Do AV suites slow down HDD performance?

[DeaconFrost]

There was a debate in another thread on a subject that I, and others, had assumed was dead and gone. With the speed and power or today's computers, I couldn't see a reason to still disable AV software during gameplay. Now, we aren't talking about scheduled scans, or full system scans, just AV software with real-time scanning. I've worked on hundreds of computers, and never once felt a performance hit by installing AV software.

I'll also admit, that not all AV software is created equal, as some can be overblown bloatware-laden crap. However, there are plenty of good alernatives, even some free, that won't yield a performance hit. I know some games have recommended disabling these software applications during gameplay, but that's long out-dated. Video editing software doesn't recommend this, and that would hit the disk much harder than a game.

Anywway, I set up a very quick test using my AV software of choice, SAV 10, on a fresh install of Server 2008 Standard x86, and I'll post the results below from HDTune. The computer is by no means high end, so I felt that actual would help my point that you aren't really causing any performance loss. It is an HP dx2000 computer with a Pentium 4 2.8 GHz processor, 1 GB of DDR400 memory, and a 160 GB Western Digital IDE drive, running at 5400 rpm. Aside from Server 2008, the only driver/software I installed was the Intel Chipset drivers. This was done before either test was run. I ran the test once, installed SAV 10, and ran the test again. Here are the results:

First run, with only Server 2008 and the Intel chipset drivers installed:

Second run, with Server 2008, Intel chipset drivers and SAV 10 installed:

The only real difference is that the minumum speeds dropped a bit. The important factors, in my opinion stayed roughly the same, such as average transfer rate, burst rate, and cpu usage. Oddly enough, the access time was even lower in the second test with SAV10 installed. The bottom line is, I see now performance loss by running AV software. I never have, to be honest, and I'm still convinced that this line of thinking is long out-dated. This computer is not a high end system, which I would easily assume would show no difference either, given the faster drivers, multi-core CPUs, etc.

 

[PTNL]

I think stepping back a little more on this topic would help put things in perspective.

Having *any* application running constantly is going to use some amount of system resources. How much used obviously varies on the software, hardware configuration, drivers, etc. But there will always be a certain amount of resources consumed when compared with a vanilla setup without said applications.

Calling it a "performance hit" is a bit of a misnomer, as it implicitly suggests something more significant of a loss than what you've reported. And to play the "devil's advocate" here, I'd look at your HD Tune screenshots and wonder why there is concern about a negligible difference in Access Time and Burst Rate? Sure there is a larger difference in Transfer Rate. And I've not seen what the purpose of this server is. Regardless of the server's role, that fact remains that desktop-grade hard drives are running in a server, so this would be the first area to address if the HD Tune reports are a concern for you.

The point is that I feel there are things not being considered to provide a deeper look into this. To really dive into this problem, you'd need to know more stuff like: SAV10 configurations, how this machine is being utilized, number of concurrent users and how users interact (whether on just the file system level, etc.), plus many other considerations. And most importantly, the impact that a serious virus could do if it got ahold of this machine. But I also feel that this is not the purpose of your thread, but rather the high-level view of having any AV package on the machine.

 

[DeaconFrost]

This server isn't really functioning as a server. No one is connecting to it, and I didn't even have the NIC drivers loaded at the time. It ws a spare, old desktop that I planned on testing Server 2008 on it, so I figured I'd use it for the test because it was clean. I'd like to rerun these tests on the same system with XP, using several other AV packages, while ghosting down a clean image between each test.