Talk:Internet protocol suite

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Computing / Networking (Rated B-class, High-importance)
WikiProject icon This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
B-Class article B  This article has been rated as B-Class on the project's quality scale.
 High  This article has been rated as High-importance on the project's importance scale.
Taskforce icon
This article is supported by Networking task force (marked as High-importance).
WikiProject Internet (Rated B-class, Top-importance)
WikiProject icon This article is within the scope of WikiProject Internet, a collaborative effort to improve the coverage of the internet on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
B-Class article B  This article has been rated as B-Class on the project's quality scale.
 Top  This article has been rated as Top-importance on the project's importance scale.

SIP As Transport Protocol in OSI?[edit]

In the TCP model, Session Initiation Protocol is at application layer because there really is nothing else over the transport layer. But given what is listed on OSI, it seems SIP should be included at the transport layer there. This is especially true with its growing importance of SIP in VoIP.

It just seems like it should be here somewhere... The preceding unsigned comment was added by (talk • contribs) 12:51, 26 August 2005.

Scheme vs. protocol[edit]

HTTPS (actually https:) is a URI scheme, not a protocol. The scheme describes a different protocol stack that includes SSL or TLS, but the protocol is no different. For that reason, might it be better to keep the distinction clear? The preceding unsigned comment was added by (talk • contribs) 05:15, 29 August 2005.

See Talk:HTTPS Hrvoje Šimić 06:55:44, 2005-09-13 (UTC)

Layers in the Internet Protocol stack[edit]

The above captioned section groups a lot of TCP/IP and non TCP/IP protocols in the OSI model. There should be something to identify the TCP/IP protocols from these, especially when the article does not provide a list of TCP/IP protocols anywhere else. RMehra 25 January 2006


the diagram "Sample encapsulation of data within a UDP datagram within an IP packet" overlaps with the list to it's left, on medium sized windows. I suggest moving the diagram downwards, or replacing it with something smaller.

Since fixed. -- Beland (talk) 00:46, 25 April 2012 (UTC)

Please Remember The 1960s![edit]

A student of mine approached me with the view that ARPANET was created in the 1970s. The view was taken from Wiki:

"Today's IP networking represents a synthesis of two developments that began in the 1970s, namely LANs (Local Area Networks) and the Internet, both of which have revolutionized computing."

I have inserted "1960s and 1970s" as 1970s alone, when referring to the beginnings of the development of the internet, is misleading.

Service/Model Layer[edit]

I think this exists above the application layer. DarkNRG (talk) 11:44, 8 February 2012 (UTC)

Protocols need written for this layer. DarkNRG (talk) 16:40, 8 February 2012 (UTC)

These claims are not supported by the well-referenced chart in the article. -- Beland (talk) 00:47, 25 April 2012 (UTC)

Requested move[edit]

The following discussion is an archived discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. No further edits should be made to this section.

The result of the move request was: request withdrawn. Dabomb87 (talk) 02:45, 6 September 2011 (UTC)

Internet Protocol SuiteInternet protocol suite

Per WP:CAPS and WP:TITLE, and because this is a generic, common noun, not a propriety term or a title, the article title should be downcased. Tony (talk) 03:48, 3 September 2011 (UTC)

No, this is not a common noun, it is the specific set of protocols that adhere to the concepts discussed in the article. There can be many Internet protocol suites, as even OSI could be called an Internet protocol suite since its purpose was clearly for the same goal. The situation is much like the word Internet (vs. internet), it can be used in both forms of capitalization, but the discussion of this article is about the specific model, including the layered architecture. Kbrose (talk) 04:27, 3 September 2011 (UTC)
  • Oppose. No, it is used here as a proper noun to refer to the specific Internet Protocol Suite, not any other arrangement nor configuration (suite) of protocols for inter-networking. — Dgtsyb (talk) 05:04, 3 September 2011 (UTC)
  • So let me get this right (because it's not entirely clear from the text, at least at the top, to a reader who doesn't already know this field): there's one and only one such suite, called the IPS. It might be modified from time to time, but people always know what specific protocols and structures it comprises at any one time? If so, yeah, it should be upper-cased, and thanks for clarifying. I note two previous attempts to downcase it:

Here are the ngrams. The green line is interesting:

ngrams. Tony (talk) 05:43, 3 September 2011 (UTC)

As Kbrose pointed out, 'Internet protocol suite' can be used to refer to a general inter-networking protocol suite, such as OSI; however, that is not the context of this article. The green line peaks in the mid to late '90s, which is the period of time that OSI was considered by many to be a viable replacement for TCP/IP. They all fall off after this period, likely because the Internet Protocol Suite is more commonly refered to as simply as 'IP' after that period. — Dgtsyb (talk) 22:54, 3 September 2011 (UTC)
OK, thanks. I accept your reasons. Tony (talk) 03:45, 4 September 2011 (UTC)
The above discussion is preserved as an archive of a requested move. Please do not modify it. Subsequent comments should be made in a new section on this talk page. No further edits should be made to this section.

Requested move, revived[edit]

FYI: User:JCScaliger has been indef blocked as a sockpuppet of User:Pmanderson (blocked for another year for abusive sockpuppetry).
The following discussion is an archived discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. No further edits should be made to this section.

The result of the move request was: page moved. Vegaswikian (talk) 17:11, 2 November 2011 (UTC)

Internet Protocol SuiteInternet protocol suite – The RM above was prematurely withdrawn after Tony accepted a flaky rationale. Since the defining official RFC documents all say "Internet protocol suite", I see no basis for doing otherwise. And a better n-gram comparison that eliminates title context is made by putting "the" in front, and shows a more consistent majority usage as non-proper: [3]. If most books treat it as generic, why should we not follow our MOS:CAPS and make it lower case? There are also a fair number of "Internet Protocol suite", which seems logical for those who think that it's IP modifying suite, but that's not how the IETF defines it. And these book hits are still clearly in the context of the Internet protocol suite, not an internet protocol suite or an Internet protocol suite. They all refer to the topic of this article, possibly unlike some in the previous ngram search that Tony did. Dicklyon (talk) 20:25, 26 October 2011 (UTC)

  • Support as nominator. Dicklyon (talk) 06:51, 30 October 2011 (UTC)
  • Support—I have to agree. It's just a suite of things. The wording of the title cannot be mistaken for anything else. It seems like another case of upcasing the expanded form solely because it's abbreviated with caps. Tony (talk) 01:01, 27 October 2011 (UTC)
Tony, the suite is actually not commonly abbreviated. --EnOreg (talk) 07:52, 27 October 2011 (UTC)
  • SupportRuud 20:29, 27 October 2011 (UTC)
  • Oppose. While Tony's move requests to lower case are generally justified, which is why I have not bothered to comment on them, this is an exception. He was right the second time: this is not about a class of suites, this is about a specific entity, using the name proper to that entity. JCScaliger (talk) 03:09, 30 October 2011 (UTC)
    • My move rationale cited evidence in sources for the generic use, including in the official document of the IETF. What evidence do you have to support your assertion that it's a proper name? Dicklyon (talk) 06:50, 30 October 2011 (UTC)
  • Support although weakly. For a long time it could be considered a proper name of sorts, since there was one real protocol (Internet Protocol of course) at its center. However, by now at least there are two, IPv4 and IPv6. So it could be interpreted more as "a set of protocols specified by IETF" which would argue for the lower case of the "protocol" and "suite" but capitalize "Internet" since one particular one of those is meant. And yes I do strongly agree that all the generic "layer" names should be lower except for very specific ones. W Nowicki (talk) 05:06, 2 November 2011 (UTC)


I'm afraid I'm not sure about this one. The lemma indeed refers to one particular suite of protocols, namely those specified by the IETF around the Internet Protocol. --EnOreg (talk) 07:52, 27 October 2011 (UTC)

But the IETF documents don't capitalize "Internet protocol suite", even though they do capitalized "Internet Protocol" when referring to IP. They also distinguish generic uses of the latter, as in "Internet protocol experts", which are experts on Internet protocols, not on IP. The suite is generic. Dicklyon (talk) 14:50, 27 October 2011 (UTC)
The intro to RFC1123 makes it clear that IP overlaps the suite, via the IP layer; it doesn't define the suite, which is why it's not "Internet Protocol suite":
  This document is one of a pair that defines and discusses the
  requirements for host system implementations of the Internet protocol
  suite.  This RFC covers the applications layer and support protocols.
  Its companion RFC, "Requirements for Internet Hosts -- Communications
  Layers" [INTRO:1] covers the lower layer protocols: transport layer,
  IP layer, and link layer.
Dicklyon (talk) 14:54, 27 October 2011 (UTC)
I don't much trust standardization bodies on spelling, even if the IETF does better in this respect than, e.g., ISO. A search on Google Books shows very inconsistent capitalization of the term. Maybe there is no clear answer.
I don't want to stand in the way of either decision. I just wanted to explain why I cannot support the motion. I don't oppose it either. --EnOreg (talk) 09:13, 28 October 2011 (UTC)
"Very inconsistent" in sources supports lower case on wikipedia, per MOS:CAPS. Dicklyon (talk) 15:37, 28 October 2011 (UTC)
Ah, I see protocol stack and protocol suite. Are they siblings in the sense that wp:titles talks of? Tony (talk) 04:17, 29 October 2011 (UTC)
I do not think they are siblings but the more generic concepts. The question is if this refers to one particular one, or a family of related ones as indicated above. W Nowicki (talk) 05:06, 2 November 2011 (UTC)
The above discussion is preserved as an archive of a requested move. Please do not modify it. Subsequent comments should be made in a new section on this talk page. No further edits should be made to this section.

Capitalization in "internet layer"[edit]

I've downcased "Internet layer" here and elsewhere. It seems to be more the generic concept, about internetworking, as opposed to a reference to the Internet (when I use TCP/IP to talk to my sprinkler controller, I use the internet layer, but I don't use a network that connects to the Internet, for example). There are a few sources (less than 20% of books, for sure) that capitalize "Internet layer" while downcasing "link layer", but the vast majority don't treat "internet" in this context differently from "link"; the original IETF documents actually have it in sentences both ways, so that's not so helpful. Most official TCP/IP docs don't do layers at all, so that too suggests that it's generic. OK? Dicklyon (talk) 20:53, 5 November 2011 (UTC)

Ah, I see I just got reverted. I have no issue on the over-linking complaint. I'll wait for comments on the case. Dicklyon (talk) 20:55, 5 November 2011 (UTC)

As examples, RFC1122 says

             A packet is the unit of data passed across the interface
             between the internet layer and the link layer.  It
             includes an IP header and data.  A packet may be a
             complete IP datagram or a fragment of an IP datagram.

but also

             IGMP is an Internet layer protocol used for establishing
             dynamic host groups for IP multicasting.

while RFC1123 uses "IP layer" instead; same layer, different naming concept. Dicklyon (talk) 21:18, 5 November 2011 (UTC)

Internet in Internet layer does not refer to some general inter-networking: it refers to the Big-I, the Internet. You know, the one that we are exchanging these comments on. (This also has been discussed and agreed before.) — Dgtsyb (talk) 01:01, 6 November 2011 (UTC)
Then why do so many sources lower-case it? And why do my packets have to go through the internet layer to speak IP on my private net that doesn't connect to the Internet? Where has it been discussed before? Dicklyon (talk) 01:55, 6 November 2011 (UTC)
Dicklyon is correct, and the revert (diff) has produced mistakes like:
According to RFC 1122, the Internet protocol suite organizes the functional groups of protocols and methods into four layers, the application layer, the transport layer, the Internet layer, and the link layer.
The uppercase 'I' was not an issue until very recently because (as far as I can see), previous versions of the article used text like "the Internet Layer, and the Link Layer". While it would be interesting to see the previous discussion mentioned above, any conclusion that "internet layer" in TCP/IP needs to be spelled "Internet layer" is erroneous. Of course TCP/IP was developed for what was to become the Internet, but as the first sentence of the article correctly acknowledges, it was also "and other similar networks". Johnuniq (talk) 02:55, 6 November 2011 (UTC)
The internet layer of the Internet protocol suite: that's absurd. You have gone crazy lowercasing proper nouns. — Dgtsyb (talk) 05:28, 6 November 2011 (UTC)
No, it's not absurd. I would have been quite happy to leave "Internet Layer" unchanged, but that issue is not related to the current discussion, although of course it revealed the problem (and I'm sure this discussion will be repeated at least once a month). TCP/IP is a system of networking that uses four layers, and one of those layers is the "internet layer" which includes things like a destination IP address so datagrams can be copied from network to network via routers (hence "internetwork"). TCP/IP was always intended to be quite general, and "internet" is just an abbreviation for "internetwork". The reason for the capital I in Internet is to distinguish the Internet from all the other internets that also use TCP/IP. Johnuniq (talk) 07:06, 6 November 2011 (UTC)
internet is not an abbreviation—its not even a word: internet. could be, but isn't (check any dictionary). Internet is a proper name. — Dgtsyb (talk) 12:05, 6 November 2011 (UTC)
It sounds like you're too young. Back in the day, papers such as this one discussed internet routing, internet addressing, internet packets, etc., before there was "the Internet"; the techniques were developed on the ARPANET and the internal networks of various companies and universities. Or this one where Vint Cerf and Yogen Dalal define TCP for internets, before there was IP. Dicklyon (talk) 16:08, 6 November 2011 (UTC)
I was programming computers before the Internet was even conceived. Go to websters and punch in internet (lowercase) and we what you get: Internet, capital "I", the global communications network. They don't list any abbreviation. Likely because internet is not even a word. — Dgtsyb (talk) 16:16, 6 November 2011 (UTC)
So much for that guess. But here are some dictionaries that explain lowercase internet: [4] (particularly this one and this one and this one). Dicklyon (talk) 16:27, 6 November 2011 (UTC)
Those aren't dictionaries: they are books of jargon and other invented words that are not part of the English language. Only Internet appears in a real dictionary. — Dgtsyb (talk) 22:40, 6 November 2011 (UTC)
The online Oxford learners dictionary lists "internet" as an alternative. And several dictionaries say that "Internet" first appeared in 1985. So maybe the older and generic uses are indeed uses of that "jargon" as you call it. So what? Does that mean that when the IETF defined an "internet layer" they were referring to "the Internet"? I can't see the logic there. Dicklyon (talk) 23:11, 6 November 2011 (UTC)
There is a Wikipedia article on Internet capitalization conventions that you might want to look at. Jeff Ogden (W163) (talk) 18:25, 6 November 2011 (UTC)
Thanks for that. It seems that the case distinction is pretty conventional, if not universal. Dicklyon (talk) 18:29, 6 November 2011 (UTC)
Section 7.76 in the Chicago Manual of Style (16th edition) says:
Terms like "web" and "Internet." ... generic terms that are capitalized as part of the official name of a system or an organization may be lowercased when used alone or in combination.
And gives this example: "Internet protocol (IP); the Internet; the net; an intranet".
Jeff Ogden (W163) (talk) 23:37, 6 November 2011 (UTC)
  • Comment I seem to read a fallacy hereabove that because 'Internet' is now considered a proper noun, everything that is now associated with it or is prefixed by that word ought to be in upper case. What do the sources say – I'm referring to the world at large and not boffins' publications or 'technical' journals? --Ohconfucius ¡digame! 01:20, 7 November 2011 (UTC)
    • No, we got over capitalizing "layer" and "protocol suite" already, so that's not the issue. The issue is about when "internet" itself is generic. One user says never, though many docs, and some dictionaries, say often, when it's not referring to the Internet. Dicklyon (talk) 05:16, 7 November 2011 (UTC)
  • I also notice that this graphic, used in the article, seems also to be in breach of WP:CAPS. --Ohconfucius ¡digame! 01:46, 7 November 2011 (UTC)
    • Do we have guidelines for text in images? Dicklyon (talk) 05:16, 7 November 2011 (UTC)
      • It's a great pity that this confusion still exists about I verus i in internet. Yes, IT professionals often still make the distinction between the worldwide Internet and smaller, in-house networks with a lowercase i. But almost no one has heard of this distinction, and in normal writing, including the press, they decided some time ago it's all too fussy, and use internet for all. That is how it will end up, I can assure you, and in compound derivative items like this one, I think it would do everyone a favour if we used lower case. Tony (talk) 06:32, 7 November 2011 (UTC)
        • You are probably correct about the eventual outcome, but regardless, in this article there is no reason to use a cap I (apart from a mistaken belief that "internet layer" refers to the Internet). The cap I is particularly silly in text like "the link layer, the Internet layer, the transport layer, and the application layer". Johnuniq (talk) 07:16, 7 November 2011 (UTC)
        • Just because the ordinary person believes something bogus doesn't mean we have to cater to them. There's a difference between 'internet' (the technical concept), and 'Internet' (the particular worldwide network using that technology), and one day people will get it straight (just like the White House is not the only white house - and most people manage to capitalize that one correctly). Having said that, "internet" is correct here - we're talking about a generic concept, not 'the' Internet Protocol (which is capitalized because it's a particular internet protocol, not, say, PUP). Noel (talk) 21:26, 8 January 2012 (UTC)


I know it's the wrong place to start, but I did a minor change to the wording in the lead. The last para may need more work (introduced in this edit). Grammatically "This is called the link." is unfortunate as it interferes with the flow to the next sentence, namely "It provides..." (what is "It"?). Also, I'm wondering if the sentence contributes much, and I'm thinking it should be removed. I am puzzled by:

The internet layer provides communication methods between multiple links of a computer and facilitates the interconnection of networks.

The internet layer is concerned with conveying data from a source to a destination, possibly over multiple links—some rewording is needed, and I cannot see why "multiple links of a computer" should be mentioned at that point. The application layer wording needs some thought. Johnuniq (talk) 01:40, 8 November 2011 (UTC)

I have trimmed that last paragraph in the lead. We have plenty of body text and linked articles (in the previous paragraph) that cover the details. Very broad strokes are all that are required here. --Kvng (talk) 16:01, 9 November 2011 (UTC)


The content of this article is a duplication of effort in the article TCP/IP model, which is more well established.

Merge discussion is being conducted at Talk:TCP/IP model#Merge_discussion. Please post comments there. Stephen Charles Thompson (talk) 20:29, 26 January 2012 (UTC)

I implemented the merge. -- Beland (talk) 00:48, 25 April 2012 (UTC)

No mention of the notable BSD TCP/IP stack implementation[edit]

The BSD TCP/IP stack has been one of the early implementations to be borrowed by various corporations (including various UNIX and Microsoft Windows), which won notable land speed records, and that is still being used today by many systems and companies because of its flexible license, high portability, scalability and configurability (Cisco, QNX, *BSD, MacOSX, Solaris, Juniper, etc). Remains for someone who has time to do some librarian reference research and include this in the history and/or implementations section(s), though... (talk) 10:00, 8 August 2012 (UTC)

I forgot to mention that the most common API to access network services on operating systems is still the BSD IPC one (see Berkeley Sockets, with Winsock being a compatible library implementation as well), and that IPv6 deployment pioneer projects (see KAME project, WIDE Project) were also based on the BSD TCP/IP stack. (talk) 16:49, 9 August 2012 (UTC)

I revisited the article. An unreference sentence also reads: "Many companies sold TCP/IP stacks for Windows until Microsoft released its own TCP/IP stack in Windows 95.", it is known that the Windows TCP/IP stack was also based on the BSD one, at least early revisions. Moreover, the AT&T code left in the BSD TCP /IP stack, if any, was pruned in 1988, after USL v. BSDi. A full BSD OS free of AT&T code was subsequently released as Net/2 in 1991, including a TCP stack, which was widely borrowed, due to its liberal license. (talk) 22:35, 11 March 2013 (UTC)

Category for discussion: Category:TCP/IP -> TCP / Transmission Control Protocol[edit]

Everyone kindly is invited to take part in the discussion to rename Category:TCP/IP, please see: Wikipedia:Categories for discussion/Log/2012_August_9#Category:TCP/IP. -- intgr [talk] 16:07, 13 August 2012 (UTC)

Standards Bodies overseeing this model[edit]

The last sentence of the first section claims that "The TCP/IP model and related protocols are maintained by the [IETF]", which is incorrect/incomplete. Ethernet for instance is a standard maintained by the IEEE (IEEE 802.3); ISDN was developed by the ITU. I don't know enough about the various link layer protocols to spell out a complete list of standards bodies, but my gut says that protocols at the link layer are commonly not an IETF standard. Eli.heady (talk) 02:56, 23 August 2012 (UTC)

The question we need to answer is if these lower lever protocols are included in the "TCP/IP" model? I don't think that they are. The IETF does oversee the use of these lower level protocols within or by other parts of the "TCP/IP" model. Still it might be good to expand this to say something like:
"The TCP/IP model and related protocols are maintained by the [IETF]. Some of the protocols used by the TCP/IP model are maintained by other standards bodies such as the IEEE (Ethernet, 802.11), ITU (ISDN), and W3C (World Wide Web)."
--Jeff Ogden (W163) (talk) 11:47, 23 August 2012 (UTC)

Under History:Adoption, a correction[edit]

You say under the heading "Adoption": "In March 1982, the US Department of Defense declared TCP/IP as the standard for all military computer networking.[7] In 1985, the Internet Architecture Board held a three-day workshop on TCP/IP for the computer industry, attended by 250 vendor representatives, promoting the protocol and leading to its increasing commercial use.'

According to the article on the Internet Architecture Board (IAB), in 1985 the IAB, technically, did not yet exist. Here's a quote from that article, which is found at: "The body which eventually became the IAB was created originally by the United States Department of Defense's Defense Advanced Research Projects Agency with the name Internet Configuration Control Board during 1979; it eventually became the Internet Advisory Board during September, 1984, and then the Internet Activities Board during May, 1986 (the name was changed, while keeping the same acronym). It finally became the Internet Architecture Board, under ISOC, during January, 1992, as part of the Internet's transition from a U.S.-government entity to an international, public entity."

Thought you might want to adjust your paragraph in light of the above chronology. Wikifan2744 (talk) 05:06, 9 September 2013 (UTC)

Done. Thanks for pointing this out. --EnOreg (talk) 08:28, 9 September 2013 (UTC)

You're welcome. Wikifan2744 (talk) 08:26, 1 November 2013 (UTC)

Host-to-host, process-to-process communication[edit]

The article talks about the "transport layer handling host-to-host communication". The host address, however, is the network-layer address whereas at the transport layer, port numbers identify an individual process on a host. Can someone please resolve this apparent contradiction? --EnOreg (talk) 14:15, 28 April 2014 (UTC)

Where is the contradiction? Kbrose (talk) 04:02, 29 April 2014 (UTC)
Well, the communication is happening on the layer that does the addressing, i.e., the network layer uses host addresses because it handles host-to-host communication, just like the transport layer uses process addresses (port numbers) because it handles process-to-process communication. The current text, in contrast, says that the "transport layer [handles] host-to-host communication". How is that true? --EnOreg (talk) 07:02, 29 April 2014 (UTC)
I think the communication actually happens at the layer above always, Internet Layer provides addresses, but the communication between end-to-end hosts happens above as the Internet Layer has no means to establish a connection (in case of TCP for example). Same with the port numbers, provided by the Transport Layer. The protocols in that layer have no means to actually use a port number; the transport protocol provides the end point to the layer above. Only the Application Layer understands the meaning of the port. The term port is really a synonym for service. The diagram on the page also illustrates this, and pure routers don't actually need an implementation of the Transport Layer for the purpose of exchanging packets. So, a layer provides tools for the actions/communications of the higher layer in general. But as TCP/IP does not enforce such layering ideas strictly, there may well be other interpretations and contrary interpretations for specific cases. Kbrose (talk) 15:23, 29 April 2014 (UTC)
I see where you're coming from. So maybe the definition of communication is too ambiguous. In the given context I interpret the term to mean the exchange of data, not their interpretation. If that is not the intended meaning we should find a wording to make that clear.
The network layer delivers data from one host to another, whereas the transport layer delivers data from a particular process on the source host to a particular process on the destination host, right? On top of this, the program running in the process implements a particular service and exchanges data using the protocol associated with that service, representing the application layer. This appears consistent with the rest of the article, right? --EnOreg (talk) 09:50, 30 April 2014 (UTC)
The network layer is called Internet Layer in TCP/IP, it is a layer that provides internetworking, meaning it connects or bridges the scope of networks, not host-to-host. At the top, in the Application Layer we have a pipe between processes, in the Transport Layer we find a pipe between hosts, the Internet Layer is a pipe between networks, and the Link Layer is a pipe between two interfaces on the same link. The diagram illustrates this. The TCP/IP model is entirely scope-based. Whether the interaction is called communications or data exchange is not really important, where interpretation happens is outside the scope of the model, but since an application is the only logical entity that could possibly interpret payload data, it would be in the Application Layer. Kbrose (talk) 19:22, 30 April 2014 (UTC)

Sorry about the lag. RL interfered. Hm, I'm surprised we have such a fundamentally different understanding. To get out of this we could search the literature for evidence for either view, or ask a third opinion. In the IP context I can't find much about "host-to-host" at all. RFC 1122, in section "Internet Layer" says:
"All Internet transport protocols use the Internet Protocol (IP) to carry data from source host to destination host."
Tanenbaum says about the Internet layer:
"Its job is to permit hosts to inject packets into any network and have them travel independently to the destination (potentially on a different network)." And:
"The network layer is concerned with getting packets from the source all the way to the destination" and "is the lowest layer that deals with end-to-end transmission."
All this seems to say that IP (the network layer) carries data from host to host. About process-to-process communication Tanenbaum only has this to say:
"In the transport layer, entire connections can be encrypted end to end, that is, process to process."
Where does your concept come from?

As a side-note: It's generally not a good idea to edit subjects while they are being discussed. Surely you know about the potential of such actions to cause edit wars. Therefore, I would appreciate if you could revert this part of your May 2 edit until we've achieved a consensus. Thanks, --EnOreg (talk) 12:15, 9 May 2014 (UTC)


This section states "the suite has been implemented on essentially every computing platform". This is far from true, some platforms have insufficient computing power to implement the whole suite, and for some platforms there are no implementations. May I suggest this is reworded to 'implemented on nearly every platform from (some date) onwards.' No citations or references either.

Also is there a list of IP suite implementations on WP? John a s (talk) 20:45, 19 February 2015 (UTC)