Anti-spam smackdown finds best junk filter McAfee has claimed the crown in a run-off of anti-spam products organised by Virus Bulletin, the independent security certification body. A total of 12 products took part in the anti-spam comparative review, which rated the effectiveness of the products against each other rather than against a fixed standard. The latest …
"means that its spam catch rate came out as twice as effective as the average in the test, and its false positive rate is twice as low as the average in the test"
So if enough rubbish filters are tested, some of the also-rans could make gold and some golds make platinum. At the moment all we can say is which one is the best, not if it is good enough.
How about an absolute standard for quality, such as 98 out of 100 spams caught. False positives no more than 5%.
When are we going to see a genuine industry move to eliminate the spam scourge, which has turned a promising new communications media into a daily tirade of shite? Come on, there have been ideas batting around for years now, I can't see why no concrete plans have never been put in place!
I note with some amusement that the "achievement" is to be twice as good as average, so if you increase the number of companies participating in this test, taking particular care to include some not very good ones, you'll get *even better* results from the winner without actually improving its performance at all.
That's not platinum. That's "tin foil", as in "essentially fake". The intellectual bankruptcy of the AV industry gives even Microsoft's "innovations" a run for its money.
The Mcafee spam / web filter appliances; when they work are wonderful; though we have one we don't use anymore. Support from McAfee in our experience is atrocious for all products. The stock response is upgrade to the latest micro patch or full version; which is a massive risk on an enterprise email system. On one occasion the upgrade hadn't even been tested properly on our appliance and we were without the filter for over a week whilst they desperately fixed the issues - not wanting us to downgrade as they were basically developing on our production box.
We have switched to Postini which is marginally cheaper although a considerably worse spam filter system but its reliable.
Shame they didn't include dSpam - it would have beaten the rest hands down. No doubt at all. But it doesn't run on Mickey Mouse OS's like MS Windows so maybe that's why it wasn't included, which seems to be a common theme in the test.
As for "ClamAV failed to catch better than 86 per cent of junk mail" - well, yes, it would wouldn't it? It's an anti-virus filter, and NOT an anti-spam filter!!!
Are any of these products based on or use Spam Assassin? Ever since we started using that beauty our spam problem has gone right down - though now we augment it with other techniques, I would be interested to see how it measures up.
And if they have not tested it directly or as part of one of these products, I don't think it says much for their methods.
The problem is those pesky 'false positives'. Even if you put the onus on the ISP to do the filtering then as soon as some critical email from Auntie Flo about how her from Eastenders is now advertising nappies in Guatamala fails to make it through, up goes the cry of "damned ISPs are blocking my email".
Anyone in the industry knows that 100% is a pipedream but when has that ever stopped the media from jumping on a wobbly bandwagon?
...So how do these commercial products perform in relation to open source solutions such as ClamAV, SpamAssassin, Razor, etc ...?
Personally I don't run desktop anti-spam software - my combination of server-side Mail Toaster running the above combo and Apple Mail's spam detection automatically filters out about 98% of spam with no intervention at all. I haven't seen a false positive for months.
Completely agree about dodgy criteria. Its all marketing bullshit. "Don't believe the hype"
The false positive rate is important, but also the number of emails tested. A small company, let's say with 25 staff, can live with that and manage it cos it's only one email now and then.
But when you have 5 million+ emails a day, a tiny false positive rate can cause huge support problems, as well as "it doesn't work properly" type comments.
I agree with Kevin Johnston that 100% is a pipedream, so I live with that.
And along with Martin Burns I would like to see dspam, spamassassin, etc tested.
Rich 2: well, yes, it would wouldn't it? It's an anti-virus filter, and NOT an anti-spam filter!!!
Correct.... but you can add signatures to ClamAV, using the ClamAV engine to actually detect spam....
In short...
Clam AntiVirus is a GPL anti-virus toolkit for UNIX and was *coded to detect email viruses*, it's scanning engine is very flexible and has been used to provide *add-on* signatures.
Sanesecurity add-on signatures provide enhanced email security against the following email types:
Phishing, Spear phishing, Fake lottery, Ecard malware, Casino, Fake jobs, Fake loans, 419s, Fake Diplomas, porn, emailed malware and other general spam.
http://sanesecurity.co.uk/index.htm
Cheers,
Steve
Sanesecurity
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
