Error in "the switch book"

Error in "the switch book"

Post by Ariel V. Burbaicki » Mon, 26 May 2003 00:11:30



On the page 260 footnote 8 following
is written: "... and discard frames
whose Source Source Addresses have the
first set to 0.". 0 shoulb be IMHO
replaced by 1.

Yours sincerely
Ariel Burbaickij

 
 
 

Error in "the switch book"

Post by Rich Seifer » Mon, 26 May 2003 03:19:36




Quote:> On the page 260 footnote 8 following
> is written: "... and discard frames
> whose Source Source Addresses have the
> first set to 0.". 0 shoulb be IMHO
> replaced by 1.

You are correct, of course.

Anyone who wants the complete errata sheet for "The Switch Book", please

you. (The publisher has been unwilling to post the sheet to their web
site.)

--
Rich Seifert              Networks and Communications Consulting

(408) 395-5700            Los Gatos, CA 95033
(408) 395-1966 FAX

 
 
 

Error in "the switch book"

Post by Ariel V. Burbaicki » Mon, 26 May 2003 08:13:45





> > On the page 260 footnote 8 following
> > is written: "... and discard frames
> > whose Source Source Addresses have the
> > first set to 0.". 0 shoulb be IMHO
> > replaced by 1.

> You are correct, of course.

While we are on the issue. Could you explain
what do you mean with the word "session"
in 6.5.1? What "session" has lost on the 2
level of OSI? Normally session is somewhere
above  transport level in OSI in level
5. Or put it another way, where does the
route discovery process know from whether
it has to deal with one session or two
different sessions with same participating
parties?

Yours sincerely
Ariel Burbaickij

 
 
 

Error in "the switch book"

Post by Rich Seifer » Tue, 27 May 2003 07:50:50




Quote:> While we are on the issue. Could you explain
> what do you mean with the word "session"
> in 6.5.1? What "session" has lost on the 2
> level of OSI? Normally session is somewhere
> above  transport level in OSI in level
> 5. Or put it another way, where does the
> route discovery process know from whether
> it has to deal with one session or two
> different sessions with same participating
> parties?

Chapter 6 is about Source Routing, a (historic) technology used
exclusively in IBM Token Ring systems. In this model, applications ran
directly over the Link Layer (typically using an interface known as
NetBIOS); IBM's Token Ring LAN architecture did not have any Transport
or Network Layer. The network-layer functionality (i.e., routing) was
pushed down into the Link Layer, implemented as Source Routing; the
transport-layer functionality was also pushed into the Link Layer,
implemented as LLC Class 2 (connection-oriented LLC).

Thus, I am not using the term "session" in this Chapter in the strict
ISO-architectural sense. When a station wishes to communicate with
another station using Source Routing, it must first establish a path
through the Token Ring catenet. This is done at "session" establishment
time, typically when a user connects to a server.

In the ISO context, a user can establish multiple, simultaneous
"sessions" with the same server. With Source Routing, there would be
only one Route Discovery for all of these ISO-sessions. When a
subsequent ISO-session is established, the device would note that there
is already an established route through the catenet to the intended
target destination, and would not re-invoke Route Discovery.

Of course, the IBM architecture didn't really implement the ISO Session
layer (unless you consider the NetBIOS interface to be an ersatz Session
layer protocol).

--
Rich Seifert              Networks and Communications Consulting

(408) 395-5700            Los Gatos, CA 95033
(408) 395-1966 FAX

 
 
 

Error in "the switch book"

Post by Ariel V. Burbaicki » Tue, 27 May 2003 18:12:21


Thank you very much for the explanation.

Yours sincerely
Ariel Burbaickij