Problem: in the past few weeks, we have been getting an increasing level of spam. The ones we get most have some type of colour or image background. How can we get this under control?

The fix: you can handle this at either the client or network level – the network level might be easier if you have a lot of users with relatively little PC experience.

You can either purchase a turnkey solution from companies such as Barracuda or you can come up with your own solution using Linux and some open-source packages such as Postfix and SpamAssassin. Which way will be the best for you depends on how much time you have to devote to managing the system and how much you can handle on your own in terms of managing Linux and some of the associated administrative tasks.

If you go for the turnkey approach, look carefully at the hardware support. One reader I have known for years got burned recently when their spam system failed and was told by the vendor it only came with 90 days' coverage, since no additional hardware warranty was purchased.

Blocking spam will be a never-ending task. The type of spam you are talking about, image spam, is the latest entry on the scene and is proving to be one of the harder ones to block, because the spammers are using graphics files instead of text. This means that the 'signature' used to block this type of spam is rendered useless by making even a minor change to the background of the image, thus requiring a new signature in order to be able intercept it. If you go with the roll-your-own approach, there is a Optical Character Recognition add-on that can help you look inside the graphics file. There is no 100 percent solution to blocking all of the spam all of the time.

At some point you will hit a point of diminishing return to where you are getting more false positives where good email is getting falsely tagged as spam. You will have to decide where that point is and what level of spam getting through you can accept.