Top Security News

For the past several years, I have been archiving Bugtraq emails and placing them on LinuxRocket, mainly for my own use. It was a good resource, but needed more functionality, especially with regard to grouping related information. Recently, I decided to build a custom tf-idf engine, using LinuxRocket's Bugtraq archive as the corpus. Preliminary results were promising, so LinuxRocket now leverages that index along with a new engine that creates mutual document associations for those emails. The functionality was then extended to also tie in relevant documents from the AP Security feed.

On the front page, you now see the results of this mutual document association engine. Below each AP security news article are the top 10 most-relevant Bugtraq emails. Upon clicking on one, you are taken to a page where you can read the entire message. At the top of that page you will see a group of related terms. These terms are ones that have been identified as the most relevant to that page and link to other such pages within LinuxRocket. Clicking on one of those links you are then taken to a Term Landing Page that lists the top 30 emails, each with a snippet that includes the relevant terms.

This engine is a custom application written in C++. It uses libexpat, htmltidy, libcurl, and the snowball stemmer. The index generation is completely written from scratch. It runs as a background process and provides regularly updated results in the form of a feed for the frontend webserver to retrieve and display.

People have used LinuxRocket to monitor vulnerabilities, malware reports and software security. Often, emails within Bugtraq suggest security measures that are needed to thwart these issues. LinuxRocket is not a security service, but an information portal for the security minded.