The purpose of web filtering in a complicated confidence architecture

In a many elementary and simplest form, filtering entrance to calm on a web can be achieved by rather blunt instruments such as DNS black-holes. And, in a early 2000s, this was some-more than enough.

Over time, though, a web became some-more renouned and useful, both for users and attackers, and web filtering providers indispensable to adult their diversion again: elementary URL filtering was used to concede entrance to example.com/news and retard entrance to example.com/badstuff. Again, for a time, that was enough.

“But web filtering is an arms race, and a expansion of a unknown substitute was in full swing. The customarily remaining apparatus in a web filter provider’s arsenal was full energetic calm analysis, that can be achieved in mixed graphic and (sometimes) hostile ways,” recounts Craig Fearnsides, Operations Technical Authority during Smoothwall, a UK-based developer of firewall and web calm filtering software.

“One routine (which pristine firewall vendors employ) is Layer 7 signature analysis, looking for patterns on a handle and restraint packets. URL and domain settlement matching is subsequent – permitting news.*.com though restraint ads.*.com. Finally, there are regular expression-based methods that concede for calm to be scored and categorised according to a customer’s requirements. This involves comparing positive/negative scores to phrases, as good as some-more nuanced categorisation, e.g. essex vs sex vs sextuple.”

No web filtering is no longer an option

“Web filtering has had a rather stubborn story and has been vastly misunderstood for many years,” Fearnsides tells me.

But, thankfully, a IT/network confidence village now generally understands and accepts that entrance to a web is an essential partial of many of their associate employees’ operative day, and that something needs to be in place that will keep a corporate sourroundings protected though preventing people from doing their jobs.


“This is where pure (filtering though requiring pithy substitute settings) and normal web filtering can assistance massively, giving already bustling IT/network administrators a collection to keep a business moving, though bogging them down in low-level doing details,” he noted.

Real-time web filtering challenges

The biggest plea for a web filtering businessman is always going to be speed, followed closely by comprehension, he says.

“There are many shortcuts that can be used to boost throughput of energetic filtering solutions (this is customarily along a lines of tying escalation of computational effort), though they mostly lead to bad categorization or fake positives. The customarily genuine approach to urge speed of throughput is by optimising any of a graphic layers of categorisation over many iterations,” he forked out.

“One routine involves regulating appurtenance training to know what calm has been formerly identified as one or mixed categories (based on a subset of existent elementary lists). We afterwards indicate a apparatus during a store of requests/responses to have it find out pointed differences that would be scarcely unfit for a singular person. This appurtenance training routine produces vastly softened patterns for use by a reticent unchanging countenance engine, augmenting throughput and efficacy during a same time.”

He says that fake positives are inevitable, though can be mitigated by permitting business to prioritize one or mixed categorization formula over another.

For example, a resolution can be done to retard adult calm BEFORE permitting audio/video content. This would have a suitable outcome of permitting entrance to audio and video content, while still preventing entrance to adult calm sites that might be (appropriately) categorised as audio/video calm sites.

Back to Top