1 00:00:00,150 --> 00:00:00,983 In this lesson, 2 00:00:00,983 --> 00:00:03,060 we're going to discuss data redundancy. 3 00:00:03,060 --> 00:00:04,470 Now, one of the most effective ways 4 00:00:04,470 --> 00:00:07,290 to create data redundancy is to use a RAID. 5 00:00:07,290 --> 00:00:10,080 A RAID is a redundant array of independent disks, 6 00:00:10,080 --> 00:00:11,190 which allows an organization 7 00:00:11,190 --> 00:00:13,320 to combine multiple physical storage devices 8 00:00:13,320 --> 00:00:15,150 into a single logical storage device 9 00:00:15,150 --> 00:00:17,400 that's recognized by your operating system. 10 00:00:17,400 --> 00:00:18,990 There are many types of RAIDs out there, 11 00:00:18,990 --> 00:00:20,100 but the most common ones 12 00:00:20,100 --> 00:00:21,930 and most popular ones we're going to discuss 13 00:00:21,930 --> 00:00:25,140 are RAID 0, RAID 1, RAID 5, RAID 6, 14 00:00:25,140 --> 00:00:28,230 and RAID 10. 15 00:00:28,230 --> 00:00:30,090 First, we have a RAID 0. 16 00:00:30,090 --> 00:00:33,030 A RAID 0 provides data striping across multiple disks, 17 00:00:33,030 --> 00:00:34,860 and it's used to increase your performance. 18 00:00:34,860 --> 00:00:36,090 The keyword that you need to remember 19 00:00:36,090 --> 00:00:38,250 for a RAID 0 is striping. 20 00:00:38,250 --> 00:00:40,740 Now, a RAID 0 should be used when you need performance, 21 00:00:40,740 --> 00:00:43,080 but you don't really care about fault tolerance. 22 00:00:43,080 --> 00:00:45,630 In fact, a RAID 0 is the only kind of RAID 23 00:00:45,630 --> 00:00:48,570 that does not provide you with any kind of data redundancy, 24 00:00:48,570 --> 00:00:50,760 but instead, it provides us with faster read 25 00:00:50,760 --> 00:00:51,720 and write speeds 26 00:00:51,720 --> 00:00:54,060 because your data is being split across multiple drives 27 00:00:54,060 --> 00:00:55,290 inside of that RAID 28 00:00:55,290 --> 00:00:57,120 to overcome the physical speed limitations 29 00:00:57,120 --> 00:00:58,470 of traditional hard disk drives 30 00:00:58,470 --> 00:01:00,570 and other slower storage devices. 31 00:01:00,570 --> 00:01:02,130 For example, if I was building 32 00:01:02,130 --> 00:01:04,170 a high-end video editing workstation, 33 00:01:04,170 --> 00:01:05,670 I might want to use a RAID 0 34 00:01:05,670 --> 00:01:08,040 to have two hard disk drives working together, 35 00:01:08,040 --> 00:01:09,720 so I could work on my raw video files 36 00:01:09,720 --> 00:01:12,900 faster than I could by using a single hard disk drive. 37 00:01:12,900 --> 00:01:15,570 To create a RAID 0, you need to have at least two disks 38 00:01:15,570 --> 00:01:18,330 that can be configured to work in tandem with each other. 39 00:01:18,330 --> 00:01:20,280 Second, we have a RAID 1. 40 00:01:20,280 --> 00:01:22,470 Now, a RAID 1 is going to be used to provide redundancy 41 00:01:22,470 --> 00:01:25,440 by mirroring your data identically onto two hard disk drives 42 00:01:25,440 --> 00:01:27,090 or solid state devices. 43 00:01:27,090 --> 00:01:29,640 This way, if one of those storage devices fails, 44 00:01:29,640 --> 00:01:31,470 the other one can continue to operate 45 00:01:31,470 --> 00:01:33,180 because it has a full copy of everything 46 00:01:33,180 --> 00:01:35,220 that was included on those drives. 47 00:01:35,220 --> 00:01:36,360 The key word you need to remember 48 00:01:36,360 --> 00:01:38,580 for a RAID 1 is mirroring. 49 00:01:38,580 --> 00:01:41,040 Now a RAID 1 provides the least amount of downtime 50 00:01:41,040 --> 00:01:42,690 because there's always a complete copy 51 00:01:42,690 --> 00:01:44,880 of your data ready at a moment's notice 52 00:01:44,880 --> 00:01:47,070 that can be served up to your end users. 53 00:01:47,070 --> 00:01:48,510 A RAID 1 is going to provide us 54 00:01:48,510 --> 00:01:50,010 with some great fault tolerance, 55 00:01:50,010 --> 00:01:52,740 but it can only be used with two physical storage devices, 56 00:01:52,740 --> 00:01:54,180 which, in turn, only provides you 57 00:01:54,180 --> 00:01:56,190 with one single logical storage device 58 00:01:56,190 --> 00:01:59,310 that can be read or written to by the operating system. 59 00:01:59,310 --> 00:02:01,980 For a RAID 1, you can only lose one storage device 60 00:02:01,980 --> 00:02:03,690 and still be able to operate the array 61 00:02:03,690 --> 00:02:06,630 without suffering from any data loss or downtime. 62 00:02:06,630 --> 00:02:09,479 For example, on that high-end video editing workstation 63 00:02:09,479 --> 00:02:11,460 I mentioned, I might also want to add 64 00:02:11,460 --> 00:02:13,380 two other physical storage devices 65 00:02:13,380 --> 00:02:15,240 that are configured as a RAID 1. 66 00:02:15,240 --> 00:02:16,350 This way, I can use that 67 00:02:16,350 --> 00:02:18,450 to store all of my finished videos on. 68 00:02:18,450 --> 00:02:20,160 This will ensure that I always have an online 69 00:02:20,160 --> 00:02:21,570 and on-demand backup ready 70 00:02:21,570 --> 00:02:23,460 in case one of my storage devices fails 71 00:02:23,460 --> 00:02:25,530 because both devices will have a fully mirrored 72 00:02:25,530 --> 00:02:28,170 identical copy of the data on them. 73 00:02:28,170 --> 00:02:30,090 Third, we have a RAID 5. 74 00:02:30,090 --> 00:02:32,940 Now, a RAID 5 is known as a striping disk with parity, 75 00:02:32,940 --> 00:02:33,810 and it requires the use 76 00:02:33,810 --> 00:02:35,940 of at least three physical storage devices 77 00:02:35,940 --> 00:02:37,380 in order for it to work. 78 00:02:37,380 --> 00:02:39,450 A RAID 5 will provide us with fault tolerance 79 00:02:39,450 --> 00:02:40,380 by striping the data 80 00:02:40,380 --> 00:02:42,450 across multiple physical storage devices, 81 00:02:42,450 --> 00:02:45,510 as well as writing the parity data across the disk array. 82 00:02:45,510 --> 00:02:46,800 The keywords that you need to remember 83 00:02:46,800 --> 00:02:49,500 for a RAID 5 is striping with parity. 84 00:02:49,500 --> 00:02:51,870 Now, if one disk fails, the other two are going to be able 85 00:02:51,870 --> 00:02:54,360 to reconstruct the data based on the parity, 86 00:02:54,360 --> 00:02:56,760 and that way they can continue their normal operations, 87 00:02:56,760 --> 00:02:58,110 but at a slower speed 88 00:02:58,110 --> 00:02:59,580 because the missing portions of data 89 00:02:59,580 --> 00:03:01,560 have to be calculated on the fly 90 00:03:01,560 --> 00:03:04,080 when the user's requesting the data from the server. 91 00:03:04,080 --> 00:03:06,240 So if one of those drives fails, 92 00:03:06,240 --> 00:03:07,950 you can actually remove the failed drive, 93 00:03:07,950 --> 00:03:09,810 install a new drive into the server, 94 00:03:09,810 --> 00:03:11,430 while the server continues to operate, 95 00:03:11,430 --> 00:03:12,930 which we call a hot swap, 96 00:03:12,930 --> 00:03:14,850 and then the server will rebuild the missing data 97 00:03:14,850 --> 00:03:16,470 onto the newly installed drive 98 00:03:16,470 --> 00:03:18,480 by calculating all those missing pieces 99 00:03:18,480 --> 00:03:20,040 by using the Stripe data pieces 100 00:03:20,040 --> 00:03:23,040 and the parity to calculate the missing dataset. 101 00:03:23,040 --> 00:03:26,040 Now, for a RAID 5, you can only lose one storage device 102 00:03:26,040 --> 00:03:28,050 and still be able to rebuild the data in the RAID 103 00:03:28,050 --> 00:03:29,700 without suffering from any downtime. 104 00:03:29,700 --> 00:03:31,470 So keep that in mind, too. 105 00:03:31,470 --> 00:03:33,450 Fourth, we have a RAID 6. 106 00:03:33,450 --> 00:03:36,360 Now, a RAID 6 is a modified form of a RAID 5 107 00:03:36,360 --> 00:03:37,230 that relies on the use 108 00:03:37,230 --> 00:03:39,480 of data striping across multiple storage devices, 109 00:03:39,480 --> 00:03:40,950 like a RAID 5 does, 110 00:03:40,950 --> 00:03:43,320 but instead of having a single piece of parity data, 111 00:03:43,320 --> 00:03:45,330 we're actually going to have two identical pieces 112 00:03:45,330 --> 00:03:46,590 of parity data. 113 00:03:46,590 --> 00:03:47,640 I like to joke around and say 114 00:03:47,640 --> 00:03:49,980 that "RAID 6 is one better than RAID 5, 115 00:03:49,980 --> 00:03:51,960 so instead of having one set of parity, 116 00:03:51,960 --> 00:03:53,490 we now have two sets of parity 117 00:03:53,490 --> 00:03:55,830 because it is one more than RAID 5." 118 00:03:55,830 --> 00:03:57,420 Now, to achieve this double parity, 119 00:03:57,420 --> 00:03:59,700 your RAID must have at least four storage devices 120 00:03:59,700 --> 00:04:01,830 contained within your RAID 6 array. 121 00:04:01,830 --> 00:04:03,360 Now, the benefit of using a RAID 6 122 00:04:03,360 --> 00:04:05,220 over a RAID 5 is that you can now lose 123 00:04:05,220 --> 00:04:06,480 two of your storage devices 124 00:04:06,480 --> 00:04:08,160 and still be able to rebuild the data 125 00:04:08,160 --> 00:04:10,020 without suffering from any downtime. 126 00:04:10,020 --> 00:04:13,290 So it does give you some more resiliency and redundancy. 127 00:04:13,290 --> 00:04:14,550 Now, the keywords you need to remember 128 00:04:14,550 --> 00:04:17,550 for a RAID 6 is striping with double parity. 129 00:04:17,550 --> 00:04:21,570 Fifth, we have a RAID 10, also known as a RAID 10. 130 00:04:21,570 --> 00:04:23,580 Now, a RAID 10, or a RAID 10 131 00:04:23,580 --> 00:04:25,470 will combine the advantages of a RAID 1 132 00:04:25,470 --> 00:04:28,500 and a RAID 0 into a single RAID array. 133 00:04:28,500 --> 00:04:31,590 Essentially, a RAID 10 is going to be a striped array 134 00:04:31,590 --> 00:04:34,500 that contains two sets of mirrored arrays within it. 135 00:04:34,500 --> 00:04:36,780 This is why we call it a RAID 10 136 00:04:36,780 --> 00:04:38,370 because it first mirrors the two sets 137 00:04:38,370 --> 00:04:40,380 of storage devices into a RAID 1, 138 00:04:40,380 --> 00:04:42,360 and then it stripes the data across both sets 139 00:04:42,360 --> 00:04:45,660 of mirrored storage devices using a RAID 0 configuration. 140 00:04:45,660 --> 00:04:49,440 So RAID 1 plus RAID 0 gives us RAID 10, 141 00:04:49,440 --> 00:04:51,720 or RAID 10 as we write it out. 142 00:04:51,720 --> 00:04:52,980 Now, the keywords you need to remember 143 00:04:52,980 --> 00:04:54,390 for a RAID 10 is that 144 00:04:54,390 --> 00:04:57,030 it is a striped array of mirrored arrays. 145 00:04:57,030 --> 00:04:59,880 Now, a RAID 10 or RAID 10 also supports the use 146 00:04:59,880 --> 00:05:02,190 of even numbers of physical storage devices 147 00:05:02,190 --> 00:05:03,840 to create this kind of an array, 148 00:05:03,840 --> 00:05:06,360 and you have to at least force physical storage devices 149 00:05:06,360 --> 00:05:08,730 in the RAID to create a minimum of two sets 150 00:05:08,730 --> 00:05:10,740 of mirrored RAID 1s to be able to use 151 00:05:10,740 --> 00:05:13,050 inside of your RAID 10. 152 00:05:13,050 --> 00:05:14,580 A RAID 10 will provide us 153 00:05:14,580 --> 00:05:16,500 with fault tolerance due to the mirrored arrays 154 00:05:16,500 --> 00:05:18,390 inside of the RAID 10. 155 00:05:18,390 --> 00:05:20,310 But it also provides us with faster speeds 156 00:05:20,310 --> 00:05:21,660 because of the striping of the data 157 00:05:21,660 --> 00:05:24,030 across two different sets of mirrored arrays 158 00:05:24,030 --> 00:05:25,773 inside of the RAID 10. 159 00:05:25,773 --> 00:05:28,050 Now, a RAID 10 can support the loss 160 00:05:28,050 --> 00:05:29,820 of up to two of your storage devices 161 00:05:29,820 --> 00:05:31,410 and still be able to rebuild the data 162 00:05:31,410 --> 00:05:33,300 without suffering from any downtime, 163 00:05:33,300 --> 00:05:35,220 as long as the loss of the storage devices 164 00:05:35,220 --> 00:05:38,040 are not both located in the same mirrored array. 165 00:05:38,040 --> 00:05:39,930 So if you lose one storage device 166 00:05:39,930 --> 00:05:41,910 from each mirrored set, that's fine. 167 00:05:41,910 --> 00:05:43,800 The RAID 10 will continue to operate 168 00:05:43,800 --> 00:05:44,970 without any issues, 169 00:05:44,970 --> 00:05:47,400 but you will have lost your fault tolerance capability 170 00:05:47,400 --> 00:05:49,830 if you lost a drive from each set of the mirrors. 171 00:05:49,830 --> 00:05:52,050 So you'll want to replace those failed storage devices 172 00:05:52,050 --> 00:05:53,550 as soon as possible. 173 00:05:53,550 --> 00:05:55,320 Now, when we talk about the classification 174 00:05:55,320 --> 00:05:56,970 and categorization of our RAIDs, 175 00:05:56,970 --> 00:05:59,301 we call these either failure-resistant, fault-tolerant, 176 00:05:59,301 --> 00:06:00,134 or disaster-tolerant. 177 00:06:00,134 --> 00:06:01,648 Each of these terms describes a level of resilience 178 00:06:01,648 --> 00:06:03,893 that a RAID system can provide against data loss 179 00:06:03,893 --> 00:06:07,470 and system downtime. 180 00:06:07,470 --> 00:06:09,690 When we're dealing with failure-resistant systems, 181 00:06:09,690 --> 00:06:10,890 these are systems that are designed 182 00:06:10,890 --> 00:06:12,930 to withstand certain hardware malfunctions 183 00:06:12,930 --> 00:06:14,670 without losing the associated data. 184 00:06:14,670 --> 00:06:16,470 And failure resistance is going to be achieved 185 00:06:16,470 --> 00:06:18,540 through redundancy of your storage devices 186 00:06:18,540 --> 00:06:21,480 by mirroring data across multiple storage devices. 187 00:06:21,480 --> 00:06:23,280 This means that if one device fails, 188 00:06:23,280 --> 00:06:24,750 the system can continue to operate 189 00:06:24,750 --> 00:06:26,940 using the data on the remaining devices. 190 00:06:26,940 --> 00:06:29,520 So when you think about a failure-resistant system, 191 00:06:29,520 --> 00:06:30,960 we're really talking about any of the systems 192 00:06:30,960 --> 00:06:33,090 that use mirroring, which is a RAID 1 193 00:06:33,090 --> 00:06:36,240 or a RAID 10. 194 00:06:36,240 --> 00:06:38,220 Now, a fault-tolerant system, on the other hand, 195 00:06:38,220 --> 00:06:39,660 is going to be designed to not only allow 196 00:06:39,660 --> 00:06:41,010 for the continued operation 197 00:06:41,010 --> 00:06:43,050 without any downtime being experienced, 198 00:06:43,050 --> 00:06:45,240 but also allows the system to maintain its operations 199 00:06:45,240 --> 00:06:47,070 in the event of a hardware failure 200 00:06:47,070 --> 00:06:49,500 due to either mirroring or striping with parity, 201 00:06:49,500 --> 00:06:53,880 such as in a RAID 1, RAID 5, RAID 6, or RAID 10. 202 00:06:53,880 --> 00:06:56,610 With fault tolerant systems, your system has the ability 203 00:06:56,610 --> 00:06:58,230 to quickly rebuild any lost data 204 00:06:58,230 --> 00:07:00,060 from the remaining healthy storage devices 205 00:07:00,060 --> 00:07:02,280 to ensure your systems remain resilient. 206 00:07:02,280 --> 00:07:04,050 Now, a disaster-tolerant system 207 00:07:04,050 --> 00:07:05,910 includes a broader level of protection 208 00:07:05,910 --> 00:07:08,760 by safeguarding your data against catastrophic events. 209 00:07:08,760 --> 00:07:11,880 Disaster-tolerant RAIDs utilize two independent zones 210 00:07:11,880 --> 00:07:14,400 that always have full access to all of the data. 211 00:07:14,400 --> 00:07:17,070 A RAID 1 and a RAID 10 are both examples 212 00:07:17,070 --> 00:07:18,720 of disaster-tolerant RAIDs 213 00:07:18,720 --> 00:07:20,580 because they have full mirrors of the data 214 00:07:20,580 --> 00:07:23,550 on at least one of the storage devices within the array. 215 00:07:23,550 --> 00:07:25,950 So remember, RAIDs are a great thing to use 216 00:07:25,950 --> 00:07:26,820 when you're trying to ensure 217 00:07:26,820 --> 00:07:28,770 you have a good redundancy of your data, 218 00:07:28,770 --> 00:07:31,590 both online and available at all times. 219 00:07:31,590 --> 00:07:34,050 A RAID is a redundant array of independent disks, 220 00:07:34,050 --> 00:07:35,160 which allows an organization 221 00:07:35,160 --> 00:07:37,200 to combine multiple physical storage devices 222 00:07:37,200 --> 00:07:38,543 into a single logical storage device 223 00:07:38,543 --> 00:07:41,280 that is recognized by your operating system. 224 00:07:41,280 --> 00:07:43,260 A RAID 0 is going to enhance the performance 225 00:07:43,260 --> 00:07:44,700 by striping your data evenly 226 00:07:44,700 --> 00:07:46,470 across two or more storage devices 227 00:07:46,470 --> 00:07:48,240 without any kind of redundancy 228 00:07:48,240 --> 00:07:50,340 to facilitate a high speed access, 229 00:07:50,340 --> 00:07:53,070 but you're not going to have any fault tolerance. 230 00:07:53,070 --> 00:07:56,190 A RAID 1 is going to maintain an exact copy of the data, 231 00:07:56,190 --> 00:07:57,390 which is known as a mirror, 232 00:07:57,390 --> 00:07:58,890 on two or more storage devices 233 00:07:58,890 --> 00:08:00,390 to increase your read performance 234 00:08:00,390 --> 00:08:02,550 and ensure data integrity and availability, 235 00:08:02,550 --> 00:08:05,490 even if a single storage device in the RAID fails. 236 00:08:05,490 --> 00:08:08,490 A RAID 5 is going to distribute data and parity information 237 00:08:08,490 --> 00:08:11,100 across three or more disks to improve the read 238 00:08:11,100 --> 00:08:14,190 and write performance while also providing fault tolerance. 239 00:08:14,190 --> 00:08:16,410 A RAID 6 is similar to a RAID 5, 240 00:08:16,410 --> 00:08:18,090 but it adds an extra parity block 241 00:08:18,090 --> 00:08:19,680 so that the data and double parity 242 00:08:19,680 --> 00:08:22,380 is going to be striped across four or more storage devices 243 00:08:22,380 --> 00:08:23,550 to help withstand the failure 244 00:08:23,550 --> 00:08:25,560 of two storage devices simultaneously 245 00:08:25,560 --> 00:08:27,690 without having any kind of data loss. 246 00:08:27,690 --> 00:08:29,700 A RAID 10, 247 00:08:29,700 --> 00:08:32,669 will combine the advantages of a RAID 1 and a RAID 0 248 00:08:32,669 --> 00:08:35,909 by striping data across mirrored sets of storage devices 249 00:08:35,909 --> 00:08:38,010 to offer higher performance, fault tolerance, 250 00:08:38,010 --> 00:08:40,169 and improved data redundancy. 251 00:08:40,169 --> 00:08:41,669 By relying on the use of RAIDs, 252 00:08:41,669 --> 00:08:43,679 we can increase our overall availability 253 00:08:43,679 --> 00:08:45,660 when we're designing high availability systems 254 00:08:45,660 --> 00:08:47,073 for our enterprise networks.