WEBVTT 00:00.420 --> 00:07.110 Keeping data secure on our individual systems is absolutely critical and it's probably the cornerstone 00:07.110 --> 00:12.900 of all I.T. security there is but let's make sure we understand what we're talking about when we talk 00:12.900 --> 00:14.470 about securing data. 00:14.650 --> 00:21.420 Now in this case what I'm avoiding is permissions based issues I'm not talking about how do we secure 00:21.420 --> 00:26.040 data to keep somebody from editing a file or something like that. 00:26.130 --> 00:29.610 What I'm talking about is just within the system itself. 00:29.700 --> 00:31.680 How do I keep all the data happy. 00:31.680 --> 00:37.530 So what we're really talking about here is number one data integrity of some form or another making 00:37.530 --> 00:43.490 sure that if a hard drive dies or if a power supply dies or something like that I still have that data. 00:43.500 --> 00:44.870 We're also talking about speed. 00:44.880 --> 00:46.840 So we want quick access. 00:46.920 --> 00:51.210 So generically when we're talking about these types of things what we're talking about is what I'm going 00:51.210 --> 00:58.200 to call high availability so we'll save a keeping out the hackers and keep it out the malware for other 00:58.200 --> 00:58.770 episodes. 00:58.770 --> 01:00.760 What we're talking about is the puer system. 01:00.840 --> 01:05.560 What can we do to keep that data happy so let's think about high availability. 01:05.640 --> 01:08.160 So probably the best place to start is RAID. 01:08.190 --> 01:12.330 RAID is a great way to secure data depending on the raid level you use. 01:12.330 --> 01:14.460 You can provide good integrity. 01:14.460 --> 01:20.010 You can provide good speed and the nice part is is Raid 9 times out of 10 is one of the cheapest ways 01:20.010 --> 01:27.060 you can go raid is a fantastic tool but raid is only good for the drives or the storage within a system. 01:27.060 --> 01:31.200 So if you lose a hard drive rate is a great way to take care of these types of issues. 01:31.200 --> 01:39.300 However what happens if I lose a power supply what happens if the entire server explodes. 01:39.300 --> 01:41.360 Now I've got a really big problem. 01:41.370 --> 01:47.550 So what we're talking about securing data we need to stop thinking about just the drive storage and 01:47.550 --> 01:49.740 start thinking about the systems as a whole. 01:49.820 --> 01:55.140 And probably the best most core way we do that is something called clustering clustering is a simple 01:55.140 --> 01:55.870 concept. 01:55.980 --> 01:59.640 So here I have a single computer that might have Raiden it might not. 01:59.720 --> 02:06.480 Dono clustering simply means instead of having one computer doing the job you have two computers or 02:06.480 --> 02:11.810 more doing the exact same job sharing the same database sharing whatever it might be. 02:11.880 --> 02:17.490 The nice part about clustering is that if one of these systems dies the other one can automatically 02:17.490 --> 02:18.400 kick in. 02:18.420 --> 02:24.060 The downside to clustering is that they constantly have to keep each other updated clustering is absolutely 02:24.060 --> 02:29.670 fantastic in fact you can see some pretty interesting setups where two servers that are acting as clustering 02:29.670 --> 02:34.320 servers will literally have their own dedicated little network between them because they have to update 02:34.320 --> 02:35.720 each other so much. 02:35.720 --> 02:42.210 However number one clustering is really expensive to literally not just recreate a couple of extra drives 02:42.210 --> 02:47.360 but to recreate an entire system can add to the price of securing data dramatically. 02:47.370 --> 02:54.270 The downside to clustering also is that you really usually have one primary cluster machine and that 02:54.270 --> 02:57.090 way he's always doing all the work. 02:57.090 --> 03:04.110 You'll have other machines that act more like a backup system so we could really use these systems a 03:04.110 --> 03:05.470 little bit more efficiently. 03:05.490 --> 03:08.280 And when you do that you get load balancing. 03:08.280 --> 03:10.080 Here are my two clustered servers. 03:10.170 --> 03:14.440 Now normally with clustering one is the primary and one is kind of like the backup. 03:14.460 --> 03:19.950 However with load balancing all you're doing is you're distributing the workload across all of these 03:19.950 --> 03:21.330 clustered servers. 03:21.330 --> 03:25.480 That way no one machine is totally overwhelmed. 03:25.680 --> 03:31.070 Clustering and load balancing are absolutely powerful tools but they do tend to be a little pricey. 03:31.080 --> 03:36.950 So one of the things we also like to do for securing data is to simply virtualize the servers. 03:37.050 --> 03:43.730 In that case if a server were to die in one fashion or another we can simply bring it back from a snapshot. 03:43.830 --> 03:49.620 Now just because a server is virtualise don't think that you don't still have the ability to do raid 03:49.620 --> 03:53.870 with virtualise drives or to do clustering with virtual servers. 03:53.970 --> 04:16.530 But really when it comes to the ultimate in securing your data virtualizing this stuff really pays off.