Talk:Reverse proxy/Archive 1

Archive 1

List of servers

I've added Cherokee to the "Servers used commercially" section, as I myself am using it commercially (as a web server and reverse proxy). The title of said section suggests some servers should be included, while others shouldn't. Is this correct? Also, I think this section might be more at home in an article of its own (e.g. "List of load balancing servers" or "Comparison of load balancing servers"). PPvG (talk) 21:58, 29 June 2011 (UTC)

Grammar

• Rewrote the first couple sentences. I hope I improved it as it was a bit clunky. I tried to express what a RP does in a way all can understand. (Forgot to log in, it was me) RebelBodhi (talk) 20:31, 25 January 2011 (UTC)

This page contains what looks like advertising

From the article:

1. Fastream IQRP is a scalable (10,000 concurrent users) and secure reverse proxy for Windows. —The preceding unsigned comment was added by 203.153.228.8 (talk) 01:53, 14 December 2006 (UTC).

I agree with you. If there's one thing I HATE seeing in wikipedia, it's advertising (see my edits...). A quick check with google on the name revealed a lot of (shareware) download sites, and no "community"-type websites in the few first web pages. I think best practice is to delete the link. I'll do so immediately Balou 20:30, 28 March 2007 (UTC)

There are other commercial products listed on Wikipedia such as Cisco CacheEngine and Microsoft ISA Server. Small shareware companies should have a chance to express themselves as well since they provide less costly alternatives and fuel competition on the market. I have changed the ad-smelling part of the text and now it seems as viable information for someone looking for a Windows reverse proxy since other than ISA Server, the listed ones do not have Windows GUIs for administration. Also, while ISA Server starts at $1,499/CPU and goes up to $5,999/CPU, IQRP starts at $275/CPU and goes up to only $495/CPU. Plus unlike the statement from "Balou", IQRP has an open yahoo group (you can see their link when you go to their product page).88.226.98.123 05:04, 2 April 2007 (UTC)

I readded IQRP as there was Proto Balancer which was both an external link and commercial which has been tolerated. Users have the right to see alternatives. —Preceding unsigned comment added by 88.224.183.52 (talk) 09:38, 2 June 2008 (UTC)

Hmmm...

Doesn't Wikipedia use a reverse proxy system? - Ta bu shi da yu 08:11, 3 May 2007 (UTC)

Yep, they use Squid. See the 'Wikimedia Servers' page. RebelBodhi (talk) 20:27, 25 January 2011 (UTC)

In-bound and out-bound are relative terms

The article uses the terms "outbound" and "inbound" to describe the direction of web-traffic without stating whether this is from the client's or the server's point of view. This distinction is vital for understanding the difference between a proxy and a reverse proxy. This is confusing. Perhaps the terms "request" and "response" should be used, as that is not ambiguous. —Preceding unsigned comment added by 194.7.54.18 (talk) 10:21, 20 May 2010 (UTC)

Server network?

I'm deleting the first sentence of the article because it doesn't make sense. (Or it is just plain wrong, depending on one's viewpoint).

Sentence was, "... proxy server that is installed on a server network or on network equipment." There are networks. "Server" is not descriptive of the type of network -- a network doesn't have much use if it has no clients. Of course, there are also peer-to-peer networks, but there are not "peer" (lack of another term) networks. It this latter usage the two terms are mutually exclusive in that the "peer" necessarily is not the "nework" in question.

The term "server" refers to a process that manages a resource. A "web" server manages resources intended to be accessed from the web, usually using either port 80 or port 443. The SSH server manages access to the protocol that answers a request for connecting to it. This is usually port 22.

A proxy server manages access to a proxy which, in common usage, implies web access. In this case, it serves as a proxy and imitates a client. A reverse proxy imitates a server. For a reference, check out the Apache docs. Using Google, to find "define:reverse proxy" should result in a link to apache documents, for that server: http://rbt.phs.com.tw/manual/glossary.html. Here there i

Kernel.package (talk) 18:51, 18 December 2010 (UTC)

Uses explained

I moved the following (anonymous) comments from the top of this talk page to this section, as I believe these uses (spoon feeding and architecture shielding) are now sufficiently explained in the article itself. PPvG (talk) 21:58, 29 June 2011 (UTC)

There is another way that rev.proxy can help main web-server: it is not caching, but still lowering resource usage.

Let'a imagine servie like forum, LiveJournal, etc - where each user has his own pages with same URL like /UserProfile.phpspoon feeding Here there i Here there is no simple way (if any) to cache such pages, but still reverse proxy can be of much help.

Usually server has times faster connection, than clients do (especially dialup or mobile phones users). And server serves times more connections at same time, than any of its client does.

Apache is flexible. MOD_PHP, MOD_PERL, MOD_ASP, FastCGI, Java, MS.Net, etc, etc, etc So, when some dialup user connects to forum, Apache forkes one more thread/process wich can take several decades of megabytes of VM to run with all those plugins (and 1 or 2 Mb even without single plugin). Then this instance generates the page and... and waits. Usual modern page with scripts, images, SWFs can easily take 300-500 Kb. That is about 10 second on dialup. While Apache already did hid deed - generated page - it cannot quit or jump into free threads pool - until slow user will download the whole page. To add that many users opens a number of connection at time to download HTML, CSS, Images, etc simulateously.... That quickly makes VM requirements huge for WWW server without reverse proxying holding an umbrella over them ;)

Links: http://www.freesource.info/wiki/AntiApache http://babelfish.altavista.com/

open reverse proxy can be used for providing remote access services (see goMyPlace). Such proxy can implement data hash based cache and allow SSL encryption. In the same time the proxy can naturally solve problem of NAT and firewall penetration and prevents exposing of the IP address of the HTTP servers behind the proxy.

List/Comparison?

Should we convert the "list" section into its own page, perhaps a Comparison of reverse proxy software? It seems to me that there are enough differences to make a comparison page useful, sort of like the one for web servers.  ◉ ghoti 20:03, 28 May 2012 (UTC) Strongly agree, people are posting links to their favorites again. Better to have a comparison page or list than have this page cluttered. — Preceding unsigned comment added by 65.112.185.66 (talk) 19:26, 2 October 2017 (UTC)

HAProxy and Exceliance links

• There is no link behind HAProxy and I wanted to add one to the haproxy homepage.
• Aloha Load-balancer, edited by Exceliance is a load-balancer which acts as a reverse-proxy, since it relies on HAProxy. => How can we add a reference (and a link) on the server list? — Preceding unsigned comment added by 213.41.240.142 (talk) 09:47 2012-06-05 (UTC)

WAY too many links

This article contains way to many links. Wikipedia is not to be used as a directory or list of links (see WP:NOTDIRECTORY). Additionally, external links must be encyclopedic and improve the page (see WP:EL).

I've added a link to a search for "reverse proxy servers" at the Open Directory Project in preparation for removing the link farm. If anyone wants to comment, now is the time to do so. — UncleBubba ^{( T @ C )} 18:25, 5 June 2012 (UTC)

User Comment

I found this sentence to be an excellent example of word selection to convey meaning: "Contrary to a forward proxy, which is an intermediary for its associated clients to contact any server, a reverse proxy is an intermediary for its associated servers to be contacted by any client." As I read it, especially the last part, and as someone who has a pet peeve about such things, I immediately understood the term. Kudos to the author.Terp02 (talk) 14:23, 9 April 2017 (UTC)

Please add "API Gateway" as included concept

See https://stackoverflow.com/a/37265515/287948

apparent contradiction in first paragraph

The first sentence includes:

"a reverse proxy is a type of proxy server that retrieves resources on behalf of a client"

The fourth sentence includes:

"In other words, a proxy acts on behalf of the client(s), while a reverse proxy acts on behalf of the server(s)"

These 2 expressions seem to be in direct opposition to each other. If the use of "on behalf" is ambiguous, perhaps the term should be avoided or clarified with additional context.

Hansifer (talk) 18:38, 29 September 2020 (UTC)

I changed the later usage to 'is associated with'. You're right, 'on behalf' is confusing because everything is always happening in response to a client request, so it some ways everything is acting on behalf of a client, even the server. - MrOllie (talk) 19:20, 29 September 2020 (UTC)

Add text about reverse proxy usage in CDNs?

Reading this article, I feel like there should be some mention of how reverse proxy servers are heavily used within content delivery networks. There is mention here about reverse proxies providing caching, but there is no specific mention of usage by CDNs. I don't have cycles right now to come up with the appropriate text, but thought I would flag it here in case someone else has the time to do so before I get a chance to think about it more. A challenge may be to find neutral sources. A quick web search on "reverse proxy in a cdn" brings up a number of pages explaining how reverse proxies are used by CDNs, but almost all the initial pages I could see are written by CDN companies. More searching may be required to find more neutral sources. (Note: I also see that the companion CDN page - Content_delivery_network#Technology - could use some additional text mentioning the role of reverse proxy servers inside a CDN.) - Dyork (talk) 06:53, 18 November 2020 (UTC)