happygeek 2,411 Most Valuable Poster Team Colleague Featured Poster

Microsoft's Cybersecurity and Systems Management Research Group have created an automated tool, as part of the larger Strider Search Defender project, to combat sources of comment spam that is the scourge of blogs across the web. Because sites can get high legitimate search rankings while at the same time serving up spammed ads, it is a problem that has to be dealt with. Microsoft is on course to be the unlikely hero of the hour.

SpamHunter does this by creating a list of doorway sites, hosted on legitimate blog or forum sites and feeding ads from a central spammer target page. Rather than adopting the usual content reading approach to spam discovery, Microsoft is thinking contextual analysis of URL redirection instead. By crawling the web using search engine queries to locate sites within the same network, SpamHunter can pass the information across to the Microsoft Strider URL Tracer which then puts the pieces together and determines where the central domains fed by those doorways are. Because networks of thousands of doorway pages can serve ads from a single domain, it is possible for Search Defender to take down an entire operation in one hit. Indeed, the system has already had some measure of success during testing, determining that 97 percent of the 5,500 spam sites at Blog4Ever were the work of a single comment spammer who was using the same AdSense affiliate identifier for example.

The real clever part is that the more comments there are linking back to a spam site then the quicker SpamHunter will find them, and what is more it effectively becomes a honey forum enabling other spam URLs to be easily obtained. Of course there will always be the problem of false positives, and to try and reduce these Microsoft is making use of the whitelist of legit advertising and web analytics sites that it has compiled during work on the malicious exploit finder project.

Of course, this is not so much a cure as just another weapon to be used in the fight against spam. It is up to us, bloggers and forum moderators, to do our bit in keeping our eyes open for comment spammers and sending them to the dev.null hell where they belong. Search engines could do a little more when it comes to blacklisting sites that host comment spam, and MSN is promising just that. Working with the Search Defender team it will pursue leads and either remove proven spam sites or assign them such a low relevance rating that they it amounts to much the same thing. I would like to think that the other big search players will co-operate, but given the current rivalry between Microsoft and Google this seems unlikely.

Dani AI

Generated

As points out, network-level detection and takedowns are only one part of the picture. Practical, site-level defenses stop most comment-spam campaigns before they create doorway pages or poison search results. The notes below focus on low-risk, effective controls that remain useful years after a single tool or vendor changes.

Immediate, high-impact controls:

  • Hold any comment that contains one or more URLs for manual review, or require approval for accounts younger than a short probation period (for example, 3–7 days).
  • Strip or convert links from untrusted comments to non-clickable text and add rel="nofollow" so comment links do not pass SEO value.
  • Add a hidden "honeypot" field to comment forms; spambots that fill it can be discarded automatically.
  • Use lightweight rate limits per IP/account and block obvious repeated patterns (same destination URL posted many times in short windows).

A simple moderation rule (pseudocode) that fits many platforms:

if (comment.contains('http') && account.age_days < 7) {
  hold_for_moderation();
}

Longer-term measures: maintain a small whitelist of trusted analytics and ad domains to avoid false positives; create a "trusted commenter" role that bypasses link checks after manual verification; record and aggregate comment destinations so spikes to a single domain are visible and can be blacklisted quickly. Keep a short blacklist of affiliate/tracking IDs and block comments that include them.

Cautions and workflow tips: test filters on a staging site and monitor false positives for at least a week before fully enforcing; prefer soft measures (hold/strip links) over blunt bans to preserve legitimate discussion. Log events so recurring offenders (IP, user, domain, user-agent) can be automated into escalation. Together these steps make a site resilient to mass comment campaigns while minimizing harm to genuine visitors.

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.