Incredible traffic saving with the header "Cache-control: public"???

[GONZO]

Hi, first thanks for your attention and sorry for my English
translation.

I have a question that intrigues me a few weeks. This is the header
"cache-control" in particular the behavior of the options "private"
and "public" in Google App Engine.

First of all, this is only to serve static files (css, js, etc)

Let's go. With "cache-control: private" experiment curve normal
traffic. But with "cache-control: public" experiment Traffic
incredible savings. In both cases, everything seems to work well.

The question is:

1. How can traffic be a big savings? Is reduced to 15%.

2. Saving you a real traffic? Or is something special instead of
Google App Engine?

Better look at this diagram illustrates: http://gonzo.teoriza.com/almacen/cache-control.jpg

Thanks in advance, I hope to be clarified.

[Erik Wilson]

Hi Gonzo,

If you have a lot of authenticated users the difference you are seeing
is probably in the cached authenticated requests.

From http://www.mnot.net/cache_docs/:
> My pages are password-protected; how do proxy caches deal with them?

> By default, pages protected with HTTP authentication are considered private; they will not be kept by shared caches. However, you can make authenticated pages public with a Cache-Control: public header; HTTP 1.1-compliant caches will then allow them to be cached.

> If you’d like such pages to be cacheable, but still authenticated for every user, combine the Cache-Control: public and no-cache headers. This tells the cache that it must submit the new client’s authentication information to the origin server before releasing the representation from the cache. This would look like:
> Cache-Control: public, no-cache

> Whether or not this is done, it’s best to minimize use of authentication; for example, if your images are not sensitive, put them in a separate directory and configure your server not to force authentication for it. That way, those images will be naturally cacheable.

[Adam Sah]

yes, incredible savings. Need to include a version number or hash tag
in the URL
to ensure that new versions get pushed in sync. An example:
http://bbfdirect.com/

I'm working on the next level, which is enable cache-control for the
HTML itself,
i.e. render the page and then make AJAX calls to fetch the dynamic
content,
and detect out-of-date HTML using the AJAX calls. My lead eng
thinks I'm
psycho and I probably am. We're starting with cache-control:
private on that
stuff, but if I can resolve the issues with proxies serving cached
data during
javascript:window.reload(true), then we'll go public FTW.

wish me luck (we went live with the private version yesterday),
adam

On Dec 7, 10:07 am, Erik <erik.e.wil...@gmail.com> wrote:
> Hi Gonzo,
>
> If you have a lot of authenticated users the difference you are seeing
> is probably in the cached authenticated requests.
>

> Fromhttp://www.mnot.net/cache_docs/:

[Jason Collins]

"Cache-Control: private" only uses the end user cache to cache
resources.

"Cache-Control: public" uses any downstream cache to cache resources
(including the browser cache).

Google has a downstream cache in front of App Engine requests, so if
you serve your resources with "Cache-Control: public", Google will
cache that resource. Subsequent hits are served from there, and I'm
pretty sure they won't even show in your request log at all.

j

[GONZO]

Hi, thanks for your response.

I like your answer. Just what I expected. Confirms my theory and my
experiences.

I confirm that requests saved (90%) do not appear in the logs and not
counted in the consumption of bandwidth. It's free total. Amazing.

I appreciate all the information you know about it.

Thank you all. Greetings!

[Albert]

Hi!

This guy seems to implement something similar to yours, but he says,
"...each time a page is served cached, you'll see a 204 logged in your
Appengine dashboard."

http://www.kyle-jensen.com/proxy-caching-on-google-appengine

Did any of you guys notice that too?


Albert

[GONZO]

Quizás quisiste decir: A mi no me aparece esto. Puede garantizar que
la diferencia entre cache-control private y public es que con public
Escribe texto o la dirección de un sitio web o traduce un documento.
Cancelar
Escuchar
traducción del español al inglés
I do not look like this.

I can guarantee that the cache-control difference between private and
public is that public away 95% of traffic, almost all. However, this
is served 100% ok to customers. The traffic is not on any chart, or if
the consumption of bandwidth or in the logs.