Jump to content

Talk:Service data unit

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Talk

[edit]

I think "payload" is too vague a term to use, to describe the relationship of an n-1 SDU to an n-1 PDU. Saying that the SDU is the "payload" of the PDU may be truthful, but I don't think it expresses much about what is really going on, absent further clarification of what is meant by payload.

Contradiction?

[edit]

Background:

I understand & agree with these sentences from the article:

That is, the process of changing a SDU to a PDU, consists of an encapsulation process, performed by the lower layer. All the data contained in the SDU becomes enapsulated within the PDU.

The layer n-1 adds headers or footers, or both, to the SDU, transforming it it into a PDU of layer n-1.

For example, Layer 2 adds headers or footers, or both, to the (Layer 3) SDU, transforming it into a PDU of Layer n-1 (Layer 2).

But the following sentence seems to contradict this information:

The PDU at any given layer, layer 'n', is the SDU of the layer below, layer 'n-1'.

If this is correct, then it means, for example, that "The PDU at Layer 3 is the SDU of the layer below, layer 2". I believe this is backwards; the PDU at Layer 2 is the SDU of Layer 3. —Preceding unsigned comment added by 206.41.50.226 (talk) 16:59, 5 February 2011 (UTC)[reply]

Let me try to explain: Layer N sends a piece of data to layer N-1. This data contains information necessary for layer N addressing, error-checking, etc. , and so is a PDU regarding layer N. At the same time, this same piece of data is just payload to layer N-1 (carries no layer (N-1)-specific information), and so is a SDU regarding layer N-1. When layer N-1 encapsulates this data with its own additional info, the result becomes a layer N-1's PDU. 79.124.62.180 (talk) 16:32, 6 November 2013 (UTC)[reply]
Had a similar reaction as 206.41.50.226 initially. Pretty good description (unfortunately not obviously F/free images) here. Most expository image here. To me, it initially seemed that N-1 would "naturally" strip the N header in at least some cases, so that there could only be payload-plus-N-appropriate-headers (ie, only PDUs and the payload, no real concept of SDU) but apparently that doesn't [conventionally? ever?] happen. Rufwork (talk) 15:06, 23 October 2014 (UTC)[reply]