I would very much welcome the addition of true HTTP content filtering in Astaro.
As good as the Cobion filtering is, it is limited by the facts that:
1) The decisions of what is worthy of being filtered are made by someone (or machine?) other than me.
2) It misses a LOT because it does not (and cannot) evaluate every site out there before any of our users get to them.
3) I believe that Cobion provides URL filtering rather than content filtering.
URL filtering is definitely good but it is incomplete.
Even though Cobion offers the ability to make recommendations to the filtering list, how can anybody know that something new needs to be blocked until after they have already been exposed to the problem? Further, my users aren't going to tell me...
I wonder if it would be much of a stretch to employ the same regular expression pattern matching that the email filtering uses, to HTTP content.
Yes, there would be some overhead, but filtering should be offered with an on/off switch.
Further, it might not be difficult to offer an option to limit the scan to the first X number of characters in the stream. I expect that whatever the purpose of a web site is, the probability of a recognizable pattern showing up in the first 1,000 characters is quite high.
This could also provide a benefit to Cobion.
There are plenty of places with objectionable material that can't be pattern matched, so a URL-based filter will still be very important. Plus, I don't know if a Cobion check would be faster than a local pattern match but it would require fewer resources from the firewall. If an Astaro user chooses to accept the overhead of the content filter, they could be offered a checkbox that would allow them to participate in a background submission to Cobion's blacklist, perhaps using a behind-the-scenes set of patterns that Cobion and Astaro agree upon.
In this way:
1) Cobion would greatly enhance their database
2) Astaro would benefit by:
. . a) Providing an improved product
. . b) Possibly gaining a small revenue stream from Cobion
3) The Astaro user would benefit by:
. . a) Having the benefit of both types of filters
. . b) Improving the accuracy of future Cobion checks
. . c) Reducing the overhead of his next filter by finding it at Cobion next time. (Even if there was no speed increase, the load on the firewall could be reduced)
It is easy enough to think of many more ways to enhance this into a big development project, but I believe that just plain pattern matching would be a huge step forward.
Thank you,
Rich
This thread was automatically locked due to age.
