Image spam continues to be a serious problem for service providers, enterprises and others that process e-mail.

According to Sophos, for example, image spam increased dramatically during 2006, from 18.5% of all spam in January 2006 to 35.1% in December 2006. Image spam is particularly difficult to address with conventional spam filtering technologies, since antispam technologies that are focused on blocking text-based spam are simply not adequate for fighting image spam that uses a variety of techniques to avoid detection.

However, as bad as image spam is, that’s not the worst of it. The proliferation of botnets means that spam is extremely difficult to stop because it is coming from such an enormous number of sources, not like in the ‘good old days’ when spam came from just a relatively few sources.

The growth of unified communications will likely exacerbate the problem with new types of spam and other exploits. For example, when voicemail received in an e-mail inbox becomes more common, I expect that we will see an explosion in both voice spam and voice messages designed simply to clog mailboxes.

While today’s spam is comparatively easy to detect, voice spam will be very difficult to detect. It will also pose enormously greater problems in terms of the bandwidth and storage that this spam consumes. The result will be that perimeter-based security, reputation analysis and related systems will be much more important as a means blocking this content before it reaches corporate servers.

Osterman Research will be hosting a Webinar, sponsored by Tumbleweed, to address image spam and other security threats. While we’ll be focusing on current threats and what you can do to stop them, we will also look to the future to see what spammers and other attackers have in store for us. You can sign up for this Webinar here.

Michael Osterman is principal analyst of Osterman Research.