Talk:Internet Protocol/Archive 1

Fragmentation/reassembly confusion
[Discussion moved to Talk:IPv4 as it is no longer germane to this page after the reorg mentioned above/below. Noel 21:19, 21 Oct 2004 (UTC)]

Now I guess I should cut-n-paste this into the page! :-) Alas, since only IPv4 has fragmentation, I'm going to have to do the "move the packet format to the IPv4 page", since IPv6 doesn't have fragmentation in the basic header. Noel 20:38, 21 Oct 2004 (UTC)
 * does ipv6 have fragmentation at all and if not how are low mtu links supposed to be handled? Plugwash 22:57, 2 December 2005 (UTC)

Bund Marao...........................

I hate to add to the confusion, but I've added an exception to the Reliability section that says that IP doesn't reorder out of order packets. When IP fragments packets, if those fragments happen to arrive out of order, they are actually put back in order. While I fully realize there is a large section on fragmentation that was moved to the IPv4 page, the original IP page(which should also apply to v4) is lacking if this caveat/exception isn't mentioned. The exception isn't rare either, any large UDP data transfer is normally fragmented by IP.

If we needed to clarify my paragraph, stating it only applied to IPv4, I guess we could do this. But moving it altogether to the IPv4 page is just wrong, because it will leave people coming to the IP page only (I didn't even see IPv4 page initially) with the idea that IP *never* re-orders out of order packets, and this is not correct.

Kmwiki 19:43, 22 November 2006 (UTC)

IP Protocol article and link to TCP/IP Guide
Why revert the link to Kozierok's online book? (I tried to discuss this at User_talk:Kgfleischmann but nothing happened.) Lumpish Scholar 19:16, 30 April 2007 (UTC)

Fair use rationale for the Image:Ethereal ip header.png
Image:Ethereal ip header.png is being used on this article. I notice the image page specifies that the image is being used under fair use but there is no explanation or rationale as to why its use in this Wikipedia article constitutes fair use. In addition to the boilerplate fair use template, you must also write out on the image description page a specific explanation or rationale for why using this image in each article is consistent with fair use.

Please go to the image description page and edit it to include a fair use rationale. Using one of the templates at Fair use rationale guideline is an easy way to ensure that your image is in compliance with Wikipedia policy, but remember that you must complete the template. Do not simply insert a blank template on an image page.

If there is other fair use media, consider checking that you have specified the fair use rationale on the other images used on this page. Note that any fair use images lacking such an explanation can be deleted one week after being tagged, as described on criteria for speedy deletion. If you have any questions please ask them at the Media copyright questions page. Thank you.

BetacommandBot (talk) 21:54, 13 February 2008 (UTC)

Needs a table-diagram of the packet structure like on TCP and UDP
This page badly needs a packet layout image like on TCP and UDP.

I could have sworn there used to be one, but I don't see the deletion. —Preceding unsigned comment added by 128.61.125.102 (talk) 05:21, 11 February 2010 (UTC)


 * You'll find the image you're looking for at IPv4 or IPv6 packet. These are linked in the lead (IPv6 packet is two clicks away) but any suggestions/edits that would make this easier to find are welcomed. --Kvng (talk) 14:06, 20 September 2010 (UTC)

Wrong placed protocols
I think that RIP and especially OSPF are not placed correctly in Internet Protocol Suite scheme (table)3! An OSPF is protocol for filling and managing routing tables in routers. It is not a link layer protocol as it is written on this page! OSPF uses IP for exchanging Link State packets through autonomous system, so it has to be removed from Link Layer section of Internet Protocol Suite scheme (table)! BGP and RIP are not Application Layer protocols as placed in this scheme. Yes ,they use tcp/ip suite for transferring routing information via network, but they are not Application Layer protocols because no one application does really use them except router`s OS, and it (router`s OS) is not an application! You cannot use their services in your program!

I think that it would be better to create new section in table named for example ROUTING PROTOCOLS or something like that and put there all routing protocols OSPF,RIP,BGP and other protocols that are not especially other section protocols! —Preceding unsigned comment added by 93.87.248.72 (talk) 18:43, 22 September 2010 (UTC)


 * No, the OSI model does not apply to TCP/IP. The Internet model does not use encapsulation as a hierarchical criterion, it uses scope of operation for layering. Application is process-to-process, transport is host-to-host, Internet Layer is network-to-network, and Link Layer is link-restricted. Some routing protocol packets are routed across the Internet as any application, some, like OSPF, are not. Kbrose (talk) 18:51, 22 September 2010 (UTC)


 * TCP/IP Basic Suite was developed at the same time as the OSI Model. Therefore it is not possible to explain TCP/IP exactly by the OSI Model because TCP/IP did not follow OSI. E.g. in OSI it is not allowed to implement the functionality of a certain layer by using functionality of a higher layer. In the special case of IP routing protocols you can see it: RIP is based on UDP transport, BGP is based on TCP transport; only OSPF is based on IP transport. Therefore I would recommend to locate all IP routing protocols to layer 3 with a remark that those protocols do not strictly follow the OSI model. — Preceding unsigned comment added by 178.191.233.12 (talk) 12:39, 3 June 2011 (UTC)

Packets/datagrams
'packets or datagrams (the terms are basically synonymous in IP)' can someone explain why? Helix84 15:31, 11 Jan 2005 (UTC)


 * Does clicking on the link to Packet, and reading the article there, help?

Noel (talk) 21:13, 11 Jan 2005 (UTC)


 * I've reinforced links to network packet and datagram, and revised both articles for clarity. They could still use some polish, though. Stephen Charles Thompson (talk) 16:30, 24 January 2012 (UTC)

Quality overrated: many fundamental problems
I believe the quality of this article is vastly overstated. I am citing several fundamental problems with this article that degrades its quality. Please join me in responding to this Talk section as these problems are abated. end of initial comments Stephen Charles Thompson (talk) 16:06, 24 January 2012 (UTC)
 * insufficient reference or link to IP datagram or application of IP addressing
 * I've reinforced the introduced of the topics of datagram structure, address labeling, and headers. I believe it should be treated a bit more in-depth, but succinctly, with references to a main article. Stephen Charles Thompson (talk) 16:28, 24 January 2012 (UTC)
 * Scope of Function section seems to overlap with section IP addressing and routing with no apparent distinction.


 * I agree that quality is overrated. I have knocked it down a notch. --Kvng (talk) 13:57, 26 January 2012 (UTC)

Navbox explosion
I have reverted user:Funandtrvl's revision that added two more RHS navboxes and removed most of the content from the See also section. I agree that all of the navboxes are relevant to the topic but I think it would be best to just choose the best one. See Talk:IP_address for a related discussion of See also contents. -—Kvng 15:51, 24 December 2012 (UTC)

lowercase and uppercase
The word "internet" as used in most of this article must use a lowercase i just as used in RFC 791, unless, of course, it's the first word of a phrase or otherwise emphasized. An internet (common noun) is an assemnly of LAN working as defined in the RFC. The Internet (proper noun) is the unique worldwide network that is using this technique. --A Pirard (talk) 19:10, 10 June 2013 (UTC)


 * See Capitalization of "Internet" for a discussion. ~KvnG 14:01, 15 June 2013 (UTC)

OSIModel
Should we add the {OSIModel} thing to this page? It's on some of the others. BrokenSegue 22:23, 21 Sep 2004 (UTC)


 * No reason not to. Noel 14:44, 4 Oct 2004 (UTC)


 * Yes JustforfunJW (talk) 21:42, 24 October 2016 (UTC)


 * No, the Internet protocol suite uses a slightly different and simplified networking model. This model is summarized by IPstack which is already on this page. Also including without giving some context OSIModel may cause confusion. ~Kvng (talk) 18:57, 27 October 2016 (UTC)

removal of link to further read
I added a link to The world in which IPv6 was a good design, because it is the most comprehensive, compact and best writing about the history of the Internet and why it is what it is that I've ever read. I think it is understandable even to non-experts or complete newbies to the topic, so I think adding the link as further read would significantly improve the article about IP as a whole (especially given the fact that some people already complained about the article being too technical and not explaining to them what IP is).

My edit, however, was undone, and the given reason was "WP is not a link farm". I know that, and I added the link not because it's just another article about IP but because of the high quality of its' content. I don't want to start a edit war and undo the revert, so I'm asking the other Wikipedians: What do you think about this? Please also keep in mind that the article I linked is easily understandable by non-experts (at least it seems so to me).

Here is a link to the revert: https://en.wikipedia.org/w/index.php?title=Internet_Protocol&oldid=prev&diff=795632434 — Preceding unsigned comment added by GGShinobi (talk • contribs) 15:15, 19 August 2017 (UTC)


 * This is a self-published blog post so not something we would normally use as a source and so that makes it hard to justify as a resource we'd want to recommend to readers. It is easy to read but rambles and I'm not sure the point it is trying to make or what it adds in the context of this article. ~Kvng (talk) 14:25, 23 August 2017 (UTC)


 * Thank you for your reply. The article explains very well why IP (and the Internet) is what it is. It tells it's history. It gives the readers some important background that helps them understand this article. I don't know any source that does this better. That's why I wanted to add it as "further read". But I also understand the reasons you named why it would be unusual to recommend it to Wikipedia readers, so I can accept that my edits were reverted. GGShinobi (talk) 14:59, 23 August 2017 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on Internet Protocol. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20060701130628/http://www.gcn.com/print/25_16/41051-1.html to http://www.gcn.com/print/25_16/41051-1.html

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 13:25, 15 November 2017 (UTC)

Doesn't make sense to conflate TCP and UDP
In the first section, the terms "packet" and "datagram" are used interchangeably, something which is incorrect and will confuse both the beginner (who only saw the term "packet" at the beginning of the article, and doesn't know what a datagram is) and the knowledgeable person (who knows the difference between packets and datagrams and doesn't understand why they are being confused with each other). Packets and datagrams are not interchangeable; the former refers to TCP/IP and the latter to UDP/IP. Since the distinction between UDP and TCP is not mentioned earlier in the article at all, only TCP is mentioned, and clicking the link to the TCP page will only mention packets, not datagrams, and moreover at the top of the page there is only a link to the Wikipedia page for packets, not for datagrams, it does not make sense to mention datagrams. Maybe if it is mentioned in the introduction that IP can also work with UDP and datagrams would it make sense to mention datagrams later in the article, but right now that is not done. Even if UDP had been mentioned during the introduction, when explaining the concepts, a given paragraph, while explaining an example in depth, should choose one of the two, TCP and packet, or UDP and datagram, and stick with it. Not doing so confuses both people new to the subject (who don't know what a datagram is and don't know that it is similar/analogous to a packet) and people with more knowledge about it. — Preceding unsigned comment added by 2620:83:8001:572:0:0:1:8AF (talk) 22:39, 30 May 2018 (UTC)
 * It seems editing over the last few months has changed several "datagram" to "packet" although the former is correct for IP. I restored most of an earlier version. I'm not sure about the best way to fix the text but IP delivers datagrams for an upper layer. IP may fragment a datagram into several packets each of which may travel different paths and be re-assembled to a datagram by IP at the destination. UDP provides port numbers for applications. Johnuniq (talk) 01:21, 31 May 2018 (UTC)
 * Datagram and Network packet (also Frame (networking) and Protocol data unit) could use some love. We're not going to be able to permanently fix this terminology if we can't clearly explain the differences. I fear that things are not as cut-and-dried as what lays out here. Engineers are not that careful about terminology and so the authoritative sources don't necessarily agree. ~Kvng (talk) 13:56, 2 June 2018 (UTC)

Transmission Control Program
I have - again - corrected the information on who was involved in the Transmission Control Program. Bobn Kahn was a contributor, not an author, so could not be said to have introduced it. I even gave the link to the information, yet both my correction and link with the evidence were removed. So, I am creating this talk section in case someone really feels that the formal Transmission Control Program specification is not the definitive source for information on the Transmission Control Program. Perhaps they can explain why, rather than make the article less accurate again by vandalising my improvement. Varybit (talk) 16:11, 17 March 2019 (UTC)
 * You should read the rest of the article, not just the lead. The article by Kahn and Cerf, provided in the article, predates the RFC. Kbrose (talk) 18:03, 17 March 2019 (UTC)
 * You should read the sentence. The RFC that established the Internet Control Program was certainly helped by Kahn 'and many others' but he was not involved in drawing up the specification. Please quote sources that support your reversions before indulging in an edit war. Varybit (talk) 10:41, 20 March 2019 (UTC)
 * The "original specification" referred to was published several months before the RFC. Please stop making the false claim that the RFC was the original specification. Shamino (talk) 15:43, 23 March 2019 (UTC)
 * I see that Kbrose just added some necessary clarifications. The original "Transmission Control Program" was a single stack containing what we today would consider both L3 and L4.  It had a datagram service (called "IP") and a stream service (called "Transmission Control Protocol").  "TCP" refers to this stream service, not the complete stack.  It is the complete stack that Cerf and Kahn wrote about in "A Protocol for Packet Network Intercommunication", which is cited at the start of the "Version History" section.  It was a few years (see the link to IEN 2) before L3 and L4 were formally split up, ultimately leading to the stack we're familiar with today.  Shamino (talk) 15:36, 25 March 2019 (UTC)

IPv9 exists and it is TUBA and it was affiliated with an official standards body
TUBA was reserved with IANA version number IPv9, see: https://www.iana.org/assignments/version-numbers/version-numbers.xhtml — Preceding unsigned comment added by 61.68.21.230 (talk) 09:04, 24 July 2019 (UTC)


 * Added ~Kvng (talk) 16:30, 25 March 2020 (UTC)

History outside of IPv4 and IPV6
The history section jumps straight into IPv4 and IPv6 with no mention of IPv1, IPv2, IPv3, or IPv5. If they do not exist, it would be helpful to at least indicate why the jump to version 4 and 6 without a version before or in between. --121.45.141.19 (talk) 03:44, 2 February 2011 (UTC)
 * Or to at least not redirect from non-notable or included protocols to the Internet Protocol article. 174.30.114.102 (talk) 06:36, 10 December 2012 (UTC)

It appears that some historic documents do exist at the IETF with these. In particular, I was able to find: What would be a good way to present this on the Wikipedia page? Shamino (talk) 19:36, 22 May 2014 (UTC)
 * IEN 2, dated August 1977, describes the need to separate the TCP and IP protocol functionalities. It proposes the first version of the IP header, which uses version 0 in the packet-header.  This may, in fact be version 1, since I could not find a file in the IEN archive describing a header with 1 as the version in the header.
 * IEN 28, dated February 1978, describes version 2 of the IP header.
 * IEN 41, dated June 1978, describes what claims to be version 4 of the IP header, but it differs significantly from today's IPv4. Given that there are no other versions published in the IEN archive between this one and IEN 54 (below) I think you could make the claim that IEN 41 is in fact describing IPv3, even though its packet format uses a 4 for the version field.
 * IEN 54, dated September 1978 also claims to describe IP version 4. The header presented in this document, however, matches the modern IPv4 header (with TOS bits in place of DHCP, of course.)  This may be the first published appearance of the IPv4 header we all use today.
 * It would excellent if we could also find some WP:SECONDARY sources. Treatmented in secondary sources should guide how we treat it here. ~KvnG 14:52, 28 May 2014 (UTC)

There are a number of versions of this story around the internet - Where is IPv1, 2, 3 and 5?. Obviously not a primary source but they do suggest that IP v1 and v2 were actually TCP v1 and v2. The link above for IEN 2, dated August 1977, looks like Jon Postel's suggestion that the protocol be split. Parsley72 (talk) 08:08, 24 November 2014 (UTC)


 * Not exactly a WP:RS but I added it anyway and updated the dates to match. ~Kvng (talk) 16:37, 25 March 2020 (UTC)

Help this is too technical!
this looks like a great detailed reference, but I just want a simple explanation of what IP is (and TCP, and what's the difference). Too much jargon for the non-pro! —The preceding unsigned comment was added by 130.195.5.7 (talk) 02:33, 22 January 2007 (UTC). That sort of thing goes on the simple English version of Wikipedia. I suggest reading all the linked articles to learn about it though. Bno112300 (talk) 05:06, 13 February 2013 (UTC)


 * The lead is more accessible now than it was. ~Kvng (talk) 16:42, 25 March 2020 (UTC)