This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 1 |
A search on Google shows that this page has paragraphs found multiple times accross the internet (right down to the space in 'a re' in the last paragraph). Whether this means its in the public domain, or that they are all stealing from the same source I am unsure. If its the former, something needs to be cited. - Eean—Preceding unsigned comment added by 24.12.199.xxx (talk • contribs) 04:18, 17 December 2001 (UTC)
I don't think this graphic is appropriate for an encylopedia. 1. The reference to "The Internet" is not appropriate. The internet is not always network (ie it can be GSM, non-internet TCP/IP nets etc) 2. The figures of two human faces do not add to the information content and can infact be misleading (ie endpoints are not always, human). 3. This images does not bring anything new or significantly aid in depicting the concept, especialls since "Rm-osi parallel.png" and 2 two tables included in the article.
Recommended Action: REMOVE Image "Osi_model_com.jpg" from the article & restore formatting.
Note to Markolinsky: It seems you have spend quite a bit of effert on this drawing and it looks good, but I don't think the graphic style and content is not appropriate in this article. In the future, it is probably best to discuss in the talk page before doing changes.
--202.161.20.46 08:42, 27 September 2006 (UTC)
Removed from article. As said above - Internet usually just works on network layer. --Nux (talk)
from user talk:MyRedDice
Here's my thinking:
First on OSI you'll notice the 7-long table with some examples of things at each of the 7 OSI levels.
You'll notice ethernet is on level 2 and 10BASE-T is on level one.
goto 10BASE-T and you'll notice that its said to be a "varient of ethernet",
PLEASE help me clarify how a variant of ethernet is on a DIFFERENT OSI level than it.
(see, i think that OSI categorization is just a loosely thrown together piece of ill-defined cruft) Anyway, maybe you can educate me
I think I can help. Ethernet is level two. But it depends upon level one: the physical layer. Thus you can run ethernet either on 10BASE-T cables, or on 10BASE2 cables, or on [[please respond to user_talk:hfastedge so that I dont have to constantly look at recent changes.
anyway, "you're just changing which variant of the physical layer it works on" that still doesnt address how "10BASE-T is a variant of ethernet" as mentioned in 10BASE-T.
The confusion here is that Ethernet is both layer 1 and 2. 10BASE-T is just one implementation of layer 1 for Ethernet. It isn't a "variant", it's a specific implementation. --AMillar
I'm thinking that maybe I should add some of my comments at the end of Talk:Internet_protocol_suite#Confusing layers (about how the OSI model really divides things up in functional layers, and you shouldn't expect the actual dependency maps of real protocol suites to follow it slavishly) to this page; that will prevent a lot of confusion as to e.g. why BGP, which uses TCP, is at the network/internet layer.... Noel 12:40, 18 Sep 2004 (UTC)
BGP is an application (Layer7) that uses a TCP/IP session between two "speakers" The fact that it exchanges routing information is not relevant for Layer3. The given that those "speakers" live in devices that are called Routers, does not degrade BGP from Layer7 to Layer3. Similar to Telnet/SSH, FTP to a router. Would Telnet to a Switch suddenly make Telnet a Layer2 function? I see no problems with the OSI model, only human errors in interpretation.
The Management Annex to the OSI Reference Model defines two kinds of management, system and layer. System management is, indeed, an application, such as SNMP in the Internet and CMIP in OSI protocol stacks. Layer management protocols affect the functioning of a given layer without the need to have something at the application layer; this is also discussed in the OSI Routeing [sic], ISO/TR 9575. The key attribute of a layer management mechanism is that its payloads affect the information transfer protocols of the layer, not that it run "on" the layer. BGP runs over TCP, but is layer 3 management. Pure OSI IS-IS runs directly over layer 2, but is layer 3 management. RIP runs over UDP, but is layer 3 management. OSPF runs over IP, but is layer 3 management. Hcberkowitz 02:47, 10 May 2007 (UTC)
Sometimes a layer 0 is introduced to refer to the physical medium the data is transported accros, eg. the copper or glass cabling itself. This is because strictly speaking the cabling does not belong to the osi model itself. [JP]
If not this, layer 8 is another way to reference user error. 166.70.62.200 (talk) 22:43, 23 March 2008 (UTC)
MPLS is considerd to be a switching techonology, i.e. layer 2. However it runs on other layer 2 technologies such as Ethernet or ATM, then why is it not considerd layer 3. Then again is what defines a layer 2 Protocol one that specifies the next hop in the path while layer 3 specifies the final destination.
Possibily could MPLS be considerd a sub-layer of Layer 2, so if MPLS ran over Ethernet there would be 3 sub-layers LLC layer, MAC layer and MPLS layer. -Vec 19 April 2005
MPLS is a Layer3 function, without doubt. Not layer2, nor layer2.5 It modifies layer3 headers. The fact that it does it in Layer2 hardware (the switch) does not mean it will work in ANY switch. The switch that can act on MPLS fields in packets is relying on the routing info from routers at the edge of the MPLS autonomous system.
The confusion here is first due to people trying to coerce protocols into the original seven-layer structure defined in ISO 7498, without any of the relevant appendices, or, more importantly, later ISO work such as ISO 8648, "Internal Organization of the Network Layer". That document comes up with a more flexible three-sublayer model with the protocol-independent network layer protocols like CLNP or IP on top, a subnetwork protocol (i.e., lower layer, not an IP subnet) such as LLC at the bottom, and mapping protocols like ARP in between.
MPLS, however, doesn't neatly fit either the classic OSI model, or even the classic IP stack. It uses routing protocols (i.e., layer 3 management) to find potential paths, end-to-end management protocols like RSVP-TE to map the MPLS topology, and then hop-by-hop protocols such as LDP to manage the tables of the label switched routers.
Things get even more confused when dealing with Generalized MPLS (GMPLS), which doesn't limit itself to paths for packets/frames, but also for lambdas/optical wavelengths, time slots (e.g., SONET/SDH or other multiplexing), or even physical ports on a cross-connect device like a DACS. Hcberkowitz 02:55, 10 May 2007 (UTC)
MPLS is a routing mechanism that overrules the original protocols. The forwarding of packets or frames is done based on the lable added to the header instead of using destination address based look-up tables. Therefore MPLS is a Layer 3 protocol that changes the forwarding/switching at intermediate nodes. GMPLS extends the label switching concept to include bit-stream and capacity switching (also known as layer 1 and layer 0 in the IP centric world). However, GMPLS is mainly seen as control plane issue intended to provide a unique control plan that is able to manage the switching options of all potentially involved (sub-)network layers. Sprawl 12:39, 24 May 2010 (UTC) —Preceding unsigned comment added by 85.127.103.152 (talk)
I am not changing it though, cause I am starting a new job tomorrow and I have to read some suff to feel ready. —The preceding unsigned comment was added by 64.231.152.31 (talk • contribs) 03:28, 3 August 2005.
Guys, can we consider deleting references to dicks and pussies? It really detracts from an otherwise useful page. —The preceding unsigned comment was added by 216.184.31.184 (talk • contribs) 23:26, 3 August 2005.
As I studied the OSI model, I was quite confused, and just didn't get the idea right. Eg, I tried also to map OSI on other protocols, and it never seems to work out. As such, for a long time I saw OSI more or less as an academic toy. It took quite some time before I finally got the sense of it, something I wanna share with you all.
First of all I think the OSI concepts are much more useful then the implementation of the model itself. What are those concepts? Well there is communication, there is layers and there is language. These work in close harmony. The essence here is that each layer communicates with the same type of layer at the other site of the 'communication line'. Let me stretch this: two layers of the same type are talking to each other in the same 'language'. You can see this in almost every picture, but it is never really emphasised. What is the use of this concept? Well it simply helps you to focus on the function of a layer: what does it do, or when designing: what does it have to do.
Next concept is that in the OSI model, except for the physical layer, all layers are separated, and thus actually can not communicate with each other directly. From the model it is obvious that layers are talking too the next lower level (through the so called Service Access Point or SAP). The essence here is that the layer has some means to talk to his counterpart. In other words it tells you how a layer can talk to its counterpart.
Last concept is language. I use the concept of language in the sense of a set of words related to the same domain (for the intended usage database and road don't belong to the same language; road and car generally will). To define a language I ask myself 'what words do I need', 'is the language complete', 'is the language consistent', 'are there any ambiguities'? In every day sense it is sort of pragmatic, semi-formal tool. In the OSI model language is not explicitly defined as a concept, but it's left implicitly. I'm not going to give a precise definition, however you can see as a set of words. These words should be the only words you use to describe a function. And yes: in this way there are many, many 'languages'.
Summary: layers on the same level are virtually functionally communicating to each other in the same language ('horizontally') by really communicating technically ('vertically') to the next lower level layer. These two forms of communication are more or less independant, and to keep things simple you should always focus on one of them at a time.
To me the OSI model is a generic model which solves probably most or all network problems. For my day-to-day work, this is not very useful. So how do I use this concepts? Well, as simple and as obvious as they seem, I use them a lot. First I used them to understand network protocols. It especially helps me to keep out of the OSI-mapping discussion (try to map tcp/ip on these concepts ;-), and it helps me to explain network protocols to a lot of people with ease. Furthermore, if I have to do some work on communication I always ask myself: what layer(s) are we talking about, what language is used and how do they communicate with each other. And yes: this sounds simple, but in my experience 4 or 5 out of 10 don't grasp the idea and take much longer time then needed (if they succeed at all) to complete a job. And this is for network people.
And now the fun part. You can use the same concepts in application development: every time you're working with interfaces, you're talking about communication. The obvious usage is of course for interfaces between processes. Especially during design you can get much clearer discussions. If you 'know where you are' in a protocol stack you can much more easily focus then when several concepts are mingled. A nice example is the man/machine interface. From the above you can conclude that, functionally, only layers who 'speak the same language' can talk to each other. A second issue is that each application has its own 'language'. Working with a financial application is something completely different from working with a video-editor. Each application has its own set of functions, concepts etc. and the human to work with it has to speak the same language. From a design point of view I've turned this around: besides using use-cases etc, in a pragmatic way I formalise the language that the intended users use. Interactively I ask them what they need from an application and I teach them how to talk to a computer. Crucial here is that the language which is developed here is actually their own language, but semi-formalised, and as such it is a language which can be understood by both layers. —The preceding unsigned comment was added by 213.84.162.175 (talk • contribs) 09:02, 21 August 2005.
ISO recognized the network layer could not be monolithic, and split it into three sublayers, the lowest of which overlaps layer 2. The top sublayer is independent of the underlying technology, which is perfectly consistent with IP and CLNP. The bottom sublayer, which is called "subnetwork access" using "subnetwork" in a very different way than IP subnets, can be 802.2, X.25 (with some caveats), etc. The middle sublayer, subnetwork dependent convergence, maps between the top and bottom, with such protocols as ARP.
The bottom subnetwork layer can be applied recursively, which allows such things, admittedly now obsolete, as 802.2/802.3 LAN emulation over ATM. 802.2 is on top, 802.3 next lower, a mapping between 802.3 MAC addresses and ATM addresses, and then link-local ATM protocols.Hcberkowitz 22:19, 5 June 2007 (UTC)
Removing the mnemonic devices was a bad idea. I came here looking for these (which I'd bookmarked) and found them gone. I dont believe that just because one person finds them hard to remember that they aren't useful.
As far as I have been able to tell, not all of X.25 is used by the OSI suite (X.25 is a CCITT standard), although some of the protocols (HDLC?) do seem to have been used for OSI. Perhaps X.25 doesn't belong in the OSI column in the table of examples? StuartBrady
It doesn't fit neatly, especially given X.25 is actually a 3-layer stack. The packet level of X.25 defines the Connection-Oriented Network Protocol in OSI stacks. Its definition is a bit circular, as this was defined by ISO to be able to bring X.25 into the OSI suite. The X.25 packet level (not layer) has some functions that are more properly in transport and even application. Hcberkowitz 06:49, 17 May 2007 (UTC)
The seven layers are described on their own pages, e.g. Application layer.
May I suggest someone adds links from the OSI model article to these pages? Maybe in the section "Description of layers". -- Felix Wiemann 13:59, 14 January 2006 (UTC)
I would like to question the usefullness of discussing a "layer 2.5" in this article. It is not part of the OSI model, nor were any of the examples of a 'layer 2.5' protocol designed for the OSI protocol stack. If this *has* to be mentioned, then this should be mentioned after a definition of the formal layers, perhaps even under an 'Informal Layers' header. -- Ryanfantastic 11:01, 31 March 2006 (UTC)
It currently states that MPLS operates on packets (layer 2)... should that not be (layer 3), or frames (layer 2) Thedarxide 09:42, 22 May 2006 (UTC)
This function is generally covered by the Internal Organization of the Network Layer, although MPLS does not fit, was developed after real-world protocol developers (as opposed to pedants who don't keep up with network engineering) gave up on a seven layer model, and will never fit cleanly into OSI.Hcberkowitz 22:18, 5 June 2007 (UTC)
I have deleted all the mnemonics because how many are being made up. WP:V says the burden of evidence (sources) is on the poster. See here for the mnemonics deleted. Cburnett 05:03, 6 April 2006 (UTC)
I deleted the mnemonics again. Just because you have a citation for them doesn't make them valid for inclusion; they're just some things that a For Dummies author made up. If you want to help people remember the facts given in this article, I suggest writing a Wikibook on the subject. For the encyclopedia, please, just the facts. ~ Booya Bazooka 05:29, 9 August 2006 (UTC)
Um. If it's not referenced, but isn't stupid either, you can keep it. Seriously, there's going too far and then there's going too far. Wikipedia is NOT a bureaucracy, Wikipedia is an encyclopedia. --Kim Bruning (talk) 05:05, 12 January 2008 (UTC)
I noticed in this article the names of the layers begin with a capital letter, however in the 7 articles on the layers themselves they are in lowercase (most of the time). This needs to be rectified.—Preceding unsigned comment added by 141.157.54.202 (talk • contribs) 21:23, 19 April 2006 (UTC)
I was doing a research project on cryptography and in a certain text, (Mitchell, Chris J, Users Guide to Cryptography and Standards, Artech House inc, 2005 [p 32-33]) the authors state the following
"No security services can be provided on in layer 5 or layer 6, although layer 6 may contain facilities to support the provision of services at layer 7."
I must say that I agree with the text, that security services such as encryption are not handled in the functions of presentation layer. It is my personal view that the 6th layer serves the function of interpreter rather then decipherer and as being analogous to the interpreter should not have a hand in the content of the message, or data to be conveyed. Therefore I believe that the functions of (en/de)-cyperment should be excluded from this layer as it is seen in the word encryption.
Magus 05:19, 20 April 2006 (UTC)
My view of the layers here is different: TCP Layer 4, HTTP: Layer 5 - It provides a session going beyond one TCP connection (i.e. session cookies etc..) HTML: Layer 6 (presentation) it provides means for prensenting a web page, Layr 7; Appication: IE/streaming applications/IPC etc... —The preceding unsigned comment was added by 67.183.217.186 (talk • contribs) .
As another example, consider SMTP, which is also a session layer protocol. The article shows MIME as being at layer 6 (which I agree with). MIME encoded messages are transported over SMTP, but SMTP belongs below MIME in the model. But the article shows SMTP above MIME in layer 1/application layer. Iambk 19:53, 22 November 2006 (UTC)
Also, the statement …”TCP/IP is not same as OSI”… TCP (over IP) is a protocol just as any other protocol, there is not reason to take it out of the OSI model. The OSI model talks about protocol layers (not necessarily 7), for sure there are many layers of protocols on top of TCP/IP which makes it 100% compatible to the OSI model. For example, in MMS: MM4 is a protocol, over SMTP, over TCP/IP – see 3GPP specs. I can give you a lot of protocols very real on top of the “allegedly layer 7” in this article.
The mnemonic "A Perfect Student Needs To Drink Port" does not identify the correct letters for network and transport layers cDima 21:13, 13 June 2006 (UTC)
I just reverted someone moving RIP from layer 3 to layer 7.
It's a routing protocol. I suspect that the mover had a different RIP in mind.
There - just wanted to have it on the talk page too. --Alvestrand 20:53, 31 August 2006 (UTC)
RIP might appear as an ambigious case, but I think "RIP is a UDP-based protocol." from http://tools.ietf.org/html/rfc1058 proves that it has to be above layer 4. It has to do with layer 3, because the information send through RIP will eventually be used to reconfigure the IP routing table, which is used in layer 3. But as a protocol, it is clearly above layer 4. Matthias
I agree that routing protocols are layer 7. They contain data intended for use by the routers themselves and are never used to carry data from other end hosts to each other. One does not assign a RIP or OSPF address to a host. One assigns an IP address to a host. One never uses RIP or OSPF to transfer HTTP data. One uses UDP or TCP. RIP, OSPF, and even ICMP are application protocols that actually contain data to be used directly by the recipients. McNuttJ —Preceding comment was added at 17:55, 13 December 2007 (UTC)
At least one of these images needs to be removed, since they're so similar. I would want to remove them both, though - do the pictures tell us anything that our text tables don't? ~ Booya Bazooka 15:46, 28 September 2006 (UTC)
I agree they both should be removed. See my comments re 2nd pic above Talk:OSI_model#OSI_Graphic_-_Osi_model_com.jpg --202.161.20.46 08:11, 29 September 2006 (UTC)
I think the image should be removed. It is beyond the scope of illustrating the OSI Model. The OSI Model is a logical model. I appreciate the efforts of the author. I think a solution is 1) remove current image 2) discuss a future image on the discussion page 3) make new image from input. Joneboi 06:00, 19 November 2006 (UTC)
Is it correct, that the link http://standards.iso.org/ittf/PubliclyAvailableStandards/s020269_ISO_IEC_7498-1_1994(E).zip is pointing to a Zip-File containing a PDF, that itself contains a scan of the ISO-Specs which are under Copyright Protection? Please have a look at that. --80.131.151.219 21:57, 24 October 2006 (UTC)
I'd say "no" - Open Systems Interconnection discusses the overall OSI project, while this page discusses a specific item, the OSI networking model, that the Open Systems Interconnection page claims antedated the project. Guy Harris 19:18, 22 November 2006 (UTC)
Support for not-merge. As per Guy Harris. Visor 14:53, 30 November 2006 (UTC)
In the description section, where the seven layers are explained, why is there no detail for Layer 1 (Physical)? Sapbuckets 16:27, 13 December 2006 (UTC)
It was deleted by a vandal on 12 December 2006 and not reverted correctly. I put it back. --Rick Sidwell 05:41, 27 December 2006 (UTC)
Switches are listed as being part of layer 2 and layer 3. I am confused.
64.211.50.62 16:38, 26 December 2006 (UTC)
I just had a discussion about this with a highly qualified individual on the subject of BGP. The protocol creates tcp sessions which puts it in 5. Although the RFC [[1]] doesn't explicitly state the location in the OSI model, nowhere does it describe an attribute which would put it in any higher layer than 5. An application that administers BGP does not make BGP an application.
I have also attempted reverse arguments to qualify BGP for Layer 7, such as it's route table being a stored data stream (Layer 6), which I could not fully validate. BGP however is a protocol, not an application. Comments please.
I argue that BGP is a layer 7 protocol used to manage a layer 3 protocol. It is not possible to run BGP without tcp or sessions.69.68.125.6 (talk) 04:01, 16 December 2007 (UTC)
Having RIP and OSPF on the same layer makes no sense. RIP messages are encapsulated into UDP packets, while OSPF messages are directly in IP packets. Hence, if OSPF and IP share a layer, so should RIP and UDP. —The preceding unsigned comment was added by 84.193.228.136 (talk) 19:34, 16 January 2007 (UTC).
In the Examples table, under "Misc. examples" 802.11(WiFi) is included as a layer 2 example while 802.11b & 802.11g are listed as layer 1 examples. This is confusing and should be clarified. I believe 802.11(WiFi) should be listed under layer 1 only.
TweakerTV 20:12, 30 January 2007 (UTC)
Inconsistent with SIP as mentioned in Layer 5. In the 4-layer/5-layer model, I suppose anything above the Transport layer is sometimes called an "application". But in the 7-layer model, seems to me that SIP is certainly a session-layer protocol; I understand it can handle arbitrary data structures at the presentation layer, and has nothing whatever to say about how applications handle those what's presented. For that matter, similar issue about http discussed above was not resolved. html is obviously a data structure at the presentation layer, no? If so, how could http be at a higher layer than html?
I can't decode the note in the table entry "TCP/IP Suite" for Layer 5: ...SIP. (Not a separate layer with standardized API.) Not sure if this specifically refers to SIP. Whetever it refers to, I say that real world layers aren't perfectly independent of other layers as in the ideal model. OSI is very useful despite the fact that it does not map perfectly to the way life works. So I'm not sure what special non-separateness this refers to in terms of SIP or anything else. And I'm not sure what what point the editor was making about APIs.
I think there are nested stacks in the world: when tunnelling, inside Sonet...the possibility of nested stacks does not mean that outer stacks are necessarily at higher layers than inner stacks, and it does not necessarily mean that layers are blurred. YMMV. OT, I think SCCP is at SS7 Layer 4, not 3. --Michael Gold 00:49, 3 February 2007 (UTC)
The history section claims that ISO is european dominated. With 157 country members, how can it be dominated by one contintent? I wonder if someone is bitter because 'their' country/continent doesn't dominate ISO? Markb 13:04, 5 February 2007 (UTC)
What does this heading, between layer 3 and 4 mean? —The preceding unsigned comment was added by 62.145.29.194 (talk) 09:16, 16 February 2007 (UTC).
Just a note here to discuss my changes. 802.11, 802.11a, 802.11b, 802.11g, and 802.11n each have their own PHY, MAC and LLC entities. I.e. 802.11a/b/g/n are not only PHY layer specs, but include MAC/LLC also. And, the 802.11-97 (legacy) standard also included MAC/LLC and actually three different PHY layers: FHSS, DSSS, and Infra-red.Short description of the standardKgrr 23:33, 6 April 2007 (UTC)
There used to be a lovely [[2]] on this matter on this page. This was removed. Anyone know why? It was really useful for non-engineers to follow.—Preceding unsigned comment added by 64.194.250.99 (talk • contribs) 22:06, 18 June 2007 (UTC)
OSI architectural development did not stop with the basic OSI reference model, ISO 7498. The original work at all layers, especially Network, was strictly connection-oriented. That reflects the circuit-switched predisposition of the telephone engineers that originally developed the model, and their realization of packet switching as X.25 virtual circuit switching.
The first addendum to the OSI Reference Model proper, 7498/1, introduced connectionless communications, as in the Internet Protocol. A later document, ISO 8648, "Information processing systems - Open Systems Interconnection - Internal organization of the Network Layer (IONL)", came in 1988. The IONL document attempted to make real-world protocols fit the OSI network layer, and recognized the reality that the strict modeling of Network and Data Link didn't work. I'll introduce a simplified version. Since it costs USD $102 to download the IONL document, you'll have to trust my memory.
sorry, I don't know how to edit a Wiki table. Feel free to fix the attempt below
Abstract name................Connection-oriented stack..........Connectionless stack
Subnetwork[note 1]...........X.25 packet layer, more or less....IP or ISO CLNP Independent
Subnetwork-dependent.........None (in reality, static mapping...ARP (for 802.2 LANs) Convergence..................between NSAP address and X.121
Subnetwork access............LAP-B..............................IEEE 802.2
[Note 1: In OSI-speak, "subnetwork" is a specific technology at data link and below. It is not the same as an IP subnet]
To make things worse, real-world protocols can recursively apply these sublayers. ATM LAN Emulation, for example, at the emulated level did an ARP over LLC, but then used ATM protocols (mercifully forgotten) to map the MAC address to an ATM address.
Routing protocols, incidentally, are in a parallel stack for layer management at the network layer. See the OSI Management Framework, ISO 7498/4, and the OSI Routeing [sic] Framework, ISO/TR 9575.
People really need to recognize that IETF and IEEE protocols were not designed to fit the OSI model, as much as basic networking instructors try to force, coerce, or shoehorn them into a 7-layer oversimplification.
To give a little personal context, I was a member of the Federal Telecommunications Standards Committee in the late seventies, when the ANSI Distributed Systems (DISY) project became OSI. Later, I was the first technical staff member of the Corporation for Open Systems (COS), the industry group for promoting and testing OSI and ISDN protocols. COS knew OSI was the answer, but we were still trying to figure out the question when IP protocols, for many reasons, passed us by. I was directly involved in OSI architectural discussions including liaison with IEEE 802, and later was the team leader for developing the CONS/CLNS/X.25 conformance test system.
Hcberkowitz 03:40, 20 June 2007 (UTC)
Read this from the introduction:
Even though newer IETF and IEEE protocols, and indeed OSI protocol work subsequent to the publication of the original architectural standards that have largely superseded it
Even though what verb? Does the writer mean "Even though there are newer...", or maybe "standards have largely superseded...".
I would change it myself but I don't get what is meant. Thanks
--Sukkoth 05:41, 20 September 2007 (UTC)
My removal of the metaphor example in the Layer 3 section was contested, and I'd like to discuss it here. While the latest rewrite has dispensed with any informality, I still don't think it's optimal. Specifically, the whole second paragraph in the network layer section discusses the Internet Protocol rather than the network layer. I think that the information contained within that paragraph belongs in the Internet Protocol article itself. I also think that an example in that section should explicitly discuss only Layer 3 of the OSI model. -FrankTobia (talk) 21:41, 9 December 2007 (UTC)
Fully agree and must add that the figures in this article do not match with the pure OSI model (ITU-T X.200). Sadly the IP community spoiled the elegant OSI model, and today if the OSI model is referred actually a mixture of IP and the original OSI model is commonly meant. This stub of a model has no meaning as it lacks a clear structure; it is simply a way to hierarchically classify protocols. The OSI model per se fits to Ethernet, SDH, ATM, IP, etc. equally; all these offer a Layer 3 responsible to offer an end-to-end connection. That IP lacks it's own lower layers demands to use other network layers as so called sub-networks to create an IP network. See the ITU-T X.200 reference (page 41ff) on the concept of sub-networks and how these integrate nicely to provide a clear view.
The Bit, Frame, Packet etc. termini must be cleared from the figures - these are not in-line with the OSI model, these refer to specific implementations. It is evident that every Ethernet LAN has all Layers, accordingly Ethernet frames travel end-to-end from application to application. —Preceding unsigned comment added by 85.127.103.152 (talk) 11:10, 24 May 2010 (UTC)
Please do not put these in the application layer, where they do not belong. Let me give the relevant background, with the caveat that IETF/IP protocols do not follow strict OSI layering.
The original OSI Reference Model, ISO document 7498, did not address management at all. Annex 7498/4, however, introduced system management (e.g., CMIP and SNMP), as well as layer management. {| class="wikitable" |- ! Layer ! User information protocol ! Layer management protocol |- | Network | IP, CLNP | OSPF, IS-IS, RIP, BGP, PIM, DVMRP, etc., are all layer management between routers. ICMP, IGMP, and ARP are host-to-host or host-to-router, but still layer management |}
Other ISO documents further elaborated. The Internal Organization of the Network Layer document, ISO 8648, defined three sublayers, not counting recursive tunneling such as ATM LAN Emulation:
Sublayer | Function | Representative protocol |
---|---|---|
Subnetwork independent | Agnostic to underlying layers | CLNP, IP, CONS |
Subnetwork dependent convergence | Maps between logical network layer address and medium specific address | ARP |
Subnetwork access | Media-dependent (overlaps with L2 in some cases) | LLC, LLC/SNAP |
The OSI Routeing Framework, ISO/TR 9575, further clarified (using a European spelling for "routing") that router-to-router protocols are layer management protocols in the network layer.
For the record, I was the team leader for X.25 and ISIS protocol conformance testing at the Corporation for Open Systems, and subsequently work in the IETF OSPF and BGP (Inter-domain routing) working groups. I was the lead author of RFC 4098 on BGP control plane convergence.
If you want to move routing protocols into the application layer, please give authoritative sources that override the references I have just given.Howard C. Berkowitz (talk) 18:01, 13 December 2007 (UTC)
It seems sort of out of place on a wikipedia article but if it is there why not have one that goes in the order of the layers. We were told to remember All = Application
People = Presentation
Seem = Session
To = Transport
Need = Network
Data = Data Link
Proccessing = Physical
which goes in the protocol numbers from 7 to 1 which is better than the sausage pizza one which isn't really ordered TheGreatZorko (talk) 09:19, 31 January 2008 (UTC)
changed dhcp from network to application. It is where it should be. —Preceding unsigned comment added by Hellomarius (talk • contribs) 00:25, 12 February 2008 (UTC)
I'm talking about the comparison of the Transport layer to the Post Office.
First off, it doesn't really explain anything. Before the author actually makes his point, he then begins to ramble on about tunneling and GRE. There is probably a better way to visually explain this layer, but personally I don't know that this layer needs any visual explanations made at all.
Secondly, a "Post Office" comparison would be better suited for the Network layer. :-) Subfrowns (talk) 20:31, 22 February 2008 (UTC)
References show that the SPARC/DISY work in ANSI was redirected to work on the newly formed ISO Open System Interconnect, not that the OSI Model work "started" in ANSI SPARC/DISY, even if personal papers (box of files) were a verifiable source, which they are not.
The term "OSI" could hardly have come into use on 12 October 1979 (that just the date on the box of paper, I believe), when ISO created TC97 SC16 charged with developing an "Open Systems Interconnect" architecture framework in 1977, and the "OSI model" was ratified by ISO in July 1979.
I have removed these two items which could not possibly be correct. Dgtsyb (talk) 04:41, 19 June 2008 (UTC)
I removed all reference to this person per WP:BLP. If someone has a problem with that, please let me known and I will file it on WP:BLPN. Dgtsyb (talk) 05:58, 19 June 2008 (UTC)
The colors in the OSI model image (Image:Osi-model-jb.png) on this page could need a change, since they are bad from the perspective of someone with any common color deficiency. Especially colors in the "Transport" and "Network" fields are hard to tell apart. Would fix it myself if it wasn't that I have both a green and red deficiency and would probably just mess it up. There's a chart with safe colors here and some more info here. -Manwal 23 June 2005
I think that the spam links are gone, so I removed the cleanup spam tags. Dgtsyb (talk) 04:46, 19 June 2008 (UTC)
Give your vote here. Should the TCP/IP model template have four or five layers? And what is the name of the bottom layer in case of four layers? And is it okay to mention both the four and five layer models in wikipedia articles? Mange01 (talk) 18:17, 17 July 2008 (UTC)
Can I suggest add ASN.1 as Presentation in SS7 stack? Asn1tlv (talk) 07:54, 4 September 2009 (UTC)
OK, someone removed the additions of several more examples of stacks. Even when they have been placed in a separate section.
No talk, no improvements, just removal. This is not the way.
The current page contains no examples of OSI5, only one of OSI6 (ASN.1 with a not very good description). Interestings comments in this discussion about HTML/HTTP, SIP, ... have been ignored.
Nowadays that OSI3 and OSI4 are dominated by TCP/IP, most part of the improvements in network protocols are in OSI1-2 and in OSI5-6. But this page ignores them.
Lots of people doesn't understand the objective and usage of OSI5 and OSI6 layers. Moreover, they are not able of made a diferentiation between OSI7 and the final application. In conclusion, they decide OSI is not important, something of the pass.
It is clear this page will not help them. —Preceding unsigned comment added by 81.38.181.95 (talk) 16:06, 4 September 2009 (UTC)
Although the OSI was never implemented in its protocals, it standardization efforts in defining "layers" was accetped. This article really misses the entire point.. TCP\IP only encompasses layers 4 and 3 of the OSI, protocal wise. And its perscription for the other layers lines up exactly with the that of the OSI. The OSI is much broader and its standard model is the basis for almsot all internetworking in the world today. You cannot take a networking class without learning the OSI model, it is the foundation upon which almost all internetworking is built.. The article seemingly dismisses the OSI as a failure, when in fact it one of the most critical and important things to understand in networking. —Charles Edward (Talk | Contribs) 18:39, 29 September 2009 (UTC)
L2TP belongs to OSI Layer 2. L2TP just provides a tunnel that is then, for example, used by IPSec for for implementing a secure transfer of the packets: L2TP packet is wrapped and hidden within the IPsec packet. Also, by googling for L2TP OSI Layer it seems that Wikipedia is the only place where this protocol is placed at Layer 4. Also removed "L2TP carries PPP frames inside transport packet." from the description of Layer 4. —Preceding unsigned comment added by 195.145.148.246 (talk) 13:43, 5 January 2010 (UTC)
Most VPN technologies don't fit the OSI model well. VPN protocols provide a virtual layer 2 or 3 tunnel, but to do this they require the upper layers. L2TP provides a virtual layer 2 connection, but it requires layers 3 and 4 to establish the connection and provide transport for the tunnel. 69.68.125.6 (talk) 13:35, 4 August 2010 (UTC)
(First wiki anything... So go easy!) The OSI Model table at the top of the page indicates that SSL/TLS are transport layer protocols. This is not the case, can they be moved up to the Application layer? OzBluBoy (talk) 14:57, 18 April 2010 (UTC)
You're almost right - they belong to the Presentation layer (i.e. cyrpto) - This is where they were in the Examples table so I've changed it to maintain consistency (and accuracy)
You're both wrong, it belongs in layer 5, that is why it is called Transport Layer Security, it secures the Transport layer below it. Although, technically, it's part of the TCP/IP model which sometimes doesn't map too well when something is "Layer 4" in that stack. — Preceding unsigned comment added by 128.229.4.2 (talk) 14:04, 20 July 2011 (UTC)
I am currently studying the OSI standard documentation from the ISO (as linked at the end of the article) and have found a number of things that are different from what I have read and heard elsewhere. I am going to start a list here and when I have time I will make corrections in the article.
This list is not complete and I will add to it later as needed. Rsduhamel (talk) 14:50, 23 April 2010 (UTC)
Actually, it may be a while before I get a chance to put any of this into the article. If someone else wants to double-check me and put it in, be my guest. Rsduhamel (talk) 17:58, 3 May 2010 (UTC)
Recent edit comment by Kbrose (talk · contribs), "Main links not needed when the link is at beginning of first sentence". Is this in the WP:MOS somewhere? --Kvng (talk) 22:51, 26 August 2010 (UTC)
The Article includes SIP in both the Application Layer (Layer 7) and the Session Layer (Layer 5), both in the table and in the right-hand-side box. Is this and error or was this deliberate? The Wikipedia page for SIP and RFC 3261 both state that SIP is an Application-layer protocol. Rahul (talk) 07:47, 22 November 2010 (UTC)
SPDY is listed in the session layer. But the main article for SPDY describes it as an application-level protocol (and it solves the same problems as HTTP/HTTPS, which are app-level). Does someone know if its placement in session layer is correct? — Preceding unsigned comment added by 98.217.176.221 (talk) 17:28, 28 May 2011 (UTC)
Many questions raised in the Discussion would be answered by including a column in the Examples table, for IETF RFCs. An IETF columns belongs in it, anyway.
FWIW IETF is the source of information used for writing ANSI specs. ANSI determines what the U.S. contributes to ISO specs. Kernel.package (talk) 18:58, 14 June 2011 (UTC)
This edit request has been answered. Set the |answered= or |ans= parameter to no to reactivate your request. |
The graphics in the can be more legible by using this graphic:
http://bcamca.in/wp-content/uploads/2011/04/osi-model.png
Scottghall (talk) 22:08, 20 July 2011 (UTC)
In the sidebar, "hub" is listed as Layer 1. While a hub is a device that operates at Layer 1, I don't believe physical devices should be listed on the table. Likewise for Network Switch listed on Layer 2.
If devices are going to be listed, where does it end, pots is listed, should we then list all the devices that pots plugs into(DMS, 5E, Telica, PBX, etc...) I don't think that's the route the article should go, let the Ethernet Hub, Ethernet Switch pages describe there functionality in reguards to the Protocol they run, IE 802.2/802.3/Ethernet II, etc....
— Preceding unsigned comment added by KMurphy111 (talk • contribs) 15:26, July 12, 2012
Section Layer 1: Physical Layer, second paragraph – the paragraph actually confuses more than it clarifies. The trouble is that the reasoning "think of Layer 1 as 1 connection, but Layer 2 as 2 or more connections" confuses numerics with meta thinking – it's more informative to explain that Layer 2 peruses and remedies problems occuring in Layer 1. I don't like the paragraph. The section is better without it. Rursus dixit. (mbork3!) 08:33, 27 July 2011 (UTC)
Hi, I don't see why the layers (layer 1, physical layer etc.) should be capitalized. While they certainly are established concepts they are not proper names. I would like to correct that. Thanks, --EnOreg (talk) 10:14, 13 September 2011 (UTC)
BTW, I am surprised that the "m" in the article title isn't upcased, since it's a proprietary name, isn't it. And the opening line is weird: "The Open Systems Interconnection model (OSI model) is a product of the Open Systems Interconnection effort ...". Surely "Model" is part of the name. If it were "An Open Systems Interconnection model", sure, you'd use "m", or possibly downcase the entire item (i.e., one of many such models, or so widespread that it has lost the ownership/specificity that underpinned the original capitals, like hoovering the carpet, not Hoovering the carpet).
BTW, why is "X.200" boldfaced?
PS User:EnOreg pointed me here, since we've been discussing capitalisation in this area, and I've been looking at the sorry state of the MoS subpage on this field. Tony (talk) 02:22, 14 September 2011 (UTC)
Dgtsyb, it's generally considered polite not to start with "You presumed wrong": you also seem to be showing early signs of ownership. When you say, "You will not see any lowercasing even in the talk page", that is more likely to make me draw the opposite conclusion. Please remember that WP's readership is much wider than the experts and professionals in a particular area, and that maximising the use of caps for no good reason aside from the fact that people on this talk page might do it, and some publications haven't thought it through, needs to be questioned. On the specifics, "Layer 7", yes, caps there would be endorsed by a wide range of style guides, and WP's own house style (although many would be happy with the lowercase option). But "physical layer" is not unique to this model, surely? The distinction needs to be made, if so. Tony (talk) 04:13, 14 September 2011 (UTC)
I'll try to summarize the arguments we have so far on whether the layers are proper nouns.
I would argue as follows. The actual OSI protocols were a failure—they never played any practical role in the industry. But the reference model caught on and has been widely used to classify any protocol into one of the layers. Therefore, the layers have become generic networking terms that do not qualify as proper nouns.
A Google search produces corroborating evidence, suggesting that the vast majority of university lectures and research papers spell the layers in lower case. The same is true for the educational material by one of the largest vendors, Cisco. Even more strikingly, the standard textbooks on the subject, most prominently Tanenbaum as well as Peterson & Davie, use the lower case. Unlike standards documents, textbooks undergo rigorous copy-editing by language experts. All this seems to make a strong case for lower-case layers. --EnOreg (talk) 14:05, 17 September 2011 (UTC)
Are there any other arguments we're missing? Or are the arguments I laid out above convincing? --EnOreg (talk) 15:37, 28 September 2011 (UTC)
It's been a month since the arguments for proper nouns have been refuted and no new ones have been presented since. That leaves really no reason not to adapt the article to the spelling commonly found in the literature on the subject. --EnOreg (talk) 15:42, 16 October 2011 (UTC)
Having not read all the above, I simply consulted books, and found that the layer "names" are not typically treated as proper names: [3], [4], [5], [6] generic; [7], [8] capitalize "Application" etc. but not "layer"; in the first page of 10 google book hits for OSI and "Application Layer", exactly zero show evidence of treatment of "Application Layer" as a proper name. For OSI and "Physical Layer", did find one out of ten: [9] (it also includes "The Lemon-Pudding Layer", so clearly a great moist source). Dicklyon (talk) 04:22, 18 October 2011 (UTC)
Lacking any credible support for the assertion "they are proper names", and lots of evidence of generic lowercase usage in reliable sources about the OSI model, I think it's time for a multiple RM on the various capitalized "Layer" articles, and corresponding fixes to this one. EnOreg, you want to start that, or should I when I have time to read WP:RM about to do a multiple move request? Dicklyon (talk) 19:49, 18 October 2011 (UTC)
OK, I have gone ahead and started a multiple RM at Talk:Physical Layer#Requested move, multiple. Dicklyon (talk) 03:46, 19 October 2011 (UTC)
If you look at the table at top right corner of the article, PPTP belongs to "session layer". If you look down in the table with examples, PPTP belongs to the "data link layer". I am not sure, where it belongs more, but please fix it ASAP. --Sena (talk) 08:49, 14 September 2011 (UTC)
I hear many people calling this model the 7 layer model, could someone please redirect that term (seven layer model) to this page? — Preceding unsigned comment added by 89.138.50.78 (talk) 10:03, 10 October 2011 (UTC)
In the section OSI model#Cross-layer functions, the description of Multiprotocol Label Switching (MPLS) states, "It was designed to provide a unified data-carrying service for both circuit-based clients and packet-switching clients which provide a datagram service model." There are two ambiguities with this:
Will an MPLS expert please revise this? Stephen Charles Thompson (talk) 21:19, 24 January 2012 (UTC)
There is now Layer 0 which is seen on ROADM (Reconfigurable Optical Add/Drop Multiplexor) Fujitsu flashwaves do layer 0, as does Cienna CPL. — Preceding unsigned comment added by 50.98.213.201 (talk) 21:25, 21 April 2012 (UTC)
On 3rd graphic (Data Unit -- Layer -- Function), on 2nd Layer "Data Link", on Function column, there is a link to "Physical Layer" that points to a Wikipedia article about "Memory Physical directions", and not "Network Physical directions". I thinks it's more correct to point it to: http://en.wikipedia.org/wiki/MAC_address — Preceding unsigned comment added by 186.130.41.69 (talk) 15:04, 20 June 2012 (UTC)
According to Tanenbaum ("Computer Networks") and as far as I remember, flow control is not part of the physical layer (1), but of the data link layer (2), and also occurs to some degree within the network layer (3). Also, the physical layer has nothing to do with the sharing of resources among multiple users. That's a task of layer 2 (MAC). Layer 1 covers mainly physical concerns (thus the name) like the media (cable, radio, ...) and how a bit is represented there. --GGShinobi (talk) 21:19, 9 July 2012 (UTC)
In the layer 3 section, there is the phrase "The addressing scheme is not hierarchical.". From my basic understanding of networking, layer 3 addressing IS hierarchical, in contrast to layer 2 one.
Regards Francesco — Preceding unsigned comment added by Francesco87 (talk • contribs) 20:16, 22 October 2012 (UTC)
The "history" section doesn't mention even a single date... — Preceding unsigned comment added by 80.174.219.223 (talk) 16:58, 14 November 2011 (UTC)
I came here to say that, what use is a history (an ordering of events) that doesn't provide any dates to put it into context with other contemporaneous events. — Preceding unsigned comment added by 27.252.100.236 (talk) 12:37, 7 August 2013 (UTC)
Also, was OSI ever implemented, and what is the current state of that implementation (if any)? The intro para calls it a "conceptual model," and there's no mention in the History sxn of an actual implementation. My mid-understanding is that TCP/IP became the de facto standard instead. There's some discussion of this on pg 457 of "The dream machine: J.C.R. Licklider and the revolution that made computing personal" by M. Mitchell Waldrop. The discussion makes it sound like OSI (or perhaps support for OSI) was in part a European political response to the fact that TCP/IP was American (not European) and had been developed under ARPA (therefore US DOD) funding. That of course means this question of why TCP/IP is still around, and OSI isn't (if that's correct), is a contentious issue. I don't know how to evaluate that claim. Mcswell (talk) 14:02, 29 May 2013 (UTC)
The information in parentheses (ISO/IEC 7498-1) tells you unambiguously that "The Open Systems Interconnection (OSI) model (ISO/IEC 7498-1)" is an ISO/IEC International Standard. For precision in labeling, that label should include the date (ISO/IEC 7498-1:1994), but the absence of the date does not change its status. TCP/IP is one implementation of two of the layers of that standard. Please read for yourself about the labeling that the ISO uses to specify the status of a proposed standard as it works its way through the standardization process at: https://en.wikipedia.org/wiki/International_Organization_for_Standardization Wikifan2744 (talk) 02:25, 9 September 2013 (UTC)
I withdraw my sentence just above: "TCP/IP is one implementation of two of the layers of that standard." It's much more complicated than that, far more complicated than I can explain. The Internet protocol suite, which includes the TCP/IP networking model and related protocols (TCP, IP, and others), was developed by and is maintained by the Internet Engineering Task Force (IETF): https://en.wikipedia.org/wiki/Internet_Engineering_Task_Force. Note that there are both the model and also the protocols. Models and protocols are separate concepts that are easily conflated or confused in writing and speaking.
The article on the IETF begins as follows: "The Internet Engineering Task Force (IETF) develops and promotes Internet standards, cooperating closely with the W3C and ISO/IEC standards bodies and dealing in particular with standards of the Internet protocol suite (TCP/IP).[1][2] It is an open standards organization, with no formal membership or membership requirements."
The article on the Internet protocol suite (commonly known as TCP/IP) is long and comprehensive. https://en.wikipedia.org/wiki/Internet_protocol_suite Its opening paragraph begins: "The Internet protocol suite is the networking model and a set of communications protocols used for the Internet and similar networks. It is commonly known as TCP/IP, because its most important protocols, the Transmission Control Protocol (TCP) and the Internet Protocol (IP) were the first networking protocols defined in this standard...." Please read the article for further details, but pertinent to this discussion is the section just below.
"OSI and TCP/IP layering differences[edit source | editbeta]
"The three top layers in the OSI model—the application layer, the presentation layer and the session layer—are not distinguished separately in the TCP/IP model where it is just the application layer. While some pure OSI protocol applications, such as X.400, also combined them, there is no requirement that a TCP/IP protocol stack must impose monolithic architecture above the transport layer. For example, the NFS application protocol runs over the eXternal Data Representation (XDR) presentation protocol, which, in turn, runs over a protocol called Remote Procedure Call (RPC). RPC provides reliable record transmission, so it can safely use the best-effort UDP transport.
"Different authors have interpreted the RFCs differently, about whether the link layer (and the TCP/IP model) covers OSI model layer 1 (physical layer) issues, or whether a hardware layer is assumed below the link layer.
"Several authors have attempted to incorporate the OSI model's layers 1 and 2 into the TCP/IP model, since these are commonly referred to in modern standards (for example, by IEEE and ITU). This often results in a model with five layers, where the link layer or network access layer is split into the OSI model's layers 1 and 2.
"The session layer roughly corresponds to the Telnet virtual terminal functionality [citation needed], which is part of text based protocols such as the HTTP and SMTP TCP/IP model application layer protocols. It also corresponds to TCP and UDP port numbering, which is considered as part of the transport layer in the TCP/IP model. Some functions that would have been performed by an OSI presentation layer are realized at the Internet application layer using the MIME standard, which is used in application layer protocols such as HTTP and SMTP.
"The IETF protocol development effort is not concerned with strict layering. Some of its protocols may not fit cleanly into the OSI model, although RFCs sometimes refer to it and often use the old OSI layer numbers. The IETF has repeatedly stated [citation needed] that Internet protocol and architecture development is not intended to be OSI-compliant. RFC 3439, addressing Internet architecture, contains a section entitled: "Layering Considered Harmful".[28]
"Conflicts are apparent also in the original OSI model, ISO 7498, when not considering the annexes to this model (e.g., ISO 7498/4 Management Framework), or the ISO 8648 Internal Organization of the Network layer (IONL). When the IONL and Management Framework documents are considered, the ICMP and IGMP are neatly defined as layer management protocols for the network layer. In like manner, the IONL provides a structure for "subnetwork dependent convergence facilities" such as ARP and RARP.
"IETF protocols can be encapsulated recursively, as demonstrated by tunneling protocols such as Generic Routing Encapsulation (GRE). GRE uses the same mechanism that OSI uses for tunneling at the network layer."
Wikifan2744 (talk) 06:29, 9 September 2013 (UTC)
This edit request to OSI model has been answered. Set the |answered= or |ans= parameter to no to reactivate your request. |
DHCP is a network layer protocol and not application layer. these are all for delivering IP addresses to the node including other information. 194.213.3.4 (talk) 13:18, 31 July 2014 (UTC)
This indicates that the inception of the OSI model was 1984. I'm not sure what an inception is. ~KvnG 18:42, 6 October 2014 (UTC)
The box of networking layers here makes the mistake of conflating TCP/IP network protocols with the OSI layers. This is wrong; the two are not connected, not even remotely.
I realize that many educational institutions like to use the OSI model because it's such a neat network model and it's so nice to explain things, but it's 'wrong' to assume that the IP network protocol is situated in layer 4 of the OSI model, or that the SIP protocol is situated in layer 7 of the OSI model. They are not. The IP protocol is found in the network layer of the TCP/IP "model", and the SIP protocol is found in the applicatoin layer of the TCP/IP "model".
If you're going to show the OSI model, the protocols on OSI protocols should be used instead. — Preceding unsigned comment added by 84.197.94.196 (talk) 23:03, 19 April 2013 (UTC)
the table in Examples gives an entry for "Layer 2" X "TCP/IP protocols". This is wrong. Layer 2 protocols, like PPP, is at a level where IP doesn't exist (IP is on top of L2).
Zenkutsu (talk) 18:42, 9 February 2015 (UTC)
Also things like HTTP and HTML are backwards. HTML is transferred over HTTP, so HTML should be layer 7, not layer 6, and HTTP should be layer 6 not layer 7. The same goes for the other 6/7 layer examples.
Majenko (talk) 23:58, 14 December 2015 (UTC)
In the table in section "Description of OSI layers" there are some interesting examples for layer 6: HTML, CSS, GIF.
How so? Can somebody explain? What does an image format have to do with the presentation layer of the OSI model? And HTML/CSS? This looks like it was mistaken for "visual presentation of a web page".
Similarly SQL is a weird choice for layer 5 because that's a programming language, not a protocol. — Preceding unsigned comment added by Niko IstDerNameImmerNochZuÄhnlich (talk • contribs) 01:35, 5 October 2015 (UTC)
It seems to me that the examples in the upper layers are all completely mixed up. HTML should be a higher layer than HTTP since HTML is transferred using the HTTP protocol over a TCP socket using IP. The 6/7 layer examples are all backwards.
Majenko (talk) 00:00, 15 December 2015 (UTC)
In first paragraph of Layer 6 topic, should "big mapping" be changed to "bit mapping"? Thank you. — Preceding unsigned comment added by 198.28.128.150 (talk) 16:53, 20 May 2016 (UTC)
Numerous text-books and online sources use HTTP; FTP; RDP; SSH; Telnet as examples for Layer 7 (Application).
Even other wiki pages like List of network protocols (OSI model) and Application Layer — Preceding unsigned comment added by Beanaroo (talk • contribs) 09:48, 23 March 2016 (UTC)
The section describing the PHY layer states: "It defines the network topology as bus, mesh, or ring being some of the most common." This is not correct. E.g. there exists both WiFi star and mesh network implementations based on the same PHY's. Another example with radio PHY's you may implement star, mesh, ad-hoc networks Cebmeisner (talk) 12:56, 15 May 2017 (UTC)
The Layer 6: Presentation Layer subtopic claims (without citation) that it is a myth that the presentation layer does compression and encryption. However, the table at the top of Description of OSI Layers mentions compression and encryption as layer 6 functions. Which is it? — Preceding unsigned comment added by Marcelocantos (talk • contribs) 23:05, 13 October 2017 (UTC)
The article says in several places that the OSI model defines a standard. The official documentation specifically says it does not define a standard. Rsduhamel (talk) 16:46, 19 January 2015 (UTC)
according to the interview with Charles Bachman IBMs SNA was one of the blueprints for the OSI model
"I started looking around for useful solutions to similar problems and became familiar with the IBM project called SNA or System Network Architecture. It had some very attractive features. It looked like it was well-architected. It had six architectural layers. It had a principle that each layer only talked to the layer immediately above and to the layer immediately below it. A layer might be modified, or replaced, and its protocol changed, as long as it maintained both its upper interface and its lower interface. And so it was possible to swap out a layer with an improved protocol and a new implementation that would support different things without disrupting how the other layers functioned." http://ethw.org/Oral-History:Charles_Bachman
as well as the experiences with standardisation bachmann made with the IDS project: "Honeywell was still willing to consider the communication add-on to IDS, so I started on that project and called it “Distributed System Interconnection.” And so I took on the project, which I had started in Phoenix, under the name of “Inter-Communicator” https://en.wikipedia.org/wiki/Integrated_Data_Store
http://www.dsbaral.com.np/subject/network-programming/sna — Preceding unsigned comment added by 77.180.0.193 (talk) 22:07, 4 July 2018 (UTC)
I was looking for info on I(ternet)P(rotocol)o(ver)E(thernet), and got this page. But this page says absolutely NOTHING about IPoE. So where should a seeker after knowledge go? 58.7.61.232 (talk) 03:35, 10 August 2015 (UTC)