[rprsg] T&D Straw Poll #2 (discussion phase)
A key terminology issue has been the description of frames entering and
leaving the ring from/to the MAC client, frames entering and leaving the
station from/to the ring, and frames passing through the station while
circulating on the ring. I have collected candidates for these terms
(in three groups) below. Please post your additional candidate
definitions and provide your comments on the T&D reflector. If there
are no objections, I will conduct a straw poll starting at the end of
day tomorrow (allowing discussion today and tomorrow). The straw poll
is, of course, nonbinding.
Q1) Terms describing frames entering the ring from the direction of the
MAC client and exiting the ring towards the MAC client (respectively):
a) ingress, egress
b) ring ingress, ring egress
c) entering, exiting
d) add, drop
e) insert, copy
f) transmit, receive
g) host transmit, host receive
Q2) Terms describing frames received from the ring by a station and
frames transmitted to the ring by a station (respectively):
a) receive, transmit
b) ring receive, ring transmit
c) ring in, ring out (used by 802.5)
d) inbound, outbound
e) arriving, departing
Q3) Terms describing frames transiting the station via the ring:
a) transit
b) thru
c) passthru
d) pass
e) continue (used in combination with (add, drop) for consistency with
SONET
f) repeat (used by802.5)
Some reasonable combinations might be:
a) ingress, egress, ring-in, ring-out, repeat
b) add, drop, inbound, outbound, continue
c) insert, copy, ring receive, ring transmit, thru
Some considerations:
The terms 'transmit' and 'transit' look alike. After living with this
for a while, I think the problem is real and I suggest that we don't use
'transit'.
The terms 'transmit' / 'receive' are overloaded. They are used to
indicate the placement / interpretation of signals on the medium AND the
transfer of SDUs across protocol layers. It follows that these are not
the best terms to distinguish ring circulating traffic from traffic
to/from the MAC client. The terms could be qualified by 'host' /
'ring', but this has the potential for confusion.
Cisco currently uses the terms 'ingress' / 'egress', but with the
opposite orientation. For them, ingress traffic enters the router from
the MAC and egress traffic exits the router towards the MAC (Cisco
folks, did I get this right?). So, this may be confusing. On the other
hand, this may not be a problem if there is agreement that 'ingress' /
'egress' are always understood to be relative to the layer of interest.
If we wanted to be consistent with 802.5, I think we would be using
(receive, transmit, ring-in, ring-out, repeat)... can the 802.5 experts
confirm this?
Bob