 Tom here from Warrant Systems and I am working on a review of a Synology RS3621XS Plus with 12 Synology SSDs. The question becomes, how do you set up the rate in it? What's the best, most optimal way to do this configuration? And well, I couldn't find an answer on Google, so I figured I have to create the answer. That's where we are today, is to compare some of the different RAID configurations and file system options, because Synology uses a little vague on that. They tell you that EXT4 has lower system overhead than the BTRFS, but what does that mean? I want to put that into more context with this particular device, share some benchmarks, share some results from those, and how that may play out for your decision making. And what I decided on was the best RAID for this configuration. Before we dive into these details, if you'd like to learn more about me and my company, head over to LawrenceSystems.com. If you'd like to hire a sharp project, there's a hires button right at the top, which does include storage consulting. If you'd like to help this channel out in other ways, there's affiliate links down below to get you deals and discounts on products and services we talk about on this channel. Now, before we get started, I want to address the Synology drives. There is a myth going around and someone's probably already leaving a comment, oh, Synology is trying to force people into using only their drives. Well, no, not exactly, that's not how this is going. This is common in the enterprise market. Companies that provide you a NAS storage device want to certify that that device will perform as expected. So on some of the higher end devices, you'll see Synology recommend using their drives. And this was something they started doing a while ago. And the reason why is really simple. They have a lot of insight to these hard drive manufacturers and probably had some insight that they weren't, well, being forthcoming. And as we've learned from a few breaking news stories here in 2021, which I'll leave links to down below, that the hard drive companies were absolutely bait and switch from what they performed on a benchmark when they provided review units versus what actually came to consumers. And this is a problem when you're a NAS manufacturer and you want to give people the performance they expect and the drives they use claim one level of performance, but then have a different chip and it gives a different level of performance. Synology takes the time to verify this. That's one of the things I want to get out of the way. It's not that all systems are moving to that, but where you have high end systems where they want verification, this is the reason why. Now the test setup here and the type of RAID setup, the BTRFS and EXT were tested both in all these tests. And the important part is this is all done as a storage target for my hypervisor, which is XCP and G. So that's the context of the testing is what if we wanted to use this as a storage target. And I have this set up with the XCP and G host connected with 25 gig card to a 25 gig unified switch to the Synology system, which also has a 25 gig card, the stock system only comes with 10 gig, but a couple of the tests would have hit the 10 gig limit. We actually broke the 10 gig limit and a couple of tests I'll have highlighted in the benchmark results. Now I did do a variety of tests. And one thing that was clear is, well, no matter which rage you use, there's a couple limits we're going to hit. And let's dive right into the numbers and talk about that. Now I will leave a link to this. This is the Pheronix test bench. I ran FIO because we're just doing file system tests. It was running on Ubuntu 2004 was the virtual machine. And as I said, XCP and G and the system is a rise in with eight cores. We're not really hitting CPU limits here. And for the most part, you can kind of see that these are all very similar in speed. And I'll leave the link to this, but this is kind of hard to read. So let's actually do a better summary of all this. And what we're going to do is export this data to a CSV file. And throw it into a spreadsheet. Now, in this spreadsheet gives you a clear picture. Like I said, graphs are cool, but I just wanted to see that there's not a major variation between these numbers right here. Matter of fact, this is RAID 10 with EXT4 versus BTRFS. And it's a little faster in certain circumstances with EXT4, but not substantially 422 versus 468. Yes, that's a speed gain, but it's not substantial. But what about IOPS? Well, IOPS were a little bit faster here, but a little bit slower here. So it kind of depends on the read block versus write block, whether it's the EXT4 or BTRFS. But either way, it's not significant. It's not night and day. And the enhancements you get with the BTRFS are quite a bit more with Synology. And we'll talk a little bit more how Synology implements that later in a video. Then this is the new RAID F1. I'll leave a spec sheet below to the RAID F1 in a white paper. But RAID F1 is a new RAID type managed by Synology. Essentially, the goal of RAID F1 versus traditional RAID is you want the ware leveling on SSDs to be managed very intelligently. This is where RAID F1 comes in and understands that well, if we spread the parity and evenly, we have these out that can cause a problem. So it has a better parity management system for specifically using solid state drives. But I wanted to test the performance. That's why I added it into the list here. And while once again, whether you use BTRFS or EXT4, yes, there's some slight performance differences between them. But the performance difference between the RAID F1 and RAID 10 is not substantial. But the cost savings in drives is we get two and a half terabytes of usable with RAID 10, we get 4.7 terabytes of usable with RAID F1. Now the RAID F1, I believe, only has one driver redundancy in this particular configuration with the 12 drives because I put it all in one pool. But this right here shows that, yeah, with some redundancy and plenty of space to spare here, it's probably the best choice of any of these on here. Just for testing, I did do RAID 0 and this is where we're hitting that limitation that I talked about. You can only go so fast. And once you say the 1,789 megabytes a second, you're talking roughly close to 13 gigabits here. So we exceeded the 10, which is the onboard card and we're using a 25. And the 25 gig card wasn't saturated, but quite a bit. Now this is where the little bit of mystery came in on my testing. And when I see a little bit of mystery, the problem really was for each host connected, they seem to have a limitation. I did say each. And no matter which host I tried, whether it was this Ryzen system we have on our test bench, or the Supermicro Super server, which I reviewed before, and I can leave a link to that, both of them hit that exact same limit. Now I'm sure I figure out where that limit is, that maybe just some of the hardware in Synology can't pass that 13 gig mark, which by the way is really fast. I'm not complaining about that level of speed in here. But this is where the interesting part was, I decided to connect both simultaneously to two different ice guzzy luns. And one host was able to hit that and the next host was able to almost hit that, not quite at the same time, but it shared the data between really impressive and got me closer to about 16 gigs instead of the 12 gig limit. So the hardware can actually get a little bit faster. But I think there is some choke points in there. But that's still really fast. And this is really reasonable for this particular device. Now the last test I did was just the standard RAID six. Once again, it's in line with the other benchmarks and RAID six will give you a couple drives of redundancy. So of course, there's a little bit less in storage at 4.3 terabytes versus 4.7. But still, the performance was quite good. And now of course, all this is leading up to which one should you choose. And I would really say that the Synology F1 having a one driver redundancy, not bad. And using VTRFS gives you a lot of enhanced features from Synology. Now a little bit of back end of how Synology handles VTRFS. The way Synology handles the drives is that each one of these drives is controlled by the Linux RAID utilities that Synology has tuned and optimized to make this work. Then on top of that, just as a file system format, they're using VTRFS. VTRFS does have the ability to directly mount and format the drives, but they're using the Linux RAID on there. My understanding from everything I've read on VTRFS, no file system expert when it comes to this, I'm a little bit more un knowledgeable on some of the other ones such as ZFS. But this is the best of both worlds where you get really solid control from the tried and true Linux RAID managed system. And then you get the enhanced features that are deeply integrated in Synology and all the benefits that come with using VTRFS for well, snapshot, revisioning, self-healing, all the different features they baked in. So I think that's the best combination on this particular device. The performance is really good with these solid state drives. The wear leveling, well, that remains to be seen over so many years, but all the testing Synology's done through the white paper looks really promising and looks like a good idea because well, SSDs are not your standard spinning rust drives. So there are different parameters on there and Synology taking the time to test and certify drives specifically for it makes sense to me. So hopefully this helps and you can extrapolate this probably to several other Synology models. I have a few other that I might be doing some testing with, but I think this RAID testing that goes between there once you hit these SSDs, you're going to hit some speed limits in other parts of the hardware that need to be beefed up because SSDs are really fast. And RAID 10, although RAID 10 is really fast when you're talking about solid states, you probably want more storage because these are a little bit more expensive than the spinning rust and you're actually able to get really good performance out of the Synology RAID F1. So I'll leave links to those benchmarks so you can pour through them and dive into the details a little bit more yourself. But hey, I found this interesting and hopefully you found it helpful. See you in the forums. Thanks. And thank you for making it all the way to the end of this video. If you've enjoyed the content, please give us a thumbs up. If you would like to see more content from this channel, hit the subscribe button and the bell icon. If you'd like to hire a short project, head over to laurancesystems.com and click the hires button right at the top. To help this channel out in other ways, there's a join button here for YouTube and a Patreon page where your support is greatly appreciated. For deals, discounts, and offers, check out our affiliate links in the description of all of our videos, including a link to our shirt store where we have a wide variety of shirts that we sell and designs come out well randomly. So check back frequently. And finally, our forums. Forums.LauranceSystems.com is where you can have a more in-depth discussion about this video and other tech topics covered on this channel. Thanks again for watching and look forward to hearing from you.