Hari: Question regarding column versus word striping, R. Taborek response to S. Sabato
This note is a response to the note by Mr. Taborek to Mr. Sabato, dated
01/03/2000
Richard Taborek writes: " When Hari is scaled to higher speeds,
Column-Striping can function with smaller IPG than Word-Striping" and
(for word striping) "The minimum IPG size is 5n, where n=number of lanes".
Both of these statements are false. Mr. Taborek's argumentation is based on
a word striping scheme substantially different from the one presented by
Mark Ritter et al. at the November HSSG meeting which is the basis for my
reflector notes on the subject. The erroneous inferences result from
incorrectly applying column striping rules to word striping. In my last
Hari related note to Taborek, dated 12/27/99, I have described in detail
why
1) The framing and IPG specifics for a word striped interface is
independent of the number of lanes.
2) The IPG can be zero for extended periods of time. At boundaries of
frequency domains, there must be of course periodic Idles which can be
removed. The granularity for insertion or removal is a single word,
regardless of the number of lanes.
3) If a specific application (4-lane 10 Gbe or FC) allows a multitude of
comma words, we put them to good use for a variety of purposes. There is no
requirement for an End-of-Packet delimiter other than what Howard Frazier
proposed. If suggested changes cause any real problem, they can be dropped
with no significant impact.
Albert Widmer Phone: 914 945-2047 Email: widmer@xxxxxxxxxx
IBM T.J. Watson Research Center
Yorktown Heights, NY 10598-0218