Talk:Client–server model

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Computing / Networking / Software (Rated C-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.
C-Class article C  This article has been rated as C-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 Top-importance).
Taskforce icon
This article is supported by WikiProject Software (marked as High-importance).

Evolution of the client-server model[edit]

What about the history of client server and how the meaning has changed ?

Justin Skinner of Wachovia Corporation[edit]

"computer networking is a technology created by Justin Skinner of Wachovia Corporation."

Yeah, sure it is. RickScott 20:14, 16 August 2005 (UTC)

The initial definition ("Client/server is network architecture which separates the client from the server") is lousy. This is tountamount to saying that a c/s architecture is an architecture that involves a client and a server, which is not saying much... B. Oct 13, 2006.

In the tiered section - "1.It puts a greater load on the network." Can someone please expand on this?

Also, when explaining 3 or n-tiered set-ups, I'm actually trying to just clarify that this can be on the same physical hardware, or can be on different server hardware, and also I think this entire page just needs lots of clarification, and isn't very good at explaining the whole idea very well at all...

Advantages: Fragment?[edit]

"It is more flexible than P2P paradigm. If a server in C/S paradigm wants to update the data or other resources."

The second part of this paragraph doesn't make sense to me. The server is more flexible: If the server wants to update, then what?

Somebody may fix this as I don't know what this should say :-S --Lazer erazer 15:25, 5 March 2007 (UTC)

what is the significance of / in client/server?[edit]

any one who got ans for what purpose / is used in client/ it just for simplicity of writing or got any significant meaning? if it got plz those who got ans...tell it... or mail me at thank you —The preceding unsigned comment was added by (talk) 15:16, August 23, 2007 (UTC)

Multi Tier Architectures[edit]

1) what is use of multitire Architectures ?and how it works.

2) what is difference between multitire Architectures and three tire Architectures.

Needs much improvement[edit]

This article is an important concept, yet is is currently barely better than an stub. The lead needs to be clarified and rewritten to be more concise. The lead is also bloated with examples that ought to be the moved and improved under a new heading. Comparison with other kinds of architectures seemss appropriate to me but needs elaboration and citations.

Also, the current article only applies the term client-server to networking situations. While this is certainly the most common meaning, I frequently see "client-server" used among computer programmers in a more general sense to divide any system into the client components that want data or services and the server components that offer data and services regardless of whether the data is transmitted over a network or not. But my experience with this as a programmer is mostly anecdotal and I'm unsure if this alternate meaning is just a loose usage of the term for an analogous but different concept or if the more generalize concept is a more proper definition of client-server for which networking situations are just the most popular example.

I'll do my best to start a facelift for this article. Help appreciated. Voteformike (talk) 03:51, 22 June 2008 (UTC)

Suggestion: if there is a standard form with either a hyphen or slash joining the terms "client" and "server", it should be used throughout. If there is no known standard, at least choose one and apply consistently. Both forms are used at different points in the article, as well as some that capitalise both terms. --ozNoz (talk) 21:53, 9 July 2008 (UTC)

Agree. Seems to be very weak and confusing article. —Preceding unsigned comment added by (talk) 03:19, 13 June 2009 (UTC)

Server can't initiates requests?[edit]

In "Characteristics of Server" section it is written that servers never initiates requests or activities. While in line no. 8-9 of second para from start, is given an example of a server application forwarding request of its client program to another client program. Could anybody explain this?

At least spell the term consistently![edit]

The article switches sloppily between "client-server" and "client/server". Assuming that the title of the article was given more careful thought than the body--and perhaps some of the body has been contributed by people other than whoever it was who first wrote the title--I guess the hyphenated spelling should be adopted throughout. —Preceding unsigned comment added by AndrewWarden (talkcontribs) 10:43, 22 July 2008 (UTC)

I ran a simple find/replace on the page and replace 6 instances of "client/server" with "client-server". There might be one or two other ones left...I'll fix them if I notice them. — FatalError 03:50, 3 January 2009 (UTC)


As is, the current Advantages and Disadvantages sections are pretty POV and unnecessary. Since virtually all of the items on the lists compare the client-server model to the peer-to-peer model, I think the sections should be merged into a "Comparison with peer-to-peer" section. Thoughts? — FatalError 03:47, 3 January 2009 (UTC)

This section could use some serious help. Right now "Increases productivity" and "Developer productivity" are listed as advantages. I'm removing these until someone can put an actual explanation for these. —Preceding unsigned comment added by (talk) 03:47, 26 May 2009 (UTC)

Developers view[edit]

From a developers view the relationship between client and servers are very simple.

The server have a "listen" function, were it sits and waits for incomming connections The client have a "connect" function in witch it make a connection to the server.

After the connection is established the to parties are *completly* symetrical. And this is important! Anyone can initiate a request to the other. No one have "priority". Normal though is that one of the parties have more "responsibility".

Mutch of other client/server wording is marketing foo. Servier=mutch money vs Clients=not so expensive.... —Preceding unsigned comment added by (talk) 16:03, 24 March 2009 (UTC)

Merge proposal[edit]

User M added a merge proposal for this article to network architecture, which I removed. This is grossly incorrect. The client server model obviously has nothing to do with the architecture of a network, but is a model for Application Layer service delivery. Kbrose (talk) 20:02, 14 June 2009 (UTC)

Any objection to merging it to overlay network architectures, then?   M   20:13, 14 June 2009 (UTC)
Why are you searching for something to merge this into? That article does not exist, btw. Client-server computing is a well understood and notable term that does not need to be merged into anything. If you are thinking of the concept of overlay networks, because they are often implemented in conjunction with P2P systems, then you have to understand how the model of networking in the Internet, see Internet Protocol Suite, intentionally separates the realm of the network functions from the transported application protocols. So C/S and P2P are distinctly separate topics from the underlying network, whether conducted over IP directly, or any kind of overlay network. Kbrose (talk) 22:23, 14 June 2009 (UTC)
Because both this article and peer-to-peer define themselves by explaining how they differ from each other. There's a section here that does this, and p2p spends most of its time talking about clients and servers. If the easiest way to describe how something works is to explain how it differs from other things, then it might help to just explain all this on one page, instead of explaining the same thing twice. The more important goal, though, is to separate the network concept of p2p from the cultural/legal/etc. file sharing concept. A lot of editors here seem to have these two entirely confused.   M   22:49, 14 June 2009 (UTC)
Not all client-server networks are overlay networks. So, no. They're correlated, but not the same thing.- (User) Wolfkeeper (Talk) 22:56, 14 June 2009 (UTC)
You have an example of an application-layer client-server network that is not an overlay network?   M   23:03, 14 June 2009 (UTC)
Do you have a reference that they are the same thing?- (User) Wolfkeeper (Talk) 23:39, 14 June 2009 (UTC)
That what are the same thing?   M   23:56, 14 June 2009 (UTC)
Do you have a reference that client-server networks are inherently considered to be forms of overlay networks?- (User) Wolfkeeper (Talk) 00:00, 15 June 2009 (UTC)
If you can't back up the claim that "not all client-server networks are overlay networks", then don't provide it as an objection. If overlay networks are "models for Application Layer service delivery", then you seem to be objecting to Kbrose on this one, not me.   M   00:18, 15 June 2009 (UTC)
I've worked on projects to build client-server systems, that were clearly not overlay networks. But I'm not making the positive claim and editing on it, you are. Do you have a verifiable reference to back up your claim?- (User) Wolfkeeper (Talk) 00:26, 15 June 2009 (UTC)
No. Kbrose states that client-server is "a model for Application Layer service delivery". The application layer runs on top of other OSI layers. This is what it means to be an overlay network. Therefore, client-server is a model or type of overlay network. I'm not even involved at this point, and I have no idea where you see me making some original claim.   M   00:37, 15 June 2009 (UTC)
OK, the fact that you don't actually know what any of the words you write precisely mean explains everything.- (User) Wolfkeeper (Talk) 01:11, 15 June 2009 (UTC)
I've found that the best indicator of when someone is utterly failing or incapable of understanding the issues is when they limit their responses to vague generalities and blatant attacks on those who disagree with them.   M   19:05, 16 June 2009 (UTC)
You're unable to come up with a reference, because client/server architectures are not a type of overlay network, they can be implemented without being overlay networks; and they are not defined in terms of overlay networks.- (User) Wolfkeeper (Talk) 19:29, 16 June 2009 (UTC)
Since my statements are being referenced (twisted) here incorrectly: No, M's definition is not what an overlay network is, and neither is c/s or P2P. Overlay networks can be a lot of different things, but what M is 'deducing' here is complete rubbish. Furthermore, P2P and C/S don't define themselves in terms of the other, as M states, but it is very useful to contrast these two since they are prominent examples of distributed application design. That does not warrant any mergers. Kbrose (talk) 18:22, 16 June 2009 (UTC)
This isn't 'twisting', your statements are simply unclear and imprecise. For example, you define client-server as a 'model of design' (way of designing? design method?). It's a type of architecture. Figure out what you want to call it. After you do so, it might be useful to discuss these two as part of an article on 'distributed system architectures' (or whatever). It would waste less time on an introduction, and on repetition in the body. There's no rule that says we need one article per concept, and it is ridiculous to try and enforce such a rule if we can cover these more coherently in one article.   M   19:01, 16 June 2009 (UTC)
Um yes, there is a rule that says that there is one article per concept. It's called WP:Wikipedia is not a dictionary; each and every article is on one topic.- (User) Wolfkeeper (Talk) 19:26, 16 June 2009 (UTC)
If we had two articles on two types of elephant, and no elephant article, we might merge them there. "Types of distributed computing architecture" is a single topic. Kbrose finally, and apparently accidentally, brought up a good point against the merge - there are about 6 such articles, and merging them all would be the right thing to do, but wouldn't work.   M   05:15, 18 June 2009 (UTC)
It has become abundantly clear that not only do you not understand what an encyclopedia is, you've no clue what an overlay network is either. Kbrose has got it bang on.- (User) Wolfkeeper (Talk) 19:26, 16 June 2009 (UTC)
I don't make any reference to an overlay network in the message you're responding to, so I'm not sure what your message is about (besides you venting, which I guess is ok too).   M   05:15, 18 June 2009 (UTC)
Since M obviously hasn't done any reading of existing articles, more inquisitive users will have no difficulty in simply following the wikilinks in this article to distributed computing where this is summarized already. Kbrose (talk) 19:43, 16 June 2009 (UTC)


Should this be "Client–server model" instead of "Client–server architecture"? Or "Client–server [some other noun]"? Stephan Leeds (talk) 02:29, 23 April 2010 (UTC)

Comparison to peer-to-peer architecture[edit]

"Both client–server and peer-to-peer architectures are in wide usage today. Details may be found in Comparison of Centralized (Client-Server) and Decentralized (Peer-to-Peer) Networking.", I'm not personalty very knowledgeable of wiki standers but this dosn't even seem to pass the smell test as I ascertain from my consumption of Wikipedia articles.-- (talk) 16:25, 26 October 2010 (UTC)Forrest

What you quoted indicates that decentralized networks are peer to peer networks. This is wrong, whether a system is centralized or not is a different dimension. Peer to Peer networks may have centralized components, Napster for example had a centralized server to search for files, BitTorrent uses trackers - and both networks are by no means considered Client-Server (if you go by , they are hybrid peer-to-peer). (talk) 11:57, 4 April 2011 (UTC)

Peer-to-peer architecture vs. Client/Server[edit]

The reference to "Peer to Peer vs. Client/Server Networks" needs improvement, it is unclear and unfounded what the author was referencing. A simple search on Google for that term gives articles about home network configurations. While it is true that these can be set up to either use a server, or simply remain "peer-to-peer" (ad-hoc?) for many given services, their scale doesn't compare to P2P networks such as the now-closed Gnutella. Scalability is considered a big benefit to P2P over C/S, so I personally doubt that using a home network with five systems is doing it justice.

Maybe the author meant to reference a specific article that compares more than mere home networks. In any case the citation gives almost no clues at all about what article they meant. (talk) 11:52, 4 April 2011 (UTC)

I think you are confusing an architecture with examples of the architecture. Gnutella and some home networks are examples of peer to peer but what that section of the article refers to is simply an alternative architecture to client server. In client server the relation between the two devices is asymetric, the client primarily takes information and the server primarily provides it. Peer to peer is just another way to say traditional distributed computing, where both devices, actually I should say both systems since it's not specific to one piece of hardware, you could have a client server model between two processes on the same machine, but peer to peer communication just means neither process is defined as the "master" or "slave" they are equal. I thought the section made sense as written but I agree a reference would be nice, I'm going to look through my books and see if I have one or two. --MadScientistX11 (talk) 22:30, 19 June 2014 (UTC)

The bind disambiguation page links here, but "bind" isn't in this article[edit]

The bind disambiguation page includes:

  • bind a client to a server in client–server computing

and links to this page, but the word "bind" does not appear in this page. (talk) 20:43, 21 July 2011 (UTC)