The latest version of Brightmail's Anti-Spam software offers effective protection from the latest spamming techniques

Version 5.5 of Brightmail's Anti-Spam software, released in December, has improvements to better handle spam originating from open proxies, and has improved URL-based filters.

New in the release are personal blacklists and whitelists for users running Microsoft Outlook, and new buttons for people using Lotus Domino client software to enable them to report incorrectly blocked messages and unfiltered spam.

The software is available for Windows, Red Hat Linux and Solaris 2.8/2.9 servers, and is also offered by third-parties as a service, and by others in appliances. In each case, the spam definitions are updated every 10 minutes from Brightmail's datacentres. We reviewed the Windows version using a twin 400MHz Xeon processor system, on which we recorded an average 30 percent processor utilisation while the suite was idle. Brightmail has priced its products by the number of users, so the number or type of servers used does not affect the acquisition price.

For our tests we configured Anti-Spam 5.5 to work alongside the basic SMTP mail server built into the Windows 2000 Server operating system. The suite can also be used with any SMTP message transfer agent (MTA), and a special version is available for use with Microsoft Exchange.

Brightmail said its software generates a very low percentage of false positives - one in one million messages incorrectly identified as spam. We did not test this claim, but the downside of a low false-positive rate is that such products are likely to allow more junk email to pass through to the users' mailboxes. With this in mind, the web-based Quarantine tool should provide administrators with some relief because it allows each end-user to manage their own spam controls.

Brightmail Anti-Spam 5.5 differs from many spam filters because the suite does not include an MTA. While this will suit some firms, others may be better off with a system that combines mail server and spam filter in one package.

During setup, Brightmail installs the MySQL database and Tomcat web server, which are both required by the Quarantine module. The management interface uses Microsoft Management Console (MMC), so it will be familiar to many Windows system administrators. Built-in wizards made it simple to configure Brightmail options such as the global whitelist and blacklist.

Integrating Brightmail with the Windows SMTP service was also extremely easy. Brightmail monitors standard Windows SMTP and post-categorisation event sinks for incoming messages and, once it has processed them, it hands the messages to the Windows SMTP service, which then delivers them. Before putting our Brightmail system online, we only needed to specify the domain names for which our system was handling mail.

We had a little more trouble with the Quarantine module. This module stores suspected spam messages in a special database, and users can configure Brightmail so they periodically receive a summary of these messages. Although it is installed automatically, the Quarantine web site does not use the standard HTTP port, so some users might need a little training in how to specify port numbers in a URL. Also, the Quarantine web pages are not compatible with Microsoft's Internet Explorer 5 browser, so many firms would have to upgrade most of their desktop PCs before they could use this feature. The Mozilla browser seemed to work well with the module.

At present, Brightmail's technical support is handled from the company's San Francisco head-office, and is available only during the US working day.

Price: :$7 (£3.80 + VAT) per user for 5,000 users

Contact:Brightmail 07900 212 304