 Hello, FreeNAS fans and people who want this question answered like I did. So I did the video a few days ago, if you didn't know, I'll link it in the description below and said encrypt all the things, how to set up, import, export encrypted volumes inside of FreeNAS. Now, this is my FreeNAS machine that's kind of like our production machine that we run everything on, and I figured I'd set up some testing because a lot of people had asked me in a few direct messages and a very valid question because I had commented that the speed at which the encryption operates is not real detrimental to the system as long as you had a processor that supports AESNI. So I tried playing with this in a virtualized environment, and I don't think I got any results that made any sense to me, so I decided, okay, I'm going to do a video and try this on my system directly. Now, giving an idea, I have a 16 gigs of RAM, I have an Intel i5-4570, and the i5-4570 was released in Q2 of 2013, so it's an older 4th gen Intel processor. We'll scroll down here real quick. Get my face out of the way. So we have an Intel AES instruction set is yes on this. So older processor, 2013, 2018, three days away, so it's not new. You can find this for cheap. I've had the system for a while, I didn't even pay for the processor. It was out of an old machine that had died, and we got another board for it. So that's what's in this machine that we're testing on. And now we have our storage over here. Now, this is my four drives RAID that doesn't really have anything important on it. It's just stuff we downloaded from the internet. So nothing I care about being encrypted. It's two terabyte drives each, so we got four terabytes there. Now, this drive was the other drive we were testing on. This one is encrypted. You see the little icons down here for the encryption, and you see how they're missing down here for this one. This drive is encrypted, and here's our test data encrypted, and here's our test data for this one on the unencrypted drive. So the drives are very similar but not identical, but it's not a one-to-one test. This is mostly to look at the CPU overhead caused by having encryption. Now, when you're doing the test data, I realize this, it'll create the data almost instantly and won't give us any real results unless you turn compression off. So please note that test data here, compression off, and over here, test data, compression off. That's because, and just so you remind you again, if you're doing speed tests on Freenast, and if you're pulling from DevZero, which we will, I'll get to the testing procedure here in a second, when you do that, it ends up compressing all that data and makes it, gives you really artificial results. The other option, you're like, oh, just pull randomly from DevRandom. Yeah, but now you're taxing the CPU. So you'll mess up the other side of the data we're trying to test. So let's see our system baseline here. It's in use. There's things going on in my office. There's data being written to it. So you'll see these little spikes right here. This is, will be what our baseline is. This isn't that data. And if you're not familiar with Freenast 11.1, check out my other review I've did of that in the new features. This is in there now, which is really awesome. I love that they built this in. All right, now let's actually get to do some testing. In the top window, and let me actually move this over so you can see the top window. This is just me running, because someone's asked this before. Z-Pool space, iostat-v1. So every one second, the stats update. And I've just sized, this is Team Ox. I've scaled the window to be sized. So you can kind of just, it's actually scrolling, but it just scrolls so fast you don't notice it. And it looks like you're seeing what's being written to each of these drives. Then down here in this corner, we have our four drives RAID test data. Now, just kind of give you a quick idea. So here's four drives RAID here at the top. And you notice how there's no .eoli, but then we look at Jupyter, we have .eoli. Well, that is the telltale sign that it's an encrypted array. Actually, my NVR is an encrypted array too. So it's an encrypted array, this is an unencrypted array. And trust me on that, if you've seen the previous screen, it's all unencrypted. I really like to make sure I'm very upfront about all of my testing procedures. That way we, you know, make sure we understand what's being tested, how it's being tested. So you can try these results, try this yourself. So the drives are close in speed, they're 7200 p.m., the Tachi drives I believe. I don't have the model numbers handy and they're not relevant anyway to this exact test. But what we're going to do here is dd space if slash dev zero and the OF output file is test dot file in this local directory called four drives RAID. The data set is test data, compression turned off, and we're going to see what happens. So at the top here, we're going to see the drives get lots of writing. This is the four drives RAID one. And through the magic of fast forwarding through data, we've seen it took 27 seconds to get this done. So I've sped up the little part there because no one wants to sit here for 27 seconds in silence. And not even me. So over here, we've got the root Jupyter test data. Like I said, this is the encrypted side. So this is test data encrypted, which it's a data set with the compression turned off so we can hopefully get accurate results. Exactly the same command dd if is dev zero output file test dot file and away we go. And you can see all the rating going on right here. So this is the one that's just getting hammered. There's still some little data rates going on because there's other things happening on the server, but it's going to get you an idea here. So 23 seconds on there. Maybe the drives are a little bit faster. I don't think the encryption part has a lot to do with it, but now let's look at this. Now here is what was going on for the uncompressed or I'm sorry, unencrypted drive. So we have some CPU usage here and it, you know, it peaked up a little bit. Then we have the CPU usage here. So at its peak at the beginning here, we are at about 36% CPU. And at the peak of this one, we hit 42. Yeah, 42% CPU. So quite a bit 54% CPU at its peak. So there we go. So let me zoom in. Oops. Control shift zoom in. There we go. And look at a little more accuracy. So right there, 54% and right there, 36%. And overall, just during the right time while we are waiting for it to write out, you can see that, yeah, there's definitely some CPU spiking there. But this is a four-year-old Intel and it was able to create the data still in regional speed. We did not completely saturate the entire system. Yes, it's more overhead. I feel though the encryption is worth it. Now I don't have time to set up a resilver event. Someone said it takes a lot longer to resilver. Maybe if I have some time, I'll do that as a separate thing. But you're not resilvering that often, hopefully. If your drives fail, it's not too frequent. But let's give you an idea that, yes, definitely. So here's our unencrypted, pretty low, definitely a higher CPU usage. But nothing knocks it out of the park. I mean, this is a four-year-old processor, a new processor. It probably wouldn't even be a blip on there if it was something more, especially like a Xeon or something more powerful or enterprise-ready. But like I said, this is just a twist to show you guys that and show you that it's not that intensive to do the encryption based on my current results on a machine that's running in production real-world. So yes, there is some extra CPU, but nothing to write home about. Just wanted to show you that. And hopefully it answers the question. And it answers it for me that there is a performance hit, but it's really, really minor at that for the performance of this. All right. If you like to count in here, like and subscribe. If you have other questions about FreeNAS, that helps suggest what videos you might want to see on the channel next. All right. Thanks for watching.