1 00:00:00,400 --> 00:00:04,280 So now let's talk about the AWS Snow Family. 2 00:00:04,280 --> 00:00:08,310 So it represents a highly-secure, portable devices 3 00:00:08,310 --> 00:00:11,810 that has two use cases within AWS. 4 00:00:11,810 --> 00:00:15,170 Either it's used to collect and process data at the edge, 5 00:00:15,170 --> 00:00:19,710 or to migrate data in and out of AWS. 6 00:00:19,710 --> 00:00:21,310 So we have two use cases of data migration. 7 00:00:21,310 --> 00:00:22,950 And for this, we have three different types 8 00:00:22,950 --> 00:00:24,900 of devices within the Snow Family. 9 00:00:24,900 --> 00:00:26,243 We have the Snowcone, the Snowball Edge, and the Snowmobile. 10 00:00:29,445 --> 00:00:31,800 And for the second use case, edge computing, 11 00:00:31,800 --> 00:00:34,730 we have the Snowcone and the Snowball Edge. 12 00:00:34,730 --> 00:00:37,180 So we'll first tackle the data migration subject 13 00:00:37,180 --> 00:00:39,250 and then the edge computing. 14 00:00:39,250 --> 00:00:41,920 So why do we want to do data immigration 15 00:00:41,920 --> 00:00:44,600 with the AWS Snow Family? 16 00:00:44,600 --> 00:00:47,560 Well, if we look at the time it takes to transfer 17 00:00:47,560 --> 00:00:50,330 a lot of data over the network, 18 00:00:50,330 --> 00:00:52,080 it can take a lot of time. 19 00:00:52,080 --> 00:00:56,340 So, for example, if we want to transfer 100 terabytes 20 00:00:56,340 --> 00:01:00,170 over one gigabits per second network line, 21 00:01:00,170 --> 00:01:02,993 it will take us 12 days to achieve it. 22 00:01:03,980 --> 00:01:06,100 And so, obviously, if we do a petabyte, 23 00:01:06,100 --> 00:01:07,780 it will take forever and so on. 24 00:01:07,780 --> 00:01:08,710 So as we can see, 25 00:01:08,710 --> 00:01:12,260 sometimes we just wanna do that to get to AWS fast. 26 00:01:12,260 --> 00:01:14,080 And the challenges that sometimes 27 00:01:14,080 --> 00:01:17,350 on top of having a small network transfer, 28 00:01:17,350 --> 00:01:21,400 you have limited connectivity, limited bandwidth, sorry. 29 00:01:21,400 --> 00:01:24,450 Transferring data over the network may cost you some money. 30 00:01:24,450 --> 00:01:26,820 It's not free to use a network. 31 00:01:26,820 --> 00:01:28,820 It could be that also the bandwidth is shared. 32 00:01:28,820 --> 00:01:31,770 For example, if you download a video from AWS 33 00:01:31,770 --> 00:01:35,930 and you download 10 terabytes of data, 34 00:01:35,930 --> 00:01:38,440 maybe you're going to block your entire office, 35 00:01:38,440 --> 00:01:41,750 because you're maximizing the bandwidth within your office. 36 00:01:41,750 --> 00:01:44,490 And then maybe the connection is not stable enough, 37 00:01:44,490 --> 00:01:46,760 so you have to retry and so on. 38 00:01:46,760 --> 00:01:51,060 So all of these reasons make a case for Snow Family. 39 00:01:51,060 --> 00:01:53,520 So the Snow Family are offline devices 40 00:01:53,520 --> 00:01:55,830 that allow you to perform data migrations. 41 00:01:55,830 --> 00:01:59,030 So AWS will send you an actual physical device 42 00:01:59,030 --> 00:02:03,430 by the post office, and then you load your data onto it, 43 00:02:03,430 --> 00:02:05,003 and then send it back to AWS. 44 00:02:05,870 --> 00:02:08,600 So the rule of thumb is that if it takes more than a week 45 00:02:08,600 --> 00:02:10,490 to transfer data over the network, 46 00:02:10,490 --> 00:02:13,950 then you should use a Snowball device, for example. 47 00:02:13,950 --> 00:02:16,450 So to really explain how that works, let's take an example. 48 00:02:16,450 --> 00:02:19,570 If we wanted to directly upload a file into Amazon S3, 49 00:02:19,570 --> 00:02:22,160 we have the clients send the data into Amazon S3. 50 00:02:22,160 --> 00:02:23,490 Very easy, right? 51 00:02:23,490 --> 00:02:25,440 But with the Snow Family, for example, 52 00:02:25,440 --> 00:02:26,850 with a Snowball device, 53 00:02:26,850 --> 00:02:29,020 the clients request this novel device. 54 00:02:29,020 --> 00:02:30,390 We receive it via the post. 55 00:02:30,390 --> 00:02:33,100 AWS will deliver the device to us. 56 00:02:33,100 --> 00:02:36,660 We loaded the data directly onto the devices locally, 57 00:02:36,660 --> 00:02:40,500 and then we ship back the device to AWS 58 00:02:40,500 --> 00:02:42,620 into an AWS facility. 59 00:02:42,620 --> 00:02:44,500 Then they will take the device, 60 00:02:44,500 --> 00:02:47,510 and they will plug it into their own infrastructure. 61 00:02:47,510 --> 00:02:50,010 And then the data will be imported or exported 62 00:02:50,010 --> 00:02:53,400 based on what you wanna do to an Amazon S3 bucket, 63 00:02:53,400 --> 00:02:54,620 and you're good to go. 64 00:02:54,620 --> 00:02:57,450 So really it is a way to transfer data to AWS, 65 00:02:57,450 --> 00:03:00,533 but through the physical, not the network route. 66 00:03:01,550 --> 00:03:04,370 So now what sort of devices do we have? 67 00:03:04,370 --> 00:03:06,320 We have the Snowball Edge, 68 00:03:06,320 --> 00:03:09,210 and snowball edge is a huge box, as you can see, 69 00:03:09,210 --> 00:03:12,330 and it is going to be used to move terabytes or petabytes 70 00:03:12,330 --> 00:03:15,410 of data in and out of AWS. 71 00:03:15,410 --> 00:03:17,060 It's going to be an alternative to moving data 72 00:03:17,060 --> 00:03:18,520 over the network, how we've seen. 73 00:03:18,520 --> 00:03:21,080 We're going to pay per data transfer job. 74 00:03:21,080 --> 00:03:24,230 And the interface within the Snowball Edge 75 00:03:24,230 --> 00:03:25,960 is going to provide a block storage 76 00:03:25,960 --> 00:03:29,150 or Amazon S3-compatible object storage. 77 00:03:29,150 --> 00:03:32,060 So we have two flavors for the Snowball Edge. 78 00:03:32,060 --> 00:03:34,640 We have the Snowball Edge Storage Optimized, 79 00:03:34,640 --> 00:03:36,850 which is going to give us 80 terabytes 80 00:03:36,850 --> 00:03:41,570 of hardware of a hardware disc capacity, 81 00:03:41,570 --> 00:03:42,810 which works for block volume 82 00:03:42,810 --> 00:03:44,690 or S3-compatible object storage. 83 00:03:44,690 --> 00:03:47,750 And we have the Snowball Edge Compute Optimized, 84 00:03:47,750 --> 00:03:51,900 which is going to give us 42 terabytes of HDD capacity. 85 00:03:51,900 --> 00:03:53,930 So, really, if we want to have more storage, 86 00:03:53,930 --> 00:03:55,350 obviously, we want to get the flavor 87 00:03:55,350 --> 00:03:59,070 that is called Snowball Edge Storage Optimized. 88 00:03:59,070 --> 00:04:01,180 So the use case for a Snowball Edge 89 00:04:01,180 --> 00:04:02,820 for data transfers so far 90 00:04:02,820 --> 00:04:05,410 is to do a large data cloud migration 91 00:04:05,410 --> 00:04:07,170 to decommission a data center, 92 00:04:07,170 --> 00:04:08,860 or maybe to do disaster recovery 93 00:04:08,860 --> 00:04:10,867 by backing up your data into AWS. 94 00:04:12,060 --> 00:04:13,490 So that's for the first option. 95 00:04:13,490 --> 00:04:17,083 The second option is called AWS Snowcone, 96 00:04:17,083 --> 00:04:18,529 and Snowcone is a cute little name, 97 00:04:18,529 --> 00:04:20,000 because it is a cute little thing. 98 00:04:20,000 --> 00:04:22,100 This is a much smaller device, as we can see. 99 00:04:22,100 --> 00:04:23,960 We can see an ethernet port on it. 100 00:04:23,960 --> 00:04:27,980 So, Snowcone is much smaller than a Snowball Edge. 101 00:04:27,980 --> 00:04:31,120 And this is a small portable computing, anywhere, 102 00:04:31,120 --> 00:04:34,470 rugged, and secure, and it can withstand harsh environments. 103 00:04:34,470 --> 00:04:36,590 So it can go in desert. 104 00:04:36,590 --> 00:04:37,620 It can go in the water. 105 00:04:37,620 --> 00:04:40,540 It's really, really light and secure. 106 00:04:40,540 --> 00:04:42,510 So this device is going to be used for edge computing, 107 00:04:42,510 --> 00:04:43,690 storage, and data transfer, 108 00:04:43,690 --> 00:04:45,840 but, obviously, of much smaller size. 109 00:04:45,840 --> 00:04:49,790 And as such, you can store eight terabytes of storage 110 00:04:49,790 --> 00:04:53,150 on a Snowcone, so much less, about 10 times less 111 00:04:53,150 --> 00:04:56,890 than this Snowball Edge Storage Optimized device. 112 00:04:56,890 --> 00:05:00,140 So you should use a Snowcone where a Snowball does not fit. 113 00:05:00,140 --> 00:05:02,570 So, for example, in space-constrained environments, 114 00:05:02,570 --> 00:05:05,460 you can even put a Snowcone on top of a drone 115 00:05:05,460 --> 00:05:06,710 if you wanted to. 116 00:05:06,710 --> 00:05:09,040 You must provide your own battery and cables. 117 00:05:09,040 --> 00:05:12,140 And it can be sent back to AWS offline, 118 00:05:12,140 --> 00:05:14,200 or you can even connect it to the network, 119 00:05:14,200 --> 00:05:17,480 and it will use AWS DataSync to send the data back 120 00:05:17,480 --> 00:05:18,400 over the network. 121 00:05:18,400 --> 00:05:21,000 So imagine you have a Snowcone into an environment 122 00:05:21,000 --> 00:05:23,140 that doesn't have any network connectivity. 123 00:05:23,140 --> 00:05:25,490 It will collect and collect and collect and collect data. 124 00:05:25,490 --> 00:05:27,810 And then you go back home, you go to your data center. 125 00:05:27,810 --> 00:05:30,040 You take the Snowcone and you plug it into your data center, 126 00:05:30,040 --> 00:05:33,590 and automatically the data is being transferred to AWS, 127 00:05:33,590 --> 00:05:36,000 or you can just send it back, again, offline to AWS, 128 00:05:36,000 --> 00:05:38,743 but you have two options with the Snowcone. 129 00:05:40,010 --> 00:05:42,950 And then Snowmobile is an actual truck. 130 00:05:42,950 --> 00:05:44,320 So when they announced it, 131 00:05:44,320 --> 00:05:46,030 they actually took a truck on stage 132 00:05:46,030 --> 00:05:48,280 to show that it was an actual truck 133 00:05:48,280 --> 00:05:50,210 that is going to transfer data. 134 00:05:50,210 --> 00:05:51,620 And so with the Snowmobile, 135 00:05:51,620 --> 00:05:53,520 you can transfer exabytes of data. 136 00:05:53,520 --> 00:05:57,900 So one exabyte is 1,000 petabytes is one million terabytes. 137 00:05:57,900 --> 00:06:01,670 And each snowmobile will have 100 petabytes of capacity. 138 00:06:01,670 --> 00:06:04,230 So if you wanted to reach one exabyte of data, 139 00:06:04,230 --> 00:06:07,270 you need to order 10 Snowmobiles. 140 00:06:07,270 --> 00:06:09,360 It's high security, it's temperature controlled, 141 00:06:09,360 --> 00:06:11,490 has GPS, 24/7 video surveillance, 142 00:06:11,490 --> 00:06:13,700 so it's quite a secure way to transfer your data. 143 00:06:13,700 --> 00:06:16,490 And it's a better use case than Snowball 144 00:06:16,490 --> 00:06:20,300 if you start transferring more than 10 petabytes of data. 145 00:06:20,300 --> 00:06:22,400 So, as a summary, for data immigration, 146 00:06:22,400 --> 00:06:24,180 we have three options. 147 00:06:24,180 --> 00:06:27,420 We have Snowcone, Snowball Edge, Snowmobile, 148 00:06:27,420 --> 00:06:29,540 and each come with different storage capacities, 149 00:06:29,540 --> 00:06:33,240 so eight terabytes, 80 terabytes, and 100 petabytes. 150 00:06:33,240 --> 00:06:35,660 The migration size that's recommended by AWS 151 00:06:35,660 --> 00:06:39,140 is Snowcone is up to 24 terabytes. 152 00:06:39,140 --> 00:06:41,730 For a Snowball Edge, it's up to petabytes, 153 00:06:41,730 --> 00:06:44,260 and it's offline, because you have to send it back to AWS. 154 00:06:44,260 --> 00:06:48,681 And for Snowmobile, the use case is up to exabytes of data. 155 00:06:48,681 --> 00:06:50,623 DataSync agent is pre-installed on a Snowcone, 156 00:06:50,623 --> 00:06:52,710 because you can plug it to a network 157 00:06:52,710 --> 00:06:54,750 and have DataSync send the data over the network 158 00:06:54,750 --> 00:06:56,210 to AWS as well. 159 00:06:56,210 --> 00:06:58,780 And for Snowball Edge, you can do storage clustering, 160 00:06:58,780 --> 00:07:02,550 so you can put up to 15 Snowball Edges together 161 00:07:02,550 --> 00:07:05,030 to increase the storage size. 162 00:07:05,030 --> 00:07:08,970 Okay, so how do we use a Snow Family device? 163 00:07:08,970 --> 00:07:11,660 Well, you request the device from the console for delivery, 164 00:07:11,660 --> 00:07:13,290 and we'll see this in the hands-on. 165 00:07:13,290 --> 00:07:14,900 Then we install the Snowball client, 166 00:07:14,900 --> 00:07:18,000 or we use AWS OpsHub that we'll see in this lecture 167 00:07:18,000 --> 00:07:19,620 on your servers. 168 00:07:19,620 --> 00:07:21,490 Then we connect the Snowball to the servers 169 00:07:21,490 --> 00:07:23,890 and start copying the files using the client. 170 00:07:23,890 --> 00:07:25,780 Then we ship back the device when we're ready. 171 00:07:25,780 --> 00:07:28,430 It will go straight to the right AWS facility 172 00:07:28,430 --> 00:07:30,530 thanks to an E Ink marker, 173 00:07:30,530 --> 00:07:33,540 and the data will be loaded onto an S3 bucket. 174 00:07:33,540 --> 00:07:35,770 And then the Snowball will be completely wiped 175 00:07:35,770 --> 00:07:38,460 according to the highest security measures. 176 00:07:38,460 --> 00:07:40,160 So that's for the data migration. 177 00:07:40,160 --> 00:07:43,900 And that was originally one and the only use case 178 00:07:43,900 --> 00:07:45,350 for Snowball devices. 179 00:07:45,350 --> 00:07:47,820 But the second use case now for the Snow Family 180 00:07:47,820 --> 00:07:49,860 is called edge computing. 181 00:07:49,860 --> 00:07:52,640 And so edge computing is when you process data 182 00:07:52,640 --> 00:07:56,200 while it's being created at an edge location. 183 00:07:56,200 --> 00:07:57,980 So what is an edge location? 184 00:07:57,980 --> 00:07:59,800 Well, an edge location is anything 185 00:07:59,800 --> 00:08:01,700 that really doesn't have internet, 186 00:08:01,700 --> 00:08:03,900 or that is far away from a cloud. 187 00:08:03,900 --> 00:08:06,400 So, for example, if you have a truck on the road, 188 00:08:06,400 --> 00:08:08,130 or if you have a ship on the sea, 189 00:08:08,130 --> 00:08:10,150 or a mining station underground, 190 00:08:10,150 --> 00:08:12,500 all these things can be called edge locations, 191 00:08:12,500 --> 00:08:14,360 because they can produce data, 192 00:08:14,360 --> 00:08:17,450 but they may not necessarily have internet connectivity, 193 00:08:17,450 --> 00:08:19,580 so, either a limited connectivity, 194 00:08:19,580 --> 00:08:23,770 or no internet access, or no access to computing power. 195 00:08:23,770 --> 00:08:27,450 And so you may still want to run computation data processing 196 00:08:27,450 --> 00:08:28,730 at these locations. 197 00:08:28,730 --> 00:08:31,170 And for this, we need edge computing. 198 00:08:31,170 --> 00:08:32,600 And so to do edge computing, 199 00:08:32,600 --> 00:08:35,760 we can order a Snowball Edge device or a Snowcone, 200 00:08:35,760 --> 00:08:39,130 and have it embedded into these edge locations 201 00:08:39,130 --> 00:08:41,230 and start doing edge computing. 202 00:08:41,230 --> 00:08:44,470 So the use cases of edge computing is to preprocess data 203 00:08:44,470 --> 00:08:45,860 to machine learning at the edge, 204 00:08:45,860 --> 00:08:48,270 so without it going back to the cloud, 205 00:08:48,270 --> 00:08:50,100 transcode media streams in advance, 206 00:08:50,100 --> 00:08:51,820 and eventually, if need be, 207 00:08:51,820 --> 00:08:54,173 if you need to transfer the data back into AWS, 208 00:08:55,030 --> 00:08:56,870 you can ship back the device 209 00:08:56,870 --> 00:08:58,630 for your Snowcone or your Snowball Edge. 210 00:08:58,630 --> 00:09:01,040 So really you start processing the data very, very close 211 00:09:01,040 --> 00:09:02,300 to where it's being created, 212 00:09:02,300 --> 00:09:05,490 and then you ship it back to AWS. 213 00:09:05,490 --> 00:09:07,340 So, for edge computing, what do we have? 214 00:09:07,340 --> 00:09:10,270 We have the Snowcone, and it comes with two CPUs, 215 00:09:10,270 --> 00:09:13,620 four gigabytes of memory, has wired or wireless access, 216 00:09:13,620 --> 00:09:14,620 so WiFi. 217 00:09:14,620 --> 00:09:19,200 It's powered by USB-C and an optional battery. 218 00:09:19,200 --> 00:09:21,730 Then for Snowball Edge, it's compute optimized, 219 00:09:21,730 --> 00:09:23,320 so we have two different flavors again. 220 00:09:23,320 --> 00:09:26,680 So, the Compute Optimized one has 52 vCPUs, 221 00:09:26,680 --> 00:09:30,640 200 gigabytes of RAM, and an optional GPU 222 00:09:30,640 --> 00:09:34,300 if you want to do video processing or machine learning. 223 00:09:34,300 --> 00:09:37,870 We have 42 terabytes of usable storage. 224 00:09:37,870 --> 00:09:39,730 And for the Snowball Edge that is Storage Optimized, 225 00:09:39,730 --> 00:09:43,260 we have 40 CPU, so less and 80 gigabytes of RAM, 226 00:09:43,260 --> 00:09:44,840 so, again, less RAM as well. 227 00:09:44,840 --> 00:09:46,950 We have objects storage clustering available 228 00:09:46,950 --> 00:09:48,310 for the storage. 229 00:09:48,310 --> 00:09:52,380 And all of these devices can run either EC2 instances 230 00:09:52,380 --> 00:09:54,750 within them, or Lambda functions 231 00:09:54,750 --> 00:09:58,470 using the service called AWS IoT Greengrass. 232 00:09:58,470 --> 00:10:00,090 And if you do edge computing, 233 00:10:00,090 --> 00:10:03,350 then you may want to have these devices in your facilities, 234 00:10:03,350 --> 00:10:06,290 in your trucks, in your boats for a very long time. 235 00:10:06,290 --> 00:10:08,950 And so, therefore, you have long-term deployment options 236 00:10:08,950 --> 00:10:10,370 where you can get a discounted pricing 237 00:10:10,370 --> 00:10:14,903 if you borrow these devices for one or three years. 238 00:10:16,600 --> 00:10:18,740 Finally, for the Snow Family, 239 00:10:18,740 --> 00:10:21,160 there is OpsHub. 240 00:10:21,160 --> 00:10:24,020 So, historically, when you were using these devices, 241 00:10:24,020 --> 00:10:27,070 you needed a CLI, so Command Line Interface tool, 242 00:10:27,070 --> 00:10:29,560 to deal with them, and it was very, very difficult. 243 00:10:29,560 --> 00:10:31,370 And so AWS recognizes that. 244 00:10:31,370 --> 00:10:33,310 And so they've created OpsHub, 245 00:10:33,310 --> 00:10:34,880 which is a software that you install 246 00:10:34,880 --> 00:10:36,080 on your computer or laptop, 247 00:10:36,080 --> 00:10:37,397 so it's not something you use on the cloud. 248 00:10:37,397 --> 00:10:40,050 It's something that you have to download on your computer. 249 00:10:40,050 --> 00:10:42,040 And then once it's connected, it's going to give you 250 00:10:42,040 --> 00:10:46,220 a graphical interface to connect to your Snow devices 251 00:10:46,220 --> 00:10:49,700 and configure them and use them, which is very, very handy. 252 00:10:49,700 --> 00:10:51,600 So this is allows you to do unlocking 253 00:10:51,600 --> 00:10:54,110 and configuring single or cluster devices, 254 00:10:54,110 --> 00:10:56,970 transferring files, launching and managing instances, 255 00:10:56,970 --> 00:11:00,110 so you see two instances running on Snow Family devices, 256 00:11:00,110 --> 00:11:01,780 monitor device metrics, 257 00:11:01,780 --> 00:11:04,640 and launch compatible AWS services on your devices, 258 00:11:04,640 --> 00:11:06,330 for example, EC2 instances, 259 00:11:06,330 --> 00:11:10,000 use DataSync or a Network File System. 260 00:11:10,000 --> 00:11:11,510 So that's it for the Snow family. 261 00:11:11,510 --> 00:11:12,343 I hope you liked it, 262 00:11:12,343 --> 00:11:14,170 and I will see you in the next lecture.