Virus definitions numbers

[IBK]

Just an overview of the number of virus records of some AV products...

[RejZoR]

But we all know that these numbers don't mean anything...
And the big bump in August is for your comparatives test (because it's the On-Demand test...)

[IBK]

Yeah, that is what I wanted to show indirectly. You see a big bump in the numbers, but in the results of the august test you will not notice any big difference which could explain those numbers. (=virus records says absolutly nothing = just marketing).

[RejZoR]

Well in general i kinda belive those published by Kaspersky Lab.
They are not way too high and we all know how good detection KAV has...
But again generic signatures cannot be counted (Backdoor.Rbot.gen are quiet common)

[hbkh]

Originally Posted by IBK

Just an overview of the number of virus records of some AV products...

No Avast!?

[FastGame]

Originally Posted by hbkh

No Avast!?

I think it was left out because in previous test the signature count is N/A

[Blackcat]

Just an overview of the number of virus records of some AV products

NOD's is still top secret?

[vlk]

Some vendors (like Eset/Nod and Alwil/Avast) don't publish these numbers as they are useless anyway.

Give me an hour and I'll create a super duper virus scanner with say 10,000,000 signatures (or whatever number you pick). It won't detect any malware - but that's a different story


Cheers

[Firefighter]

Originally Posted by IBK

Just an overview of the number of virus records of some AV products...

An other interesting point of view from this IBK table above. These numbers are still only some estimates of my own from that same picture.

The growth of signatures between the period 2-2004 to 8-2005:

183 % -- BitDefender
154 % -- AntiVir
_85 % -- F-Prot
_75 % -- DrWeb
_67 % -- Kaspersky
_63 % -- McAfee
_43 % -- M$ RAV
_24 % -- Sophos
_8.5% -- Symantec

My private conclusion. Symantec and Sophos have outstanding heuristics, generic detection or what ever like these, when they are still in the game with so minimal growth of new signatures!

Best regards,
Firefighter!

[Brian N]

Originally Posted by Blackcat

NOD's is still top secret?

I had PHP count a file, and it came up with 48959. Maybe that's the max what it can count, or you gotta give or take a few K.

[IBK]

Originally Posted by Brian N

I had PHP count a file, and it came up with 48959. Maybe that's the max what it can count, or you gotta give or take a few K.

number of virus records and number of virus names in database is not the same. If you count the number of virus names in KAV database you see that they are about the half of the number of virus records.
p.s.: yes, your php did not count everything..., the number would be much higher...

[Brian N]

Ah well, worth a try

[fosius]

Three or four years ago ESET told me that they had something about 50000 signatures... This number must be a lot bigger now...

[Stefan Kurtzhals]

I don't think Symantec is giving the real number of signatures they actually have. IBK, how about also giving stats on the size of the virus definition files?

[Technodrome]

Def files are much larger now so they must be adding something. The way they count malware is an entirely different matter and possibly has nothing to do with advertising. What av product (with high virus count) is very popular among the end users or enterprises?