[squid-users] content adaptation using squid

Alex Rousskov rousskov at measurement-factory.com
Thu Sep 6 16:46:10 UTC 2018


On 09/06/2018 12:06 AM, Yosef Meltser wrote:

> Now we would like to make a content adaptation, for example to show an alert
> every time the user entered a website.

If you have not already, please familiarize yourself with the content
adaptation problems. The following FAQ entry is a good starting point:
https://answers.launchpad.net/ecap/+faq/1793

Splash pages (mentioned by Amos) can be a form of content adaptation and
several problems in the above FAQ entry apply to them as well.


> For example, in Icap there are some server frameworks that we can use (like
> c-icap, ICAP-server and etc and etc), which one is the most recommended?

There is no general winner AFAICT. All servers have their pros and cons
(cost, ease of management, performance, ease of modification, quality of
support, licensing restrictions, etc.). Use whatever works best for your
needs. For example, the server you use for prototyping may differ from
the server you use as a long-term solution.


> In contrast to the above, ECAP is not using any server, and the whole
> process in embedded into the squid, so it sounds quite easier. Is it?

Overall, there is no drastic complexity difference between

* embedding your code into an ECAP-capable host like Squid and
* embedding your code into an ICAP server like c-icap.

In some cases, even the APIs are going to be similar! Some ICAP servers
are probably easier to work with than eCAP, and some are probably more
cumbersome, but you are doing the same work on the conceptual level --
writing a "plugin" where your adaptation code/logic lives.

An ICAP server introduces an additional stand-alone service/software
that you have to take care of, of course.

Needless to say, writing an eCAP adapter is much easier than writing a
(good) ICAP server -- writing an ICAP server from scratch is usually the
wrong answer.


HTH,

Alex.


More information about the squid-users mailing list