🚀 go-pugleaf

In article 23@inode.com, andor@inode.com () writes:
> Has anyone out there been using/testing the new 100Mbps ethernet? 
> 
> Opinions? Feedback? How well does it work? What kind of thruput do you get?
> Collision rates? etc...
> 
> 
> --- Blue Wave/QWK v2.10
> #! rnews 2

Considering the fact that the IEEE Working Group has just begun to work on a
standard, anything in the field at this point is at best a "contender" - there
is no real product when there isn't even a Draft Standard...

	- John

---
+---------------------------------------------------------------------+    
|  John K. Scoggin, Jr.			Email: scoggin@delmarva.com   |
|  Supervisor, Network Operations       Phone: (302) 451-5200         |
|  Delmarva Power & Light Company       Fax:   (302) 451-5321         |
|  500 N. Wakefield Drive               NOC:   (800) 388-7076         |
|  Newark, DE 19714-6066		                              |
|  The opinions expressed are not those of Delmarva Power, simply the |
|  product of an over-active imagination...                           |
+---------------------------------------------------------------------+

Considering that the people doing this have set some ridiculous
up-front requirements, like no change in frame format or size, they
are already seriously curtailed in what they can acheive.  1500 bytes
was too small a frame at 10 mbps, it will be ridiculous at 100 mbps.
For high bits/second throughput, big frames are critical.  The
software overhead is mostly per frame, not per byte.

This silly requirement for no frame format change comes from the fact
that the PC software industry has (semi-)standardized on NDIS and ODI,
which mean that every protocol stack has to understand every MAC frame
format.  If the world had standardized on the Packet Driver, which
interfaces transparently at the bottom on the Network layer, then
adding a new class of driver with a larger frame size would have been
trivial, and transparent to applications.

The same problems happens in the STREAMS DLPI interface.  Token-Rings
and FDDI often have to masquerade as Ethernet, since the clients only
support Ethernet.

In article <JAS.93Jan26145441@mcgarret.proteon> jas@mcgarret.proteon (John A. Shriver) writes:
>
>This silly requirement for no frame format change comes from the fact
>that the PC software industry has (semi-)standardized on NDIS and ODI,
>which mean that every protocol stack has to understand every MAC frame
>format.  If the world had standardized on the Packet Driver, which
>interfaces transparently at the bottom on the Network layer, then
>adding a new class of driver with a larger frame size would have been
>trivial, and transparent to applications.
>
>The same problems happens in the STREAMS DLPI interface.  Token-Rings
>and FDDI often have to masquerade as Ethernet, since the clients only
>support Ethernet.

I have to point the above is in error, NDIS is MAC Header aware, ODI is not
nor has it ever been and is designed to isolate Protocol Stacks from
LAN Drivers and visa versa.  Protocol Stacks just utilize the WorstDataSize
field of the MLID Config. Table to see how much they can Tx/Rx, etc.

I suggest you download the PKZIP'ed postscript files from odi/specs/
on sjf-lwp.sjf.novell.com and see how ODI works.

By the way, a properly written ODI Protocol Stack can work on top of any
underlying Media, ie. Ethernet, Token-Ring, FDDI, ARCNet, etc. transparently
and does so currently. 

David_Murphy@novell.com         All opinions are my own, not the employer's
                                And I am not a support person either, so 
                                support questions should go to support people.

In article <3159366228.23@inode.com>, andor@inode.com writes:
> Has anyone out there been using/testing the new 100Mbps ethernet? 
> 
> Opinions? Feedback? How well does it work? What kind of thruput do you get?
> Collision rates? etc...
> 

Hmmm... I just got back from the 100 Mb/s Ethernet standards study group
meeting...

FYI: No one has announced any products for 100 Mb/s Ethernet. There are
no prototypes or beta test units in the field for any vendor. A few
vendors have breadboard equipment operating in their labs.

At this time, 100 Mb/s Ethernet is "paperware". Watch for announcements,
though.


-- 
Rich Seifert                    Networks and Communications Consulting
seifert@netcom.com              (408) 996-0922
                                (408) 996-2860 FAX
"... specialists in Local Area Networks and Data Communications systems"

In article <JAS.93Jan26145441@mcgarret.proteon>, jas@mcgarret.proteon (John A. Shriver) writes:
> Considering that the people doing this have set some ridiculous
> up-front requirements, like no change in frame format or size, they
> are already seriously curtailed in what they can acheive.  1500 bytes
> was too small a frame at 10 mbps, it will be ridiculous at 100 mbps.
> For high bits/second throughput, big frames are critical.  The
> software overhead is mostly per frame, not per byte.
> 

There have been no such requirements set by the study group. The group
has just barely begun to work on the OBJECTIVES for a standard.
I am sure that maximum frame size will be discussed.



-- 
Rich Seifert                    Networks and Communications Consulting
seifert@netcom.com              (408) 996-0922
                                (408) 996-2860 FAX
"... specialists in Local Area Networks and Data Communications systems"

If anyone's really interested, there's a pretty good article on some of
the 100-Mbps ethernet proposals in the November 16, 1992 issue of "Open Systems 
Today".  As Rich ETAL said, there ARE no current implementations, just
standards proposals.  Personally, I lean t'words changes at the MAC layer,
to more logically support real time services (like video) and such,
but as it's been pointed out, then you don't really have ethernet...


Mike Simon				simon@moscow.uidaho.edu
Computer Science Dept.
Moscow, Idaho