Unidirectional Low-Level Data-Pattern On Ethernet Link

[IFG][Pre-Amble][Destination Address][Source Address][Type/Length] [Payload][CRC][IFG]...

I would like to know if the diagram above is an accurate characterization of the data stream on an Ethernet link.

I would also like to know if the wire is typically packed with this structure on optimized links.

The answer is not super important. I work best when I can carry around a visualization of what is going on, and rather than have an incorrect visualization, I thought I would simply find out.

-Le Chaud Lapin-

Reply to
Le Chaud Lapin
Loading thread data ...

Not quite. There is no proceeding IFG. The way you have it would imply that there would be two IFG between adjacent packet contents. The IFG only follows packets.

No, typically packet transmission is asynchronous. If it -happened- that one particular system had a bunch of packets to send and no other system had any to send (including no acknowledgements), then you could end up with a temporally local packing of packets on the wire. And a full duplex link that had a stream of packets to send could temporarily look as if it were a packed set of the above.

Google ethernet frame structure

Reply to
Walter Roberson

Actually, it's the other way around. The IFG is enforced by senders as an idle time *before* transmission, not after. (But your conclusion that there need be only one, minimum-length IFG between frames is still correct.)

Well, "temporary" could be a rather long period of time, especially with bulk file transfers between machines capable of saturating the link (which are fairly common these days).

-- Rich Seifert Networks and Communications Consulting 21885 Bear Creek Way (408) 395-5700 Los Gatos, CA 95033 (408) 228-0803 FAX

Send replies to: usenet at richseifert dot com

Reply to
Rich Seifert

The diagrams I've seen have always placed the IFG after the packet. But perhaps what we have is a semantic difficulty.

If I understand correctly, when a system is ready to send a packet and submits the packet to its low-level ethernet layer, that that layer is permitted to immediately begin transmitting if that layer has been monitoring and there has been at least the minimum gap time of silence. I tend to think of that as the gap -following- the packet: when the gap is described as being before the packet, I would tend to interpret that as the transmitter having to wait that gap time even though there has been enough silence on the link.

But of course it isn't the case that anything is transmitting some kind of "silence signal", so whether the silence is "owned" by the transmitter or by something else is not a correct question. Possibly it is not a correct question to describe the IFG as belonging to the frame format at all (but nearly everyone does.)

Reply to
Walter Roberson

Not true; see below.

And that is precisely the case.

It is not "owned" by the transmitter, but it is "enforced" by the transmitter.

Once a station has decided that it will transmit (i.e., after assessing carrier sense following the submission of a frame by some higher-layer entity), it waits an IFG and then begins its transmission. Thus, the IFG is "enforced" (caused to be present on the channel) by the sending station at the start of each and every frame transmission. The algorithm does not allow it to start transmission immediately, even if there has been a long idle period prior to its determination that it can now send a frame.

(Of course, the decision whether to actually transmit at the end of the IFG is usually revocable if carrier sense reasserts during the first 2/3 of that IFG, but that is irrelevant to the question of whether IFG is enforced at the start or end of transmission.)

The IEEE standard includes it as part of the frame format definition; you cannot properly send a frame without including the IFG.

-- Rich Seifert Networks and Communications Consulting 21885 Bear Creek Way (408) 395-5700 Los Gatos, CA 95033 (408) 228-0803 FAX

Send replies to: usenet at richseifert dot com

Reply to
Rich Seifert

Well, time *before* the next transmission or time *after* the previous transmission doesn't make much difference, especially if both are from the same source.

-- glen

Reply to
glen herrmannsfeldt

Right. I put two IFG's (instead of one) to indicate where repetition would begin, assuming of course tight packing of frames (transmitter always has frame in serializer).

For those who are interested, I calculated that, assuming optimal adjacency of frames, with full 1500 bytes of payload for frame using "standard" format, the effective bit rate is

1500 / [8+6+6+2+1500+4] = 98.296%

of indicated bit-rate.

-Le Chaud Lapin-

Reply to
Le Chaud Lapin

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.