WEBVTT 00:00.360 --> 00:06.480 A random array of independent devices better known as Raid has been the primary way that we provide 00:06.480 --> 00:09.930 security to our stored data on our systems. 00:09.930 --> 00:16.260 Now raid is an old idea it's been around for over 25 years but basically the cornerstone of raid simply 00:16.260 --> 00:23.100 says this instead of just using one hard drive to do something I can use multiple hard drives that will 00:23.100 --> 00:29.670 work together to act as one hard drive to do one of two things either number one they're going to provide 00:29.760 --> 00:35.130 some form of data integrity if one drive dies and another one will take up for it. 00:35.280 --> 00:41.310 Or number two it will improve access and in some cases we can actually do both at the same time. 00:41.610 --> 00:45.990 Now one raid was developed way back in the day. 00:45.990 --> 00:51.930 The idea was there would be what were known as RAID levels so be RAID 0 RAID 1 Raid 2 like that and 00:51.930 --> 00:56.880 each one of these RAID levels define a specific way to do this. 00:56.910 --> 01:03.060 So let's march through all the different versions raid started with RAID 0 0 better known as striping 01:03.090 --> 01:08.890 is designed to do one thing and that is to increase the speed that you can get data. 01:09.270 --> 01:14.610 Now it does this by dispersing a piece of data across multiple drives. 01:14.640 --> 01:16.960 So normally here I've got a Word document. 01:16.980 --> 01:22.350 Normally I would just go ahead and chop that up into individual clusters in each one of these clusters 01:22.350 --> 01:27.390 would be saved to one hard drive but with striping what we do is we could use two or three year how 01:27.470 --> 01:28.880 many drives you want. 01:29.160 --> 01:36.150 And in this case what takes place is that that file has saved one cluster is saved on one drive another 01:36.150 --> 01:38.220 cluster on another and another on another. 01:38.250 --> 01:44.550 So you have to have a minimum of two drives to do this striping really speeds things up because we're 01:44.550 --> 01:46.650 not waiting for one hard drive. 01:46.650 --> 01:53.910 However the downside to striping is that if I lose any drive I've lost all my data rate zero speeds 01:53.910 --> 01:54.570 things up. 01:54.570 --> 01:57.470 But it provides no data integrity whatsoever. 01:57.630 --> 02:03.390 If you want data integrity one of the things you're going to turn to is RAID 1 that are known as mirroring. 02:03.480 --> 02:06.820 RAID 1 requires a even number of drive. 02:06.820 --> 02:09.770 So in this case let's just use two hard drives. 02:09.780 --> 02:12.350 Now what we're going to do is we're going to save that word document again. 02:12.360 --> 02:16.340 However this time what we're going to do as we save the Word document. 02:16.380 --> 02:22.160 Each individual cluster is copied or as we say mirrored to the other drive. 02:22.180 --> 02:29.700 Now RAID 0 provides speed without any data integrity and RAID 1 provides data integrity without any 02:29.700 --> 02:33.990 speed increase in fact RAID 1 slows things down because you got to do everything twice. 02:34.020 --> 02:38.220 So now we've got a challenge we would like to speed things up. 02:38.520 --> 02:41.350 But on the same token we'd like to have data integrity. 02:41.520 --> 02:47.580 So to do that we have a number of different RAID levels that into do something called a parity. 02:47.580 --> 02:53.550 So let's say I'm got a mathematical formula for say I got one plus two equals three. 02:53.590 --> 03:01.050 Now that formulas easy to do but if I lost any one part of this formula it would be trivial for someone 03:01.050 --> 03:03.450 mathematically to figure out which part isn't there. 03:03.450 --> 03:08.850 So if I lose the one I know something plus two equals three and mathematically I can figure out what 03:08.850 --> 03:09.600 that is. 03:09.690 --> 03:11.940 That's kind of how parity works. 03:11.940 --> 03:18.720 So what we do in this situation is we will save one chunk of data then we'll save another chunk of data 03:18.780 --> 03:24.300 and they will do some magic math to it and create what we call a parity data piece. 03:24.300 --> 03:30.010 Now the first three times they did this was with raid's levels two through four. 03:30.060 --> 03:31.650 Let me show you how they did it. 03:31.680 --> 03:33.480 So let's take that same word document. 03:33.500 --> 03:38.190 And this time we're going to apply raid we'll call it just raid 2 for the moment. 03:38.190 --> 03:41.160 So I'm going to use a minimum of three drives to do this. 03:41.250 --> 03:45.030 So what I'm going to do is I'm going to go ahead and start saving that document. 03:45.030 --> 03:49.620 So I'm going to put one cluster on the first hard drive then I'm going to put the second cluster on 03:49.620 --> 03:51.350 the second hard drive. 03:51.390 --> 03:55.260 But on the third hard drive I'm going to go through this parity calculation. 03:55.260 --> 03:59.910 So that way if I lose any one of these three drives I'll still be OK. 03:59.910 --> 04:06.600 Now if I keep saving I'm only putting data on the first two drives so that third drive is nothing more 04:06.600 --> 04:09.870 than what we call a dedicated parity drive. 04:10.140 --> 04:14.980 So raid's two three and four pretty much did this all exactly the same. 04:15.210 --> 04:21.530 What you would have would be two or more data drives that would store the individual pieces of data. 04:21.540 --> 04:23.750 We call it Strypes some time too. 04:23.820 --> 04:28.340 And then you would have a dedicated drive that did nothing more than handle parity. 04:28.350 --> 04:33.330 Now they worked out OK in you know years ago I used to see some Raid 3 and raid for solutions. 04:33.510 --> 04:39.690 But what we settled on was a more interesting and more unique idea and that is instead of having one 04:39.690 --> 04:45.600 drive that is just for parity go and just distribute the parity data out almost all the drives. 04:45.690 --> 04:47.680 And that's where range 5 comes into play. 04:47.700 --> 04:50.270 Range five requires a minimum of three drives. 04:50.280 --> 04:55.250 But in this example we're going to use for just because you can use more than three if you want to. 04:55.470 --> 04:57.960 So here's my word document one more time. 04:57.960 --> 05:01.900 Now what I'm going to do is I'm going to say the first chunk to the first drive. 05:02.240 --> 05:07.700 I'm going to save the second chunk to the second drive and what I'm going to do now is I'm going to 05:07.700 --> 05:10.610 stop and I'm going to generate a parity value. 05:10.640 --> 05:12.620 So now I've got the third chunk to save. 05:12.710 --> 05:14.710 We'll go ahead and put that on the fourth drive. 05:15.620 --> 05:20.390 And then the fourth piece will put on the first drive and then generate another parity so you can see 05:20.690 --> 05:25.770 it doesn't matter where the parity is because it's distributed across all the drives. 05:25.850 --> 05:31.180 Rain 5 was the most popular form of raid for a long long time. 05:31.430 --> 05:37.500 The downside to RAID 5 is that you can lose one drive but you can't lose more than one drive. 05:37.520 --> 05:43.420 So if you've got four drives that you're using as a RAID 5 array if you lose one that's fine. 05:43.430 --> 05:45.660 But if you lose to you're in trouble. 05:45.830 --> 05:52.730 So as hard drives get cheaper and cheaper and it became easier to use more and more drives it started 05:52.730 --> 05:54.100 to be a good idea to go. 05:54.110 --> 05:59.930 Could we come up with a way to do this where we could lose two drives and that's the beauty of RAID 05:59.930 --> 06:00.670 6. 06:00.740 --> 06:06.890 Now to make RAID 6 work you need a minimum of four drives so let's go and put four drives up here. 06:06.950 --> 06:13.100 Now as we save stuff we're going to do the same thing where we saved one piece then we save another 06:13.100 --> 06:13.620 piece. 06:13.670 --> 06:18.160 But here's the big difference this time what we're going to do is we're not going to make one parity 06:18.500 --> 06:24.110 but we're going to make two parity so and we'll distribute each of those on separate drives. 06:24.110 --> 06:30.770 Now what you're looking at is that you could lose any two of these drives and you can still recover 06:30.770 --> 06:31.720 all the data. 06:31.940 --> 06:36.460 So raid's zero through six are our basic RAID levels. 06:36.500 --> 06:41.750 However you can do some interesting things with raid and what we call hybrid levels where you can combine 06:41.750 --> 06:42.710 two of these. 06:42.710 --> 06:49.970 So probably the two most common that we're going to combine are known as RAID 0 plus 1 and RAID 1 0 06:50.000 --> 06:51.080 or RAID 10. 06:51.080 --> 06:58.310 Let me show you both of those RAID 0 plus 1 or sometimes just called RAID 0 1 is basically a mirror 06:58.340 --> 07:02.290 of stripes so to do this we're going to need a minimum of four hard drives. 07:02.450 --> 07:07.470 So we're going to have these two hard drives that are going to be working as one side of the mirror 07:07.760 --> 07:10.870 and these two hard drives that will be work it is the other side of the mirror. 07:10.880 --> 07:16.850 So now when I save something what we're going to do is we're first going to go ahead and Mir that which 07:16.850 --> 07:19.580 means we're going to send that data to both sides. 07:19.580 --> 07:21.850 But then we Strype it within those drives. 07:21.860 --> 07:27.620 So if I've got this one chunk of data I'm going to put two stripes on one drive and then I'm going to 07:27.620 --> 07:30.350 mirror those two stripes onto the other side. 07:30.350 --> 07:37.940 So this way what we're basically generating is a mirror of stripes RAID 10 or sometimes called RAID 07:37.940 --> 07:42.200 1 plus 0 is a stripe of mirrors as you might guess. 07:42.200 --> 07:48.170 So let's use the same setup we had before and this time what we're going to do is we're first going 07:48.170 --> 07:53.600 to go ahead and take our piece of data and we're going to stripe it however that stripe gets mirrored 07:53.600 --> 07:55.130 between each drive. 07:55.160 --> 07:57.400 So we get this one piece of data. 07:57.830 --> 07:59.240 We go ahead and stripe it. 07:59.300 --> 08:05.330 So this side gets the first stripe and then this side gets the other stripe However it's mirrored between 08:05.330 --> 08:07.220 the pair drives. 08:07.220 --> 08:11.690 Now the last type of rate I want to talk about is what I'm going to call proprietary raid 08:16.170 --> 08:21.900 there are a lot of folks out there who say there are problems with the established RAID levels and they 08:21.900 --> 08:22.800 can do it better. 08:22.920 --> 08:24.240 Well they're absolutely right. 08:24.240 --> 08:28.740 For example one of the big problems with established RAID levels is that if you don't use the same sized 08:28.740 --> 08:31.440 drives you can get real waste issues. 08:31.590 --> 08:36.730 So if I'm doing RAID 5 for example when I have a two terabyte and a 2 terabyte in one terabyte drive 08:37.020 --> 08:41.680 well you might as well just saw the one terabyte off the other two because RAID 5 in and of itself does 08:41.680 --> 08:42.930 not handle that. 08:42.930 --> 08:46.140 So that's where these proprietary protocols come into play. 08:46.140 --> 08:49.410 In fact I use one here at total seminars. 08:49.410 --> 08:54.530 I use a very popular type of box called Scientology boxes. 08:54.660 --> 08:56.590 Let me get a picture of it up here for you. 08:57.030 --> 08:58.140 Aren't those pretty. 08:58.620 --> 09:02.940 And these boxes are just filled with drives. 09:02.940 --> 09:08.640 So the cool part to these is that you actually access them through a web page and I've got that up here. 09:08.640 --> 09:09.880 Let's take a look at this. 09:10.740 --> 09:15.660 So this is by Saddam digibox and he's just another computer on my network. 09:15.840 --> 09:18.750 So what I'm going to do is I'm going to head over to the storage manager 09:21.450 --> 09:22.780 give him a moment to do his thing. 09:22.800 --> 09:24.610 Looks like everybody's happy. 09:24.820 --> 09:30.300 Now you could actually look here right now I have six physical drives stored in one of these. 09:30.300 --> 09:35.920 And then there's a backup device that I backup to with another six here and there's room for six more. 09:36.120 --> 09:42.630 So what I want to do is first off show you the hard drives so you can see here's my six drives and they're 09:42.720 --> 09:48.450 actually for Terabyte drives 3.6 Tibbits but they're all looking good. 09:48.450 --> 09:50.090 Everybody's happy right now. 09:50.340 --> 09:51.920 And these are the backups. 09:51.960 --> 09:54.620 Now what's actually kind of interesting is when we go to volume's. 09:54.690 --> 10:01.830 So right now I have a primary array of six drives that are called Volume 1. 10:01.900 --> 10:04.660 You can actually see the six drives right here. 10:04.750 --> 10:06.830 But what I want to show you is right here. 10:06.880 --> 10:09.970 So you see it says Chanology hybrid raid. 10:09.970 --> 10:15.980 That's their type of raid with protection for two disks fault tolerance. 10:16.030 --> 10:19.810 So these types of proprietary raids are very popular. 10:19.810 --> 10:20.730 They work great. 10:20.740 --> 10:23.540 So knowledge is not the only game in town for this type of thing. 10:23.650 --> 10:29.500 If you're using a more advanced version of Windows you had the famous storage spaces which is an amazing 10:29.500 --> 10:33.310 tool when it comes to proprietary raid technologies. 10:33.310 --> 10:38.800 The important thing especially for the exam that I want you to keep in your mind is that different levels 10:38.860 --> 10:41.690 of Raid do too. 10:41.710 --> 10:43.870 Or a combination of two things. 10:43.870 --> 10:51.390 Number one they in crease disk access or they improve fault tolerance slash data integrity. 10:51.580 --> 11:15.140 Using race can be one of the more inexpensive ways for you to help secure data on your system.