Making a web site cachable to cache engines

[schemery]

How do I ensure the content on my site is cachable? I have a customer that wants to mark all the content cachable so their Cisco cache engines will cache it for a performance boost.


-kev

[Spathiphyllum]

You can try with adding META-tags to header though they aren't always guaranteed to work:

<meta name="date" content="2005-05-21T14:03:00-05:00" />
<meta name="distribution" content="global" />
<meta http-equiv="expires" content="2592000" />
<meta http-equiv="Cache-Control" content="public, max-age=2592000, must-revalidate" />

These headers will let the document be cached anywhere and that it is considered expired/dated after 30 days. You would also want to set a document date in the header somewhere to give the requesting agent a time-zero reference. You might also consider setting the document date to be returned by the server. That would actually be more proper than manually dating the document.

See http://www.faqs.org/rfcs/rfc2068.html for the HTTP/1.1 rules. This will tell you more than you want to know. I'm not sure if you can specify caching to a particular engine... it might be interesting to see if this is done.

[Connie]

May be I don't fully understand what your trying to do but SEs cache pages by default. They also show the newest page rather than the cached page. Google is introducing a new service. For people who sign up and do the search through that service then they will get a pre cached page. The purpose is to make surfing faster. The program is in beta right now.

Don't have a link for more information handy right now, but I can post one later if your interested.

[Spathiphyllum]

clssam,

Best guess I can venture is that some proxy hardware/software device on his customer's network (Cisco tradename) wants a header to help it store data locally (intranetwork) on the caching device rather than bypassing the device every time a document request is made. I should think that the device could be configured to cache and not cache in innumerable ways; but maybe the device just needs simple caching header rules to store documents. By default and in the absence of any specific caching headers, I guess the device is set to get everything "fresh" on every request.

[schemery]

I think the cachabilty of content is usally controlled by the server. The CE is looking for a HTTP header that says it is ok to cache this object (jpg, html, etc). I think the idea of using META tags might work. I just need to figure out how to add in the code to my PHP site.

-kev

[Spathiphyllum]

...The CE is looking for a HTTP header that says it is ok to cache this object (jpg, html, etc)... I just need to figure out how to add in the code to my PHP site.

kev,

I kinda figured the default would be to bypass caching - it's the "safe" way to let those behind the cache retrieve everything new by default. As far as PHP code - sorry. You've tapped my knowledge and I'm too lazy to figure it out ; but there are several regulars who could guide you and I suspect they will in the not too distant future.

[Spathiphyllum]

I guess it would be helpful to provide something on using the server to include a caching tag:

See http://httpd.apache.org/docs/mod/core.html#fileetag

In .htaccess try this directive -> FileETag MTime Size

The server will respond with a header looking something like this:
ETag: "25dda0a-3967-4289b182"

So, the caching server (or browser) will have a unique ID to identify a retrieved resource. If the ID of that previously accessed resource remains unchanged, the caching server says "Hey, it's still the same document so I'll give you the locally cached version instead." Alternately, if the ETag changes, the server says "Looky here. The document has changed so I better look at it more closely. Hmm, what rules (or other headers) do I have to handle this "new" document?" At least that's what I think the server is saying.