IT Services



Web Proxy Caching FAQ


Contents



1. General Information



1.1. What is a Web Proxy Caching Server?

A proxy caching server is a system which sits between your web browser and the Internet. It acts as an intermediary (or "proxy") by fetching pages on your behalf, which it forwards to your browser. It also saves a copy of each document it retrieves to form a collection (or "cache"). The next time that a cached document is requested, the server returns it from the cache instead of going off to the internet.



1.2. What is Squid?

Squid is an open-source software package that offers high-performance proxy caching for web clients. At Oxford we are currently using version 2.7. See the Squid FAQ for more details.



1.3. My browser stores documents in its own cache on my PC. Why do I need another one?

The documents held in your PC's cache are only accessible to browsers running on your PC. The Oxford cache is accessible to all Oxford machines, but not to anybody outside the ox.ac.uk domain.



1.4. Can all web traffic be cached?

No. Examples of non-cacheable objects are pages which require authentication, responses to search requests, and dynamically-created pages. However, the vast majority of web pages can be cached.



1.5. How old is the data in the cache?

The time that objects are retained in the cache varies according to load. However, once an object has been in the cache for three days the server regards it as "stale" and will go and check the originating server if it is requested. The cache uses information in the headers returned by a remote webserver to determine how long it can retain an object without revalidating it with the server: see this flowchart for more information. You can force the server to re-fetch content with the "Reload" or "Refresh" button on your browser (or possibly shift+reload/ctrl+refresh).

The "cacheability" of a given page may be checked using the very useful tool at http://www.ircache.net/cgi-bin/cacheability.py.



1.6. What does the cache do with cookies?

The presence of a "Cookies" header in a requests does not affect whether or not an HTTP reply can be cached. Similarly, the presence of a "Set-Cookie" header in a reply does not affect whether the reply can be cached.

Squid does cache replies with "Set-Cookie" headers, but it filters out the "Set-Cookie" header itself for cache hits.



1.7. I'm being denied access to the proxy cache machines

Only hosts registered in the .ox.ac.uk domain are authorised to use the proxy cache.

If your machine is not within the Oxford network, please do not attempt to use the OUCS proxy cache, as it will not work. Refer to your local IT support officer or your Internet Service Provider for information regarding proxies.

If your machine is within the Oxford network, please first check that a local configuration problem is not to blame, and if not, then contact us on cache@oucs.ox.ac.uk with details of the problem and we will attempt to resolve it as soon as possible.



1.8. When I try to access a site, I get the error message "ERROR: The requested URL is invalid"

There are several possible reasons for this, some of which are given in the error page returned. Essentially, it means that the proxying software is unable to process the request submitted. If you are unable to solve the problem, please contact us with as much detail as possible and we will look into the problem for you.



1.8.1. Whitespace in URIs

Such URLs are technically invalid (according to RFC2396). However the caching software implements a workaround for such URLs: whitespace characters are "escaped" through conversion of the ASCII character code to hexadecimal. For instance the invalid URL http://www.ox.ac.uk/invalid url.html may be called instead with http://www.ox.ac.uk/invalid%20url.html. This conversion in itself may be considered a breach of standards, but does at least permit such URLs to be accessed.