17 Jul 2007 @ 9:12 PM 
 

Anti-Spam technologies measured and compared

 

Peter Brockmann today published a great report called The Spam Index Report (registration required but free). He has actually created a metric called the Spam Index which is really cool because it is easy to calculate and is user oriented. It is the sum of:

  • Monthly (estimated) count of Spam messages dealt with
  • Monthly (estimated) count of minutes spent dealing with Spam
  • Monthly count of messages that need to be resent due to anti-Spam tools
  • Monthly count of good messages marked as Spam (false positives).

In the report, Brockmann compared the performance of anti spam tools (filters, blacklists, and challenge-response) and their subcategories. The metric was really the Spam Index of the users after the implementation of each tool. Challenge-response was the most accurate and consistent tool.

“Challenge-response services allow known email senders to pass messages without interruption. They get these typically from users’ address books, or the outbound email stream or from the customer relationship management system. Messages from first-time senders trigger a ‘challenge’ message asking the sender to respond by clicking a link, replying to the message or visiting a website to release the message they’d sent to its final destination”.

This made me think about my own situation. I use gmail as a mail client, and gmail seems to use part of filtering and part of challenge-response (it approves people in my contacts). I estimate that my Spam Index is well under 100 with gmail’s tools.

  • Share/Save/Bookmark
Tags Categories: Security Posted By: jmiles
Last Edit: 17 Jul 2007 @ 09 13 PM

EmailPermalink
 

Responses to this post » (2 Total)

 
  1. pjbrockmann says:

    Hi there. Thanks for the post.

    I wanted to let you know that I have also made the Spam Index Calculator readily available on http://www.brockmann.com too. It uses Javascript so all the details stay in the browser and don’t traverse the network.

    Url is Calculator.

  2. richij says:

    The comment below has been edited with phrases and links removed
    to keep the focus on the design and operational issues of the technology
    and solution (not on the people).

    I found these links on Richi’s site to be very useful for my knowledge and
    understanding, and in supporting his statement:
    - http://richi.co.uk/blog/2006/12/another-challengeresponse-datapoint.html
    - http://richi.co.uk/blog/2006/12/boxbe-another-cr-spamhaus.html (This
    was of extra interest to me since I have a Boxbe address I use to
    experience this technology myself).

    And also this link to Justin Mason’s site, http://taint.org/2007/07/19/122638a.html,
    which I think shows a valid consideration about the Spam Index referred
    to in my post.

    I think it’s interesting to reflect on my own situation, with a very low spam
    index using gmail’s spam tool and not a challenge-response system (I
    only get about 1 message per week in Boxbe).

    ====================
    the edited comment follows:
    ====================
    A very flawed study: challenge/response is a dreadful idea

    Justin Mason and I have debunked this study [...].

    More at [...]

Post a Comment

You must be logged in to post a comment.

\/ More Options ...
Change Theme...
  • Users » 4
  • Posts/Pages » 191
  • Comments » 5
Change Theme...
  • VoidVoid
  • LifeLife
  • EarthEarth
  • WindWind « Default
  • WaterWater
  • FireFire
  • LightLight

About Jim Miles



    No Child Pages.

Contact



    No Child Pages.

Case Studies



    No Child Pages.

Prioritized Approach to PCI



    No Child Pages.