficient to ensure that nothing old is left on the ring. Sending for D_Max will do that. paul
Technical Details
Message-ID:
<1992Sep10.152047.641@nntpd.lkg.dec.com>Path:
arc1-ca.pugleaf.net!sparky!uunet!decwrl!deccrl!news.crl.dec.com!news!nntpd.lkg.dec.com!usenetReferences:
<1992Sep2.153938.615@hls.com> <1992Sep9.162240.25518@webo.dg.com>Show Full Headers
Complete Article Headers
Newsgroups: comp.dcom.lans.fddi Path: arc1-ca.pugleaf.net!sparky!uunet!decwrl!deccrl!news.crl.dec.com!news!nntpd.lkg.dec.com!usenet From: koning@koning.enet.dec.com (Paul Koning) Subject: Re: What's difference between Scrub & Strip Message-ID: <1992Sep10.152047.641@nntpd.lkg.dec.com> Lines: 63 Sender: usenet@nntpd.lkg.dec.com (USENET News System) Reply-To: koning@koning.enet.dec.com Organization: Digital Equipment Co., distributed systems architecture References: <1992Sep2.153938.615@hls.com> <1992Sep9.162240.25518@webo.dg.com> Date: Thu, 10 Sep 1992 15:20:47 GMT