Talk:Router (computing)

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

motorola cellphone GPRS[edit]

I'am using the motorola phone as a modem and it doing will. The phone is giving the bandwith of about 912.26 mbps but it is not stable, it lapse and at to the point of disconnection. I would like to ask you people to discuss about this matter if the cell phone could be able to use as the modem while using the router. Then what kind of router could possible and compatible to the phone.

Forwarding[edit]

Some of the material in the beginning of the Applications section can be merged into the Forwarding section. I plan to get back to this but it is no problem if someone else gets to it first. ~Kvng (talk) 21:10, 28 September 2017 (UTC)[reply]

The boxes in our homes are not really routers[edit]

The boxes in our homes and small businesses do not fit the traditional definition of a router. They simply forward information from our local networks to the ISP. It is most correct to say they are packet forwarders. They do not connect networks to networks, just a network to a single IP address at the ISP. 162.196.138.163 (talk) 11:08, 20 February 2018 (UTC)[reply]

The usual home router "box" you are referring to integrates router, (often) DSL modem or ONT, switch, and wi-fi access point. For IPv4, they usually NAT but that doesn't make them any less routers. --Zac67 (talk) 11:53, 20 February 2018 (UTC)[reply]
What kind of routing do they do then? They just act as a gateway, which I don't really think is routing, or is it? TheUnnamedNewbie (talk) 15:34, 22 July 2019 (UTC)[reply]
Routing is forwarding between different IP networks (e.g. your local network and the ISP access network). Devices are called routers or gateways interchangingly. --Zac67 (talk) 16:36, 22 July 2019 (UTC)[reply]
When we send a document from our computer to our printer to be printed, our home router is acting as a router. The printer and computer both have different IP addresses and the computer sends packets of data to the router, which forwards them to the printer. Also when a web page is downloaded from the Internet, our home router has to route the packets to the IP address of the correct device on our home network. My home Wifi network currently has 5 cell phones, 3 laptops, 2 printers, a WII, 2 Firesticks, and an Alexa registered. Modern home modem/routers do a good deal of routing. --ChetvornoTALK 16:13, 22 July 2019 (UTC)[reply]
Not usually. Your printer and your computer are commonly in the very same IP network (subnet) and don't use a router to communicate. They're connected by a switch (forwarding by MAC address), often integrated within the home router (or a repeater hub, repeating everything). Even if one node is on Wi-Fi and the other on wired Ethernet, the home router *bridges* between them like an access point, forwarding just like a switch. --Zac67 (talk) 16:45, 22 July 2019 (UTC)[reply]
But, the devices in your home network are on a different subnet than the rest of the internet and so a two-port router (and NAT) in the box are in play when contacting devices and servers outside the home network. ~Kvng (talk) 14:55, 25 July 2019 (UTC)[reply]
Exactly. --ChetvornoTALK 19:16, 25 July 2019 (UTC)[reply]

"Access, core, and distribution" section with failed verification tags[edit]

Hello!
I'm in the middle of work, but I found some possible sources for the section mentioned in this post title. I haven't read the guidelines for updating references yet, but I don't want to forget where I found some sources, so I'm making a note here.
Here are some sources for the access/core/distribution layers of the Cisco three-layer hierarchical model.
http://www.ciscopress.com/articles/article.asp?p=2202410&seqNum=4
https://www.mcmcse.com/cisco/guides/hierarchical_model.shtml
Thanks!
Imperatrixmundi (talk) 23:54, 19 February 2020 (UTC)Imperatrixmundi[reply]

Dictionaries contradict common knowledge[edit]

The dictionaries confirm that it is pronounced "rooter" or "rauter". I just asked a Cisco employee and she was pretty adamant about "rauter" and so was I, but the dictionaries tell a different story here. Elizium23 (talk) 00:39, 15 April 2020 (UTC)[reply]

But this is not a contradiction: the dictionary supports the pronunciation used by you and the Cisco employee, and it also supports the pronunciation used by other people. I'm English, but have lived somewhere else for long enough to be quite confused about it, but I think the BrE pronunciation is generally "rooter", distinguishing it from the tool for shaping wood, called a "router" (which routs). Imaginatorium (talk) 10:49, 15 April 2020 (UTC)[reply]
In the US, it's mostly a north-vs-south thing. Southerners tend to use /ˈrtər/ which includes the inhabitants of Silicon Valley and makes it more popular in the industry. Webster's Dictionary tells us it's /ˈrtər/ for the network device, but then again they're based in Massachusetts. --Zac67 (talk) 17:00, 15 April 2020 (UTC)[reply]
Anecdote for anecdote: Working in the industry for decades in the northern US Midwest and New England, I've never heard anything but /ˈrtər/. 24.31.139.196 (talk) 21:17, 15 April 2020 (UTC)[reply]
I've worked internationally and in technical circles outside N. America I've heard the "rooter" pronunciation. ~Kvng (talk) 13:35, 18 April 2020 (UTC)[reply]

I've reverted Elizium23's recent change. We don't have consensus or even a proposal on this yet. ~Kvng (talk) 14:41, 12 June 2020 (UTC)[reply]

RFC[edit]

Which reliable sources should be used to cite the pronunciation of "router" in this article?

  1. None (status quo)
  2. Merriam-Webster
  3. Dictionary.com
  4. Oxford English Dictionary
  5. Macmillan Dictionary
  6. Technical documents: ???
  7. Other

More than one answer is encouraged. Elizium23 (talk) 15:41, 12 June 2020 (UTC)[reply]

  • 2 and 4 for American and British usages. Elizium23 (talk) 15:42, 12 June 2020 (UTC)[reply]
  • Remove pronunciation per MOS:PRON#Appropriate use. This is a common word and doesn't require a pronunciation guide. The fact that pronunciation differs regionally isn't by itself sufficient reason to include a guide. If you really want, include a {{Wiktionary}} link box in the lead for the interested reader. –Deacon Vorbis (carbon • videos) 16:12, 12 June 2020 (UTC)[reply]
  • None of the above - The note on pronunciation currently claims that the pronunciation is engvar dependent. It is not sufficient to cite an American and British dictionary separately to establish this - that's WP:SYNTH. We need to find a source or two that actually discusses this from all sides. I've tried to find such a source and couldn't come up with anything reliable. Pronunciation is an expected question readers will have so I would prefer to keep this in the article. ~Kvng (talk) 13:31, 15 June 2020 (UTC)[reply]
    For AE, I've found RSes Merriam-Webster and Oxford Learner's listing both pronunciations, specifically for the network device. An interesting discussion points out there's been a shift in pronunciation in the US (or parts thereof) in the 1960s. I think if we do need to include the pronunciation here it should suffice to list the RS for the two AE variants. --Zac67 (talk)
    Merriam-Webster is claiming rau̇-tər for woodworking and rü-tər for networking. I don't think we can get a consensus of editors here to go with that alone. Oxford Learner's just lists the two pronunciations. That's one possible resolution for this: Just acknowledge that there are two pronunciations and don't get into which is used where. ~Kvng (talk) 13:38, 20 June 2020 (UTC)[reply]
  • 3: OED The 2010 3rd edition of the Oxford English Dictionary notes difference in pronunciation: Brit. /ˈruːtə/, U.S. /ˈraʊdər/. See: router, n.6, www.oed.com/view/Entry/272580. Alternatively, the pronunciation could be omitted altogether, since it is dependent on the dialectical pronunciation of "route" and follows American vs rest of the world split on that line.
  • Remove pronunciation It's a common word. Regional differences in pronunciation are not reason to include in Wikipedia. Glendoremus (talk) 16:42, 16 June 2020 (UTC)[reply]
  • Remove pronunciation per what others have already said: "router" is an ordinary English word and according to [[[MOS:PRON#Appropriate us]] we don't give pronunciation for those; that's what dictionaries are for. – Finnusertop (talkcontribs) 03:43, 18 June 2020 (UTC)[reply]
  • Remove pronunciation Phonetic transcription is enought I think. AXONOV (talk) 07:23, 27 June 2020 (UTC)[reply]
  • Remove pronunciation I agree with others that the word is now in common usage and has multiple pronunciations depending up where in the world you are located. I don't know that it really adds to the article to have the various pronunciations here. - Dyork (talk) 00:52, 30 June 2020 (UTC)[reply]
  • Comment - As we're having here, I've had not insignificant discussions with IT professionals about how this word is pronounced. There has also been significant editing associated with this statement in the article over the years. I think we're doing readers a disservice by removing all discussion of pronounciation. I assure you a new editor will fly in shortly after we've closed this discussion and readd something (uncited) about pronunciation. If we kick the can, it's just going to promptly come back to us. ~Kvng (talk) 14:23, 2 July 2020 (UTC)[reply]
  • @Zac67: I think it's good point here.
  • Counter-proposal - as suggested multiple times in this discussion, remove pronunciation. Glendoremus (talk) 15:50, 4 July 2020 (UTC)[reply]
  • 2 and 4 for American and British usages. Per the previous section, there seem to be enough people convinced there is only one correct way for it to be worth showing otherwise. Johnbod (talk) 02:48, 5 July 2020 (UTC)[reply]

End of RFC[edit]

I would assume an uninvolved closer would say there's a consensus to remove the pronounciation. And we can go ahead and do that. I've predicted that it will be restored in some form within a year's time. I wish we could agree on a more durable solution but maybe that's not possible right now. ~Kvng (talk) 20:55, 6 July 2020 (UTC)[reply]

Router definition in intro[edit]

At the very beginning of the article it only reads that the router is just a networking device that performs this and that, however a router is actually a fully fledged computer, and this article's intro omits this important fact for an unknown reason. The fact that it is a computer always used to be a more or less standard definition across textbooks, educational courses, and any other definitions.

The router is a computer in the sense that it has all the main elements of such: it of course has a CPU and RAM, it also has a storage device, input and output ports, it has a fully featured OS that runs on top of that, you can run different apps and utilities within that OS, create and edit files, etc. and etc. This in big contrast to e.g. a switch (a classic unmanaged one) which doesn't have all that and is controlled by firmware (like e.g. your washing machine, which is also not a computer in the general sense).

In fact a router is very little different from any PC running a server OS and remotely accessed (i.e. no monitor and keyboard directly attached). In another fact, the router functionality is essentially realised simply by software on any PC with at least 2 NICs (network adapters), highlighting the fact that the hardware router is merely a computer with hardware optimized for the task, but you could in theory use any desktop PC given that it runs the relevant software.

I propose amending the intro with the fact that a router is foremost a specialised computer.

77.85.6.55 (talk) 10:25, 2 December 2022 (UTC)[reply]

The fact that most routers are specialized computers isn't all too relevant for their router functionality. The majority of digital devices are specialized computers today: smartphones, smartwatches, MP3 players, TVs, Bluray players, managed network switches, firewalls, etc – even cars or bikes. On most routers you can't run generic apps and utilities however, you can only use what's provided in the box. And many routers include hardware for forwarding or filtering, only very simple or virtual routers are software only. --Zac67 (talk) 11:04, 2 December 2022 (UTC)[reply]
I don't agree with these arguments. The important facts that a device is a computer are that you can run arbitrary programs written for that computer, and also that the device is under the control of an OS and not a firmware. Most devices you listed (players, smart-devices, etc., etc.) are electronic stuff, but not computers in the sense that they are running a fixed firmware which doesn't allow starting and running of applications (programs).
You do can run generic apps and utilities on routers, when those are written for that platform and OS, of course (similarly to being only able to run apps made for Windows in the Windows OS), and if the OS is not some closed/prohibitive type. Routers are classic computers, not just some sort of electronics.
Now, of course when I say "router" I don't mean the home "Wi-Fi routers" most private people have, but those aren't true routers anyway.
Also I do not agree that these particulars are not relevant because they don't affect the functionality. We are running an encyclopedia here which is supposed to describe what stuff truly is and how it internally works -- not just describe its "consumer merits" outside of the "black box" of the device.
77.85.6.55 (talk) 11:48, 4 December 2022 (UTC)[reply]
Well, let's agree to disagree and see what other editors might say. --Zac67 (talk) 12:20, 4 December 2022 (UTC)[reply]

I agree with 77.85.6.55, I think it should be characterized as a “specialized computer” in the intro. I feel the present intro isn’t sufficiently clear for nontechnical readers about what a router is. All routers are run by microprocessors. The word ‘computer’ is an adequate umbrella term for what it is, suitable for general readers --ChetvornoTALK 12:38, 4 December 2022 (UTC)[reply]

I can agree with the argument about the generic non-tech reader. For laypeople the umbrella term "computer" is appropriate here, as for them it usually implies a machine that can run arbitrary code (i.e. random apps and utilities), while "non-computer" electronics usually imply that their functionality is fixed (think of unmanaged switches or various modems).
I also wanted to reiterate, that the router function is predominantly a software one (think of a server, like Apache). It is just that manufacturers like Juniper or Cisco sell their router software (e.g. IOS) running atop specialized hardware (just like with the hardware servers), but otherwise all of that (i.e. Cisco IOS, Apache, Microsoft IIS) can run atop an ordinary desktop PC, it's just that it isn't practical.
However, lots of non-computer electronics' (a switch or modem) predominant feature is not a software, but a hardware one. And the software is not the main thing, it is just to enable the hardware, hence it is a "simple" firmware. You cannot practically make a switch out of a PC because the main thing about a switch are the 24 network ports, not its software. Same with a modem -- main purpose is hardware one, not the software.
So by this logic, a router is like a "network server". You wouldn't call a server a "networking device" or some other ambiguous term, would you? Servers are outright computers that run special server software. Same with routers. They are outright computers running special router software. In both cases that software runs atop a normal, albeit specialised, OS, and in turn this OS, generaly speaking, can run on any general-purpose PC.
77.85.6.55 (talk) 14:33, 4 December 2022 (UTC)[reply]
I'm entirely OK with mentioning that a router is (usually) a specialized computer, although I fail to see the importance. However, some of your claims are plainly wrong, including
  • the router function is predominantly a software one – Many/most enterprise-grade routers use a hardware forwarding plane, just the control plane is software. They're very similar to a layer-3 switch with added functionality, much of which is realized be using TCAM hardware. You simply can't run a BGP router with the entire global routing table efficiently in software, not even today.
  • It is just that manufacturers like Juniper or Cisco sell their router software (e.g. IOS) running atop specialized hardware (just like with the hardware servers), but otherwise all of that (i.e. Cisco IOS, Apache, Microsoft IIS) can run atop an ordinary desktop PC – partially true but very often even virtualized routers use special hardware features present in the specialized hardware, e.g. Cisco ASA on Firepower. And you can't run IOS on generic hardware.
  • You cannot practically make a switch out of a PC because the main thing about a switch are the 24 network ports, not its software. – you can add as many network ports as possible into a standard PC (a dozen doesn't seem hard) and run a bridging software, making it a (software) switch. Also, virtual switches used by hypervisors are software generally.
  • Same with a modem -- main purpose is hardware one, not the software. – Mind you, some time ago telephone modems were added using AMR or CNR/ACR boards, running most functions on the host CPU, in software. DSL/cable modems very often include management functions, requiring some kind of CPU.
The fact that you can use a standard PC with a bit of software as a router doesn't mean that a router is nothing more than a PC. 90%+ of all routers are purpose designed and there's no way to run arbitrary software on them. It's really not all black and white like the distinction you seem to be trying to make (somewhat arbitrarily imho). --Zac67 (talk) 15:49, 4 December 2022 (UTC)[reply]
Zac67 has described the situation well. Is there a proposal for precise wording or an edit with the proposal? If someone produced a box which performed routing functions yet was not a specialized computer, the box would still be a router. It may be true (given the number of home appliances) that most routers are specialized computers but it's not a defining characteristic. Johnuniq (talk) 06:51, 5 December 2022 (UTC)[reply]
Yeah, you're right in your comments Zac67, and I was aware of some of the facts you mentioned. However, your comments are more on the practical side, while I was purposefully trying to pinpoint the principal (theoretical) difference between a specialized computer and "non-computer electronics". Think of a class in Theory of computation where they don't discuss any practicalities at all, just the underlying fundamentals.
To clarify what I say here -- yeah, you couldn't run a router fully in SW, or you could build a PC with 30 NICs, but that doesn't change the primary theoretical fact, that routing is a programmatic (applicational) task and the routing hardware is just an optimization to help with that. While layer-2 switching isn't like that, the SW there is just to control the hardware "connections" between the ports. Think of the washing machine. It has a SoC and firmware (much like any non-computer electronic device), but its task is mechanical (washing your clothes) and the SoC is merely a controller. The washing machine isn't a computer because of that (because it has a SoC).
Please concentrate more on the theoretical differences, and not the practical ones, because we're trying to get the fundamentals right, not the practicalities, Johnuniq.
(about the modems, I meant dsl/docsis standalone devices)
77.85.6.55 (talk) 11:23, 5 December 2022 (UTC)[reply]
Sources:
--ChetvornoTALK 07:14, 6 December 2022 (UTC)[reply]
I work professionally near the cutting edge of this stuff and I agree with Zac67's description of the state of things. But, hey, this is Wikipedia and we go by what the reliable sources say and in that sense, Chetvorno makes a strong point. So, assuming Chetvorno has not cherry-picked sources, everyone's right here. Sources are lagging or fail to appreciate the state of the art. ~Kvng (talk) 15:56, 11 December 2022 (UTC)[reply]
Be my guest, check my cites.
To me, this is an example of a very common debate on WP on the level of terminology to use in the introduction: jargon or plain language. Of course someone who works in the field like Kvng is not going to call a router a “computer”, he uses more specialized terminology. But the introduction should be written for general readers. Virtually every router is run by software, even if it’s only microcode. It is a “computer” in the broadest descriptive sense to nontechnical readers.
I agree this is not consistent terminology, as pointed out above if software is the criterion, washing machines could be called computers. Contributing to the router’s “computerness” to ordinary people is the fact it is mounted in a rack in a data center, or next to a PC, plugged into networking cables, and does computer networking tasks, not wash your clothes.
--ChetvornoTALK 17:49, 11 December 2022 (UTC)[reply]
I souped up the 'specialized computer' text a bit, from half a sentence to a short paragraph. I think it's quite appropriate in the lede but to not give undue weight I've moved the text to its end. Surely there's room for improvement but I think the additional context is worth it. --Zac67 (talk) 19:17, 11 December 2022 (UTC)[reply]
In practice we call this class of equipment Network hardware. I don't think that's too jargony. If you want to emphasize the computer in them, they are Embedded systems which is jargony but we have a bluelink to help; that's a more useful description than specialized computer. ~Kvng (talk) 15:14, 12 December 2022 (UTC)[reply]
My main point in all of this is that we should put fundamentally/theoretically correct descriptions into the intro of the article, not practical ones. As Chetvorno says above, we can get a host of cutting edge field practitioners who will speak professional jargon and describe things from the practical side, however I think we should rather take the opinion of computer scientists who work "with the papers" at a desk in a plain room (not tech guys from a NOC). In other words, it doesn't matter at all whether the subject of our discussion here makes any practical difference. Please understand this last sentence. If it looks like a router, smells like a router and behaves like a router it doesn't mean at all that it is a router. So this is one half of our discussion here.
Now the other half. I think I have sufficiently argumented the implied difference between a "computer" and "microprocessor-powered electronics", and that difference is not the presence or absence of any code or a processing unit. (but whether the computing machine is there to sit and wait to run "arbitrary" software written for its (also present) OS, or it is a firmware-powered controller of some different sort of activity). So I think I'm not going to argument this anymore.
Also, the whole point of writing that a router is a specialized computer is exactly to highlight the fact I just mentioned above, and I believe this is very close to a layman's understanding of a computer, as marking a router a "computer" brings their understanding of a computer much closer to the router's abilities, than marking it a "device" or similar.
I am sorry I am not giving any concrete or real-world examples like the guys above, I am doing this on purpose to depart from any practical implementations.
77.85.6.55 (talk) 13:27, 12 December 2022 (UTC)[reply]
@Kvng: I have no objection to having the term network hardware in the intro, as you say it should be there. But as a sole description it is too jargony, it is insufficiently specific and won't mean anything to many readers. To many (probably most) laypeople, anything from a coaxial cable up would be "network hardware".--ChetvornoTALK 16:14, 12 December 2022 (UTC)[reply]
@Chetvorno It would be a stretch to classify wiring as network hardware. Using the common synonym network equipment instead may make the scope of the term more clear. Your references clearly demonstrate that specialized computer is a common description and so I'm not opposed to using that term but I just wanted to be clear that this is not a term I or my peers that practice in this area use with any frequency. ~Kvng (talk) 16:55, 12 December 2022 (UTC)[reply]
@Kvng: You know that and I know that, but the majority of readers coming to this page, who are not familiar with networking terms, won't know that. So describing a router as a "networking device" tells these readers almost nothing. Thats why "specialized computer" needs to be in the lead sentence. --ChetvornoTALK 21:53, 12 December 2022 (UTC)[reply]
@Zac67: I think the term "specialized computer" should be in the lead sentence, to tell readers what a router is. This is clearly not WP:undue weight, as the 6 sources I gave above have it in their defining sentence. --ChetvornoTALK 16:14, 12 December 2022 (UTC)[reply]
I've improved this new paragraph with a link to Embedded system and trimmed stuff that is not well supported in the body of the article. ~Kvng (talk) 15:44, 17 December 2022 (UTC)[reply]

I see that after so many people and arguments in favour, and time elapsed, the lead sentence still doesn't name the router as a computer :D At least the last one says that... 77.85.6.55 (talk) 10:14, 17 July 2023 (UTC)[reply]