Latest / Nutanix Weekly / Nutanix Weekly: Objects 3.3: Improved Efficiency and Time-to-Value

Transcript

Nutanix Weekly: Objects 3.3: Improved Efficiency and Time-to-Value · Nutanix Weekly

WEBVTT 1 00:00:03.000 --> 00:00:09.809 Andy Whiteside: Everyone welcome to new tannic weekly i'm your host Andy whiteside got Harvey green and direct costs when the rv how's it going. 2 00:00:10.410 --> 00:00:13.080 Harvey Green: I am doing well, just a few pounds bigger. 3 00:00:13.799 --> 00:00:15.809 Andy Whiteside: Few pounds bigger oh man. 4 00:00:16.830 --> 00:00:22.950 Andy Whiteside: i've had such a whirlwind of a day I feel like the holiday was well not yesterday. 5 00:00:24.810 --> 00:00:26.340 Andy Whiteside: Like the long weekend was not yesterday. 6 00:00:26.790 --> 00:00:29.100 Harvey Green: yeah agreed agreed. 7 00:00:29.670 --> 00:00:34.770 Andy Whiteside: and got jarrod Cox with us still in his layer his basement and is just not a basement ISM. 8 00:00:36.120 --> 00:00:37.680 Jirah Cox: database correct now, not a basement. 9 00:00:37.950 --> 00:00:40.320 Andy Whiteside: OK, I can tell from the window that you probably weren't in a day. 10 00:00:42.210 --> 00:00:50.610 Andy Whiteside: Maybe a walkout basement he's got it is a most recent, I guess, maybe new tannic shirt on what's that one got on what's what's under the word new tannic there. 11 00:00:51.000 --> 00:01:00.480 Jirah Cox: Oh, this was for a little internal training team, so we set up some some internal teams all the team leaders got it got a T shirt so through a picture of a superhero on it so. 12 00:01:01.410 --> 00:01:02.970 Andy Whiteside: Is that cam new tactics. 13 00:01:03.420 --> 00:01:04.440 Andy Whiteside: I cam Newton Newton. 14 00:01:06.750 --> 00:01:14.580 Andy Whiteside: For for our listeners it's a picture of a guy with his shirt spread open with the new tannic logo, I was trying to make a joke about cam Newton but it didn't work out. 15 00:01:15.750 --> 00:01:18.630 Jirah Cox: Sorry, I don't I probably, and I fail to get it to yourself, my fault. 16 00:01:19.260 --> 00:01:19.800 Andy Whiteside: that's all right. 17 00:01:20.970 --> 00:01:22.320 Andy Whiteside: You didn't you're not missing much. 18 00:01:25.740 --> 00:01:39.870 Andy Whiteside: gyro and Harvey brought forth a article today from the titanic's blog the name of it is objects three dot three improved efficiency and time to value by phil Gilbert from 29th of November so that's like today. 19 00:01:40.020 --> 00:01:40.650 Jirah Cox: that's today. 20 00:01:41.100 --> 00:01:41.910 Harvey Green: is today. 21 00:01:41.970 --> 00:01:43.200 Jirah Cox: hot off the presses. 22 00:01:43.380 --> 00:01:45.330 Andy Whiteside: off the price do. 23 00:01:45.660 --> 00:01:47.010 Jirah Cox: People even know what presses are anymore. 24 00:01:47.070 --> 00:01:47.520 Jirah Cox: Right like. 25 00:01:47.550 --> 00:01:48.030 Jirah Cox: replace. 26 00:01:48.120 --> 00:02:01.950 Jirah Cox: Our presence right thing, I was at a birthday party for one of my kids friends turning eight or nine and after the after the hat birthday song right like like we all used to as a kid they saying all the extra versus right like and see we do on channel two. 27 00:02:03.030 --> 00:02:03.870 Jirah Cox: And I was like. 28 00:02:03.960 --> 00:02:14.700 Jirah Cox: I asked the parents like to your kids know what channels are because my kids have no clue what channels are they have no idea what these words even mean like what is channel two and they all laughed like we had never thought about it that way. 29 00:02:15.810 --> 00:02:16.320 Harvey Green: yeah. 30 00:02:16.500 --> 00:02:17.070 Jirah Cox: that's. 31 00:02:17.640 --> 00:02:23.910 Jirah Cox: The one remote that goes to Amazon and netflix and plex and that's it yeah. 32 00:02:24.630 --> 00:02:27.030 Harvey Green: that's like trying to discuss downtown with them. 33 00:02:28.050 --> 00:02:28.680 Harvey Green: Was that. 34 00:02:29.070 --> 00:02:36.210 Andy Whiteside: yeah yes know when whenever the cell phones came out they actually had to invent dial tone for the soft phones because, by default, it didn't have it. 35 00:02:36.630 --> 00:02:36.990 yeah. 36 00:02:38.520 --> 00:02:49.920 Jirah Cox: I remember when I worked for my first job when we did our like digital pbx to voice transition and someone told me like yeah the handset is just playing the dial tone like. 37 00:02:50.340 --> 00:02:52.770 Harvey Green: MP3 or wav It is this an MP3. 38 00:02:53.010 --> 00:02:55.710 Jirah Cox: It didn't work i'm like wow that just blew my mind man. 39 00:02:58.890 --> 00:02:59.820 Harvey Green: yeah thanks. 40 00:03:05.520 --> 00:03:06.570 Jirah Cox: hrp so I just. 41 00:03:06.630 --> 00:03:09.300 Harvey Green: I know you're out on wasn't bad enough. 42 00:03:09.300 --> 00:03:16.380 Harvey Green: Right, we had a dial tone MP3 and and so not only dedicated saying what's down there like was an MP3. 43 00:03:17.400 --> 00:03:21.960 Jirah Cox: you're saying you feel like the same age man, so I just i'm maybe i'm more immune to the effects of feeling old. 44 00:03:22.500 --> 00:03:23.580 Harvey Green: yeah you must be. 45 00:03:25.350 --> 00:03:32.280 Andy Whiteside: And i'm sure listeners don't want to hear us talk about it too much, but just think about how much technology has changed in our generation. 46 00:03:34.740 --> 00:03:36.000 Harvey Green: yeah it's crazy. 47 00:03:36.750 --> 00:03:49.020 Jirah Cox: I heard it I heard an interesting study that it was talking about how you know gen X and millennials and all that all that jazz and zoomers like the traditional age based grouping actually loses value. 48 00:03:51.420 --> 00:03:59.130 Jirah Cox: For for whatever not tearing but beginning right and it's actually more pertinent to us. 49 00:03:59.640 --> 00:04:11.160 Jirah Cox: When did you discover the Internet right where you like the 80s generation of Internet like bbs is where you mid 90s with like dial up transitioning to like initial broadband where you like to thousands. 50 00:04:11.610 --> 00:04:20.520 Jirah Cox: And, and that, regardless of age has much stronger correlation with like, how you will relate to your technological peers yeah. 51 00:04:21.150 --> 00:04:25.380 Andy Whiteside: So it's probably timing and then how open, you were to the concepts that came with it. 52 00:04:25.710 --> 00:04:33.150 Jirah Cox: yeah yeah right if you can remember, like the Bulletin board days right you're a different generation of Internet users and like MySpace days. 53 00:04:33.330 --> 00:04:33.630 yeah. 54 00:04:35.310 --> 00:04:37.830 Andy Whiteside: Well, and question for you guys, do you. 55 00:04:38.340 --> 00:04:50.040 Andy Whiteside: Like did you have some point feel like you were going to be always an adopter whatever the latest greatest thing was or have you got to the point where you're like I don't want anymore, I like it where it is right now and I don't want it to change. 56 00:04:56.250 --> 00:05:08.130 Harvey Green: that's a hard question, I mean I I don't as a technologist I don't think I ever want to feel like number two, but there are times that I feel like number. 57 00:05:10.020 --> 00:05:10.650 Andy Whiteside: I think it was a. 58 00:05:11.100 --> 00:05:13.170 Jirah Cox: there's a few areas of my life, where I. 59 00:05:14.460 --> 00:05:22.590 Jirah Cox: I use a bit less like chewing gum and twine right like I no longer flash my my Internet router right, and you know. 60 00:05:23.550 --> 00:05:34.650 Jirah Cox: there's certain things that I need stability from but I don't want to hear about from like my family members at two in the afternoon about like hey this thing crashed, you know so encouraged me to do more robust engineering with the technology in my life. 61 00:05:36.780 --> 00:05:52.170 Harvey Green: I can see that, I mean it i'm I am sort of sort of that way for well i'm definitely that way for other people I am sort of the opposite, for me, my I was telling you guys about right before we jumped on to record. 62 00:05:53.340 --> 00:06:13.440 Harvey Green: I I had a few little issues, I was dealing with windows 11, and so I just wiped it and started over and I did that you know very much within the course of a day, and here I am back on that same machine and I don't have a little issues and so that that was a good exercise but. 63 00:06:14.580 --> 00:06:18.540 Harvey Green: I kind of pride myself on being able to do stuff like that. 64 00:06:19.290 --> 00:06:29.010 Andy Whiteside: But today standards, this is a great example, what does wipe it mean did you actually format, the hard drive and put the os back down or did you hit the reset my machine but. 65 00:06:31.680 --> 00:06:40.740 Harvey Green: So I did hit the reset my machine button, but I did tell it to wipe all the settings and files and all that. 66 00:06:40.890 --> 00:06:49.470 Harvey Green: yeah um but yes, that I do come from the day, where you did have to go break out the the the CD or set of CDs. 67 00:06:50.010 --> 00:06:50.370 Harvey Green: But to. 68 00:06:50.490 --> 00:06:55.710 Jirah Cox: Do that you know to do it from the CDS do you trust the whole reset system, I have a hard time. 69 00:06:56.730 --> 00:07:02.640 Harvey Green: Well, so you no longer have a choice because you don't get a CD. 70 00:07:04.260 --> 00:07:04.560 Harvey Green: No. 71 00:07:05.220 --> 00:07:09.720 Andy Whiteside: But anyway, the convenience versus the cleanliness of it. 72 00:07:09.780 --> 00:07:10.830 Andy Whiteside: Would I rather just. 73 00:07:11.130 --> 00:07:26.610 Andy Whiteside: wipe out the at least the partition table and reinstall my ios yes Is there something really nice about hitting the reset delete everything except the core files button and having a backup and somewhat functional and 30 minutes that's it's hard to. 74 00:07:27.150 --> 00:07:30.300 Andy Whiteside: it's hard to resist that level of convenience and you think it'll be good enough. 75 00:07:30.510 --> 00:07:34.080 Harvey Green: And and Jared to your point I did have it go and download a fresh. 76 00:07:35.400 --> 00:07:35.940 Jirah Cox: There you go. 77 00:07:36.000 --> 00:07:49.860 Jirah Cox: that's that's what I mean yeah i'm all MAC right, but like I can like wipe the hard drive and at the bios i'm looking go download a fresh I was from the Internet it's like a guarantee it's totally wipe that machine couldn't boot totally fresh I was laid back down and. 78 00:07:49.920 --> 00:07:50.940 Jirah Cox: that'll that'll trust. 79 00:07:51.300 --> 00:07:52.830 Andy Whiteside: yeah yeah. 80 00:07:54.990 --> 00:07:55.470 Andy Whiteside: All right. 81 00:07:55.800 --> 00:07:56.670 Jirah Cox: yeah we're off topic. 82 00:07:56.790 --> 00:07:58.620 Harvey Green: right way, more often. 83 00:08:00.330 --> 00:08:10.950 Andy Whiteside: You know I will finish my part of that topic by saying this every time there's new evolutions and technology, the people around me some of them in my household it's like starting over. 84 00:08:11.430 --> 00:08:18.900 Andy Whiteside: And so maybe I can adopt the newest latest greatest technology or maybe the whole new concept that comes with it. 85 00:08:19.380 --> 00:08:33.300 Andy Whiteside: it's just now i'm having to revolve those around me and sometimes I wish it would just stay still at least long enough for me to get some some momentum out of what happened before so then i'm not now. 86 00:08:34.680 --> 00:08:39.510 Andy Whiteside: Trying to re educate totally different concept that replaced previous concept. 87 00:08:40.740 --> 00:08:42.720 Harvey Green: yeah agreed. 88 00:08:43.890 --> 00:08:57.420 Harvey Green: i'm yeah i'm kind of saying weighing by can say, you know as soon as I have access to those things, I will try them, and I mean like the windows 11 example if I don't like what's happening i'll wipe it. 89 00:08:57.480 --> 00:09:00.570 Harvey Green: And i'll try again and if I don't like it by the end of that then. 90 00:09:01.170 --> 00:09:02.280 Harvey Green: I move on with something. 91 00:09:05.160 --> 00:09:08.730 Jirah Cox: yeah speaking of new features, you can get just by installing updates. 92 00:09:09.900 --> 00:09:19.320 Andy Whiteside: yeah let's talk about objects three dot three improved efficiency and time devalue in new tannic so jarrod can you re educate, some of us on what objects in general is. 93 00:09:19.740 --> 00:09:32.820 Jirah Cox: Totally so objects, is a is the name of our functionality for on Prem s3 storage s3 storage, you know invented by Amazon I don't know 1015 years ago something like that back when the when the earth was young. 94 00:09:33.870 --> 00:09:48.900 Jirah Cox: For you know, storing large datasets at scale with web native technologies right, so the way you know, an fs rights over the network as a network file system, the way SMB rights over you know to a window share. 95 00:09:50.310 --> 00:09:53.580 Jirah Cox: To like a windows file server that your client speaks. 96 00:09:56.160 --> 00:10:07.410 Jirah Cox: The web age right, where I can put it in place via like an API request and do like an http upload download and then my client can go, you know download it right, so I can store like website images that way. 97 00:10:09.210 --> 00:10:10.860 Jirah Cox: For backup data that way, I could store. 98 00:10:12.150 --> 00:10:23.100 Jirah Cox: You know its own its own file system right, so I could store anything records log data, you know 10s of thousands of files whatever I want vast scale available to me. 99 00:10:24.240 --> 00:10:26.280 Jirah Cox: But it's a it's fundamentally for that kind of a. 100 00:10:27.780 --> 00:10:30.630 Jirah Cox: web based you know upload download kind of format. 101 00:10:31.890 --> 00:10:42.750 Jirah Cox: And so, so with this three three right which you can get if you're running objects now just by go running your lifecycle manager to pull down the latest version and automate that upgrade for you. 102 00:10:43.860 --> 00:10:54.660 Jirah Cox: Some updates here for you, that you'll get are you you've already had an object three two, I think it was the ability to tear out older data to aws. 103 00:10:55.170 --> 00:11:07.080 Jirah Cox: So you can have a data store on Prem keep only the most recent data send older data out to public cloud if you wanted to with this release we also add azure that I compatibility, so now, you can send it up to. 104 00:11:08.580 --> 00:11:11.790 Jirah Cox: To s3 and aws or azure blob up an azure. 105 00:11:12.750 --> 00:11:23.580 Andy Whiteside: So gyro I was with a client this morning it's got a lot of video please camera video yeah they can use this technology to store it in mass and when it gets to a certain point where they got to keep it even longer. 106 00:11:23.880 --> 00:11:34.110 Andy Whiteside: And they run out of room where it's more cost effective put in the cloud, this is going to allow them to aggregate it or push it up to the cloud and they've got the best of both the on premises part of storing it. 107 00:11:34.440 --> 00:11:39.300 Andy Whiteside: As well as the scalability of the cloud to store it in the same format. 108 00:11:39.690 --> 00:11:53.940 Jirah Cox: Totally yeah and and video footage can be a fantastic use case for that right because it's large it's already compressed right we we often care about it's often you know written once and like read rarely right. 109 00:11:55.020 --> 00:11:58.170 Jirah Cox: Only if you're scanning back you know to scan for motion or something like that. 110 00:11:58.590 --> 00:12:07.530 Jirah Cox: And so yeah usually that's a data type that seeks for like the most cost effective, you know lowest cost per gig type of functionality and so yeah objects. 111 00:12:07.890 --> 00:12:14.100 Jirah Cox: s3 storage is a fantastic you could use case for that, and if I had very, very long retention requirements yeah I would want to make sure I had. 112 00:12:15.510 --> 00:12:16.530 Jirah Cox: You know the. 113 00:12:17.670 --> 00:12:22.980 Jirah Cox: The most optimized place to put that right now don't count objects out for even that whole workload right like. 114 00:12:23.550 --> 00:12:35.760 Jirah Cox: You know, get your own to study, because it's actually even cost competitive against you, using public cloud, but you know use your own numbers right and do your own study, make sure that you're using the most effective storage location for that. 115 00:12:36.720 --> 00:12:43.710 Andy Whiteside: Well, and that was the complaint this morning in the meeting is that you know sticking it out and aws solved, a lot of problems but it's still got expensive over time. 116 00:12:44.340 --> 00:12:52.830 Jirah Cox: Totally right and then also you know all the traditional stuff around using public cloud right cost of access for retrieval transit in and out, you know. 117 00:12:53.940 --> 00:13:01.590 Jirah Cox: And then of course it's just it's it can be latency right if you need to scan for it remotely compared to it could be, you know, in the next room if it ran on mechanics yeah. 118 00:13:02.400 --> 00:13:19.050 Andy Whiteside: And dumb question for me how does this change, I mean Do you still back this stuff up the replicated somewhere else, or is it is somewhat resistant resilient where its resting here, then, do you maybe replicate that to another new tactics cluster somewhere else is that common so it. 119 00:13:19.170 --> 00:13:27.990 Jirah Cox: yeah so it's actually super resilient and a lot of ways right not only like all of the foundational goodness of ios right with the rapid rebuilds software defined storage. 120 00:13:29.250 --> 00:13:32.730 Jirah Cox: scale out everything that we do for all the data we ever host. 121 00:13:33.210 --> 00:13:39.750 Jirah Cox: On top of that s3 itself supports stuff like the warm functionality so right once read many and legal compliance. 122 00:13:39.990 --> 00:13:47.550 Jirah Cox: So with with s3 data, you can even get to a point on an engine X cluster where you know that data really truly is immutable. 123 00:13:47.790 --> 00:13:55.830 Jirah Cox: No one can delete it before it hits its retention lock, of whatever you want three years, seven years, whatever you're you're mandated for there so. 124 00:13:56.100 --> 00:14:06.720 Jirah Cox: Usually, if you need a multi site availability that's going to be your application right you don't typically back it up and then restore it somewhere else, but you can just replicate it or write it to both places if you wanted to yeah. 125 00:14:07.920 --> 00:14:16.530 Andy Whiteside: So Harvey this first announcement is about not only being integrated with Amazon, but also integrated with azure any additional points you think we ought to make there. 126 00:14:17.760 --> 00:14:25.230 Harvey Green: yeah I mean this is just the tenants continuing their journey right this you know a lot of features that were set up. 127 00:14:25.710 --> 00:14:31.530 Harvey Green: With Amazon and we're ready, and then there were some things and azure that lag behind some. 128 00:14:32.040 --> 00:14:43.500 Harvey Green: Because the integration needed to be written there and some because azure didn't have all of the same feature parity with what Amazon had in place already so. 129 00:14:44.370 --> 00:14:55.170 Harvey Green: glad that everyone is getting back together and working again, and this is, you know some fruits of that Labor so that you have options available to you. 130 00:14:56.670 --> 00:14:59.160 Jirah Cox: What I think you don't forget like depending how you want to use it. 131 00:14:59.670 --> 00:15:06.960 Jirah Cox: It can even be your backup storage location right so like think about that if you're using something as your vm target is your haiku target. 132 00:15:07.410 --> 00:15:15.330 Jirah Cox: Would you think of that as needing its own backup right probably not you would simply have a second copy of your backups if you wanted like an offsite copy of that data. 133 00:15:15.600 --> 00:15:15.930 Right. 134 00:15:19.050 --> 00:15:25.440 Andy Whiteside: um yeah it's it's what we talked about several the recent podcasts where you have best of breed new Technics. 135 00:15:26.250 --> 00:15:40.800 Andy Whiteside: This case objects now able to talk to the public cloud infrastructure, which has the scalability that most of us long, for I take those two things and mix them together now you've got best solution for customer and question yeah. 136 00:15:42.780 --> 00:15:48.600 Andy Whiteside: um next section talks about improved time to value for analytics diver what's this announcement related to. 137 00:15:49.170 --> 00:15:55.560 Jirah Cox: So the the the short version is supporting that s3 select protocol, but what that what that really means is. 138 00:15:57.000 --> 00:16:07.050 Jirah Cox: In in my super layman's mind is like I can request parts of a file right, so if the data this blog article says semi structured I think it's a good way to go about it think about it, if it's like a csv file. 139 00:16:07.440 --> 00:16:10.320 Jirah Cox: Rather than, say I want that csv file back like customer records. 140 00:16:10.830 --> 00:16:16.890 Jirah Cox: This is saying I can make an API requests or a call that says send me back just the rows in that spreadsheet where. 141 00:16:17.190 --> 00:16:29.970 Jirah Cox: The customer country equals USA right or something like that I can get back a subset of the data which is even within perhaps a single file rather than have to pull down the entire file just something like what I want. 142 00:16:31.650 --> 00:16:39.570 Jirah Cox: Since we support stuff like spunk smart store like sending all my analytics and logs and security data to an s3 storage location. 143 00:16:40.020 --> 00:16:56.520 Jirah Cox: Well now, I could say just something back the logs from that one client or something like you know website hits matching this source IP address and just give me that back faster so faster retrieval of data without unless reversal as well, like less data has to come across the wire. 144 00:16:57.780 --> 00:17:08.790 Andy Whiteside: And that's all doable through api's does new tannic provide the front end for this or that's up to you to the application provider you to write the front end for the API integration. 145 00:17:09.540 --> 00:17:11.040 Jirah Cox: So I would think of it more like. 146 00:17:12.180 --> 00:17:18.750 Jirah Cox: You know the same way, like you would deploy files on new tonics as a as a nas right to be a shared file storage. 147 00:17:19.470 --> 00:17:27.420 Jirah Cox: We rarely are new tannic is almost never the client for that right we're not counting those file shares and sending our data to it that's for like your workstations to do. 148 00:17:28.080 --> 00:17:36.150 Jirah Cox: same thing here like an s3 storage is for your applications talk to that are sending and receiving that s3 data so they'll usually be the one to be a body cam data. 149 00:17:36.600 --> 00:17:47.940 Jirah Cox: You know, writing that data in a structured format, now that client, which is an s3 supported client can now pull back just the data it wants to from like camera 123 or whatever right. 150 00:17:49.080 --> 00:17:52.200 Andy Whiteside: Yes, a super time me, we were having some of these conversations this morning and. 151 00:17:53.220 --> 00:18:00.750 Andy Whiteside: we're talking about what new tannic brings to the mix above and beyond hyper converge in terms of solving problems for people yeah. 152 00:18:02.040 --> 00:18:06.960 Andy Whiteside: Poverty he thinks, you will come in about on this section, the improved time devalue peace. 153 00:18:08.610 --> 00:18:20.850 Harvey Green: No, I mean, I think you guys hit it pretty well, this is just better, faster by doing using the technology that have less come across and be more specific around what you're getting. 154 00:18:22.800 --> 00:18:33.120 Andy Whiteside: And jarrod the third of the three points made in this article is the ingest large volumes of data efficiently help us understand that that piece. 155 00:18:33.990 --> 00:18:43.470 Jirah Cox: yeah so this isn't is talking about time to move data down to our ratio coded tier so ratio coding. 156 00:18:44.790 --> 00:18:52.110 Jirah Cox: Unlike traditional workloads right we're will simply write two copies with three copies rf to rf three in our parlance of storing data. 157 00:18:53.040 --> 00:18:59.640 Jirah Cox: We also work rich coding right where if it's cold data it's you know, especially if only been written written once and isn't going to be updated. 158 00:19:00.120 --> 00:19:12.420 Jirah Cox: Then will erase your code at right, we can actually store less than two full copies, but still be able to recreate the data if we had say a disk failure or node failure so there's no risk added to the data, but we can also give you back more usable space. 159 00:19:13.620 --> 00:19:20.190 Jirah Cox: Since objects, is one of our in theory like largest scalable workloads right, you can get to like a multi petabyte objects workload quite easily. 160 00:19:21.690 --> 00:19:30.060 Jirah Cox: Then, if I needed to rf to that full data that's a lot of capacity, if I can sex it faster, if I can raise your code at faster. 161 00:19:30.480 --> 00:19:41.520 Jirah Cox: than I get more usable space back on the cluster right, so this talks about doing it seven times faster usually that that sex threshold is commonly one week right so after it's been one week written cold. 162 00:19:41.910 --> 00:19:52.560 Jirah Cox: Then we'll move from rf to down to a ratio coded seven X faster, you can imagine, this sounds to me like I think it's happening after about one business day what one one calendar to actually write 24 hours so. 163 00:19:53.400 --> 00:20:00.330 Jirah Cox: Doing that faster means I now have six more days worth of capacity of ingest right whatever my rate is right if i'm doing 100 terabytes a day. 164 00:20:00.960 --> 00:20:10.230 Jirah Cox: that's 600 terabytes usable that I just got back from the cluster so erasure coding I think of it as like a hipster rate almost it's. 165 00:20:11.010 --> 00:20:21.030 Jirah Cox: Not disk based it's it's data block based right so rather than rather than saying it's this 1234 and five is all parody or even distributed parity. 166 00:20:21.450 --> 00:20:36.510 Jirah Cox: it's more like I wrote pieces of data, a B, C D and then I wrote a fifth piece of data data, you know he that is enough to reconstruct any of those missing pieces if I needed to, but I can do that at like the one megabyte level, or whatever you want to do, for your regularity. 167 00:20:38.640 --> 00:20:41.010 Andy Whiteside: Alright, so pronounce it for me again. 168 00:20:41.130 --> 00:20:41.910 Jirah Cox: But straight. 169 00:20:42.360 --> 00:20:44.970 Andy Whiteside: razor coding eraser coding. 170 00:20:45.120 --> 00:20:46.800 Jirah Cox: Your race your coding yep. 171 00:20:47.100 --> 00:20:53.040 Andy Whiteside: You your race your coding yep I have never understood what that was and all the time. 172 00:20:53.340 --> 00:20:53.670 Jirah Cox: yeah. 173 00:20:53.760 --> 00:20:55.860 Andy Whiteside: it's a real fast and move on, nobody ever asked. 174 00:20:56.460 --> 00:21:10.920 Jirah Cox: So so because right because Member rate is that we're done didn't array of independent discs yeah and so erasure coding doesn't care about the underlying dis geometry it occurs at the data level right so it's saying, if I right. 175 00:21:12.030 --> 00:21:24.990 Jirah Cox: Did you use a terrible analogy that probably will make every actual engineer on the call like cry if I wrote like word document 1234 and then word document five with like a bit wise math it's still you know, a different thing. 176 00:21:26.400 --> 00:21:30.060 Jirah Cox: What is it it's not a binary X or i'm sorry don't don't don't email me. 177 00:21:31.200 --> 00:21:36.240 Jirah Cox: You know math so I can reconstruct the missing piece of data, but it's occurring at the data level, not the disk level. 178 00:21:36.810 --> 00:21:37.110 yeah. 179 00:21:38.610 --> 00:21:48.270 Andy Whiteside: it's enough like a synopsis of what was written enough to put it back if necessary, is it identical to what needs to go back or is it just enough to get the general idea. 180 00:21:48.840 --> 00:21:51.630 Jirah Cox: So it's also if this is all in the tanks Bible. 181 00:21:52.740 --> 00:21:57.480 Jirah Cox: Will even adjust the erasure coding usually will say like strip size versus. 182 00:21:58.560 --> 00:22:05.250 Jirah Cox: With raid you know there were all the the hard coded we memorize like raid zero Grade one grade five grade 10th grade six. 183 00:22:07.230 --> 00:22:13.590 Jirah Cox: With a ratio coding typically people refer to like strip sizes of like a two plus one, a four plus one strip size. 184 00:22:14.010 --> 00:22:29.250 Jirah Cox: And that refers to if I do a ray of ideal ratio coding four plus one and it's the same thing it's me for actual data bits and then a fifth parody bit right, but I would use to reconstruct any one of those missing bits so 20% overhead for my. 185 00:22:31.290 --> 00:22:31.560 Jirah Cox: You know. 186 00:22:31.890 --> 00:22:35.670 Jirah Cox: redundancy redundancy data right the parody data dirty data yeah. 187 00:22:37.290 --> 00:22:40.560 Andy Whiteside: Okay, well, no that was that was worth the price of admission right there. 188 00:22:41.130 --> 00:22:41.520 Jirah Cox: There you go. 189 00:22:44.250 --> 00:22:46.170 Andy Whiteside: Harvey thoughts comments. 190 00:22:48.420 --> 00:22:51.090 Harvey Green: I said he explained it well. 191 00:22:52.440 --> 00:23:06.240 Harvey Green: And how it works in my head is kind of how you explained it is probably a little different in my head, but i'll leave it at how he explained it, it is just not at the disk level is that the data level so it's. 192 00:23:07.410 --> 00:23:18.090 Harvey Green: Much much closer and you're much better off in a failure situations and she don't have to have all the necessary disk in order to make that work. 193 00:23:18.960 --> 00:23:34.830 Andy Whiteside: And all of this around objects as a way to conveniently store a massive amount of data and be able to access it and move it around and not over consume the disk subsystem and that's basically objects, in a nutshell, and it just keeps getting better. 194 00:23:35.730 --> 00:23:38.760 Jirah Cox: Totally yeah that's that's a fantastic summary yeah. 195 00:23:39.930 --> 00:23:41.550 Andy Whiteside: Fantastic summary course. 196 00:23:42.690 --> 00:23:44.010 Andy Whiteside: Thank you for helping me understand that. 197 00:23:44.400 --> 00:23:44.880 Jirah Cox: Of course man. 198 00:23:46.470 --> 00:23:54.300 Andy Whiteside: Well guys I guess that's it I think we've covered this blog gyro any additional topics or comments or concepts, you want to bring up. 199 00:23:57.000 --> 00:23:58.800 Jirah Cox: That that'll do it for me. 200 00:24:00.990 --> 00:24:04.110 Jirah Cox: As always, you can you can kick the tires on objects on the next test drive. 201 00:24:05.940 --> 00:24:13.320 Andy Whiteside: yeah we should always bring that up the ability to do a test drive with new tactics and you're just that functionality, as a way to get engaged and. 202 00:24:13.530 --> 00:24:27.420 Andy Whiteside: learn how the technologies work firsthand that's it's amazing that we have that we've had a whole podcast on that just topic alone, but the fact that new tactics brings near real time lab actions for their customers and partners is extremely valuable. 203 00:24:28.560 --> 00:24:32.460 Andy Whiteside: Speaking of labs have you got a workshop coming up sometime soon right. 204 00:24:32.670 --> 00:24:35.460 Harvey Green: I do, it is two days away. 205 00:24:38.100 --> 00:24:41.700 Andy Whiteside: Registration daya again, that was a client this morning i'll make sure that they get to. 206 00:24:42.210 --> 00:24:43.530 Andy Whiteside: Get invited to that one as well. 207 00:24:43.950 --> 00:24:52.770 Harvey Green: yeah absolutely um you know that I think, as I say, think I think this one might be our last one for 2021. 208 00:24:53.910 --> 00:25:03.480 Harvey Green: But we'll see there was definitely lots of demand for this, so I don't think will fit another one here but we'll see what happens after tomorrow. 209 00:25:04.830 --> 00:25:11.460 Andy Whiteside: yep well guys, I appreciate it thanks for jumping on on the Monday after thanksgiving and look forward to what you guys bring next week to talk about. 210 00:25:11.760 --> 00:25:12.630 Jirah Cox: cool talk to them. 211 00:25:13.620 --> 00:25:14.190 Andy Whiteside: Thanks guys.