Talk:Frame (networking)

Remove redirect from Data Frame

There really ought to be an article on the "data frame" data structure that is becoming common in scientific/statistical computing. R has had the concept since forever, but now that it's spreading into Python (via Pandas) and Julia (via DataFrames.jl) I think its time it had its own page. Anyone opposed to a disambiguation page instead of a redirect, as a placeholder until I or someone else writes an article on the other kind of data frame?

8.37.179.150 (talk) 15:51, 11 October 2018 (UTC)

I agree. Qzekrom (she/her • talk) 23:54, 13 January 2022 (UTC)

Any editor may convert the Data frame redirect to an article on the computing concept. We'll figure out what hatnotes or disambiguation are required once that's done. ~Kvng (talk) 15:10, 18 January 2022 (UTC)

All the hits in a Google search are to explanations of the R language. We could link to the R article. Gah4 (talk) 18:31, 18 January 2022 (UTC)

It is not a great option but not out of the question. There is one mention and a terse explanation there, ...data frames (similar to tables in a relational database)... ~Kvng (talk) 22:40, 18 January 2022 (UTC)

Yes, I was hoping it would say more. It would be nice to redirect to an actual explanation. Some years ago, I did some work with R, and my favorite comment about it is that (at the time) I still didn't understand data frames. They are a convenient structure to put all the important things for some data together, especially including row and column headers. When you use one, your plots come out with the right labels on the axis, without much extra work. But they are also confusing when you want to do something a little different from the ordinary (which I usually did). It might be worth its own article, or maybe just a section with a good description in R_(programming_language). Gah4 (talk) 00:24, 19 January 2022 (UTC)

Really confusing article

After reading this article I am really confused. I simply dont know what a frame is.

• It says that " (...) a frame is a packet (...)". To me it is not! And not according to the article about [packet].
• The illustration hints that frames are of fixed size, but the article does not say that. I also hints that a frame is definitely not a packet.
• And of what size is a packet?

I believe this article needs a total makeover, untill then stick to the article [packet].

Velle 10:19, 17 August 2006 (UTC)

Comparison to packet

It's not a reliable source, but this post offers a reasonably convincing explanation. It suggests that the term "frame" derives from framing bits used to demarcate a time slot during which a set of bits are sent. (it goes further, suggesting that a "frame" isn't really a frame once it's been statically captured, but instead "frame" only refers to a specific "time frame" during which a set of bits are sent). I won't add that in the article, since IANACCNA, and the article is already sorely lacking references for what's currently there.

Anyway, the standard breakdown of:

• hub = layer 1 = sees things only at the bit-level
• switch = layer 2 = sees things only at the frame-level
• router = layer 3 = sees things only at the packet-level

is clear enough at least. A frame is something that's closer to what's actually happening on the wire, but is higher-level than bit-at-a-time.

(also, the NASA image discussed above was removed at some point, but it may still be marginally helpful) --Twinkie Assassin (talk) 17:18, 13 February 2008 (UTC)

-- Dc sharp (talk) 19:09, 21 September 2011 (UTC) Agree- specifically with respect to mixing "frame" and "packet" -- http://en.wikipedia.org/wiki/Protocol_data_unit says that an OSI Layer 2 PDU is a "frame" and a Layer 3 PDU is a "packet". But then when you follow the link to this "frame" page, it says that a frame *is* a packet. Either this page needs to not equate packet and frame, or the PDU page needs to be updated.

-- The OSI model is not the end-all be-all definition of network architecture. Please consider the TCP/IP model as well. The current version of the network packet article is entrenched in OSI terminology, but needs to be revised to include TCP/IP definitions as well. Stephen Charles Thompson (talk) 19:58, 24 January 2012 (UTC)

I agree with Velle and Stephen Charles Thompson that saying a frame "is a" packet is confusing and misleading, and attempted a fix.

If I'm reading the Ethernet frame article correctly,

an Ethernet transmitter takes an Ethernet frame, and wraps it in more bits, forming an Ethernet packet, and then sends that packet over the wire.

If I'm reading the OSI model and Internet Protocol articles correctly,

the transmitter takes a packet, and wraps it in more bits, forming a frame, and then that frame is sent over the wire.

Is one of these articles exactly backwards and needs to be fixed, and if so, which one?

Or do the Ethernet people really use exactly the opposite definitions of "frame" and "packet" from the OSI and Internet people? --DavidCary (talk) 19:47, 24 December 2015 (UTC)

The Ethernet people (IEEE) do like to use frame for Ethernet transmission units. OSI and IETF prefer packet or datagram. These preferences can be explained by the different technologies the different groups work with but it can also be explained as a cultural thing. I don't think the reliable sources agree on this. ~Kvng (talk) 16:23, 27 December 2015 (UTC)

Multiple overlapping articles

We have at least four distinct articles covering the same basic topic:

• Frame (networking)
• Datagram
• Network packet
• Protocol data unit

While I appreciate that there may be technical differences between the different terms, I don't think we do readers a good service by having multiple articles on the same general topic. I don't feel so bold as to perform or formally propose a merge but I would like to discuss what we can do to improve WP coverage in this area. --Kvng (talk) 17:39, 3 February 2012 (UTC)

I came here after reading in 802.11 the statement The datagrams are called frames. My understanding, mostly from Ethernet, is that frames are at the hardware (layer 1) level, and packets or datagrams in software, or layer 2 and higher. (Not completely obvious, as some of layer 2 is in hardware.) I believe that means that we need at least Frame (networking) separate from the rest. Each should concentrate on its part, and link to the others as much as possible when that makes sense. Gah4 (talk) 06:27, 30 January 2017 (UTC)

Ethernet transmissions are definitely done in "frames" at least according to IEEE terminology. Other networking technologies also use this terminology, sometimes in different ways: Frame Relay, Fiber Distributed Data Interface, Token Ring... Some editors want to beleive that there is a simple mapping of terminology to network technology or protocol layers but, the fact is there is not good consensus or consistency in reliable sources about which terms are used in which contexts. I'm more convinced in the 5 years since my original post that readers would be best served by a single article that describes a transmission unit on a packet-based network and describes how the different terminology is used in different contexts. ~Kvng (talk) 15:21, 7 February 2017 (UTC)

I suppose I didn't say what I meant quite right. I believe it is ethernet frames, carrying IP packets or datagrams. At the frame level, it includes preamble, framing bits, (start/stop in asynchronous protocols), CRC, and other things that the software never sees. Gah4 (talk) 17:19, 7 February 2017 (UTC)

"Networking" and "telecommunications" frames?

This page was renamed from "Frame (telecommunications)" to "Frame (networking)" because "Article currently only references frames in networking, not in any other form of telecom.", but, at the time it was renamed, the third paragraph said:

In telecommunications, specifically time-division multiplex (TDM) and time-division multiple access (TDMA), a frame is a cyclically repeated data block that consists of a fixed number of time slots, one for each logical TDM channel or TDMA transmitter. In this context, a frame is typically an entity at the physical layer. TDM application examples are SONET/SDH and the ISDN circuit switched B-channel. TDMA examples are the 2G and 3G circuit switched cellular voice services. The frame is also an entity for time-division duplex, where the mobile terminal may transmit during some timeslots and receive during others.

which sounds as if it's discussing frames in "other forms of telecom". Is that not the case? If it is the case, should the third paragraph be moved to "Frame (telecommunications)", and, perhaps, see also hatnotes added to this page and "Frame (telecommunications)", to point people to the other page (i.e., splitting this page into a page for the two types of framing)? Guy Harris (talk) 19:27, 10 March 2014 (UTC)

Hm, that's a good point. Frames have somewhat different meanings in computer networking and general telecommunications, though they still represent pretty much the same thing – a unit used for reserving time and bandwidth. Having that in mind, I'd say that splitting off another "Frame (telecommunications)" article would be pretty much an overkill, especially considering the amount of content currently available. I'd suggest we rename this article to "Frame (data communication)" (or something similar) and call it a day. :) — Dsimic (talk | contribs) 22:22, 10 March 2014 (UTC)

Instead of twiddling titles, let's discuss how to better organize this whole topic (See above). ~KvnG 22:35, 10 March 2014 (UTC)

I have expanded the lead to discuss the use if frames in TDM systems. It does not appear that a separate Frame (telecommunications) article is required. A frame serves the same purpose in TDM and packet switching applications. ~Kvng (talk) 00:40, 26 February 2018 (UTC)