Learning Library

← Back to Library

Object Storage: Core Components Explained

Key Points

  • Object storage provides low‑cost, low‑performance storage optimized for internet workloads like web apps, content delivery, and long‑term archival that traditionally relied on tape.
  • An “object” is any file that includes four essential parts: a unique identifier (ID), the data itself, metadata describing the file (e.g., creator, type, size), and attributes that control access and actions on the object.
  • Metadata makes objects searchable and indexable, while attributes define permissions such as who can read, overwrite, download, or delete the object.
  • Objects are organized in flat containers called buckets, where all unstructured data is placed for easy retrieval and management.

Full Transcript

# Object Storage: Core Components Explained **Source:** [https://www.youtube.com/watch?v=ZfTOQJlLsAs](https://www.youtube.com/watch?v=ZfTOQJlLsAs) **Duration:** 00:10:29 ## Summary - Object storage provides low‑cost, low‑performance storage optimized for internet workloads like web apps, content delivery, and long‑term archival that traditionally relied on tape. - An “object” is any file that includes four essential parts: a unique identifier (ID), the data itself, metadata describing the file (e.g., creator, type, size), and attributes that control access and actions on the object. - Metadata makes objects searchable and indexable, while attributes define permissions such as who can read, overwrite, download, or delete the object. - Objects are organized in flat containers called buckets, where all unstructured data is placed for easy retrieval and management. ## Sections - [00:00:00](https://www.youtube.com/watch?v=ZfTOQJlLsAs&t=0s) **Object Storage Basics and Use Cases** - An overview of low‑cost object storage for web workloads and long‑term archiving, explaining what objects are and their essential components like unique IDs. - [00:06:18](https://www.youtube.com/watch?v=ZfTOQJlLsAs&t=378s) **Cold‑ish Object Storage Use Cases** - The speaker outlines pricing tiers for cold object storage, introduces a “cold‑ish” tier for regularly accessed website assets, and highlights API‑driven benefits like global replication for video streaming. ## Full Transcript
0:00if we think about object storage which 0:02is kind of a newish tier 0:05object storage the idea is for 0:07relatively low performance storage at a 0:10relatively low cost that is designed to 0:13serve the needs of the internet workload 0:17and so by internet workloads i'm talking 0:19about web applications i'm talking about 0:21websites and website hosting and 0:23delivering content across the internet 0:26but also a way to find a new home for 0:30all of that data that we used to store 0:32on tape right we have to have a way to 0:34take big volumes of data write that down 0:37put it somewhere that is safe that is 0:39secure and where we can hold on to it 0:42for a long time for regulatory reasons 0:44legal reasons and other 0:47requirements that make it necessary for 0:50us to keep that data for a long amount 0:51of time so let's start off with what is 0:55an object and why do we call it object 0:57storage so 0:58the concept in object storage is that 1:01you have an object right now 1:04that object can be any kind of file 1:07there aren't really any file 1:09restrictions or anything like that there 1:11are some characteristics that are very 1:13important but as a general rule any kind 1:16of file can be an object so 1:20we'll throw a little object in there now 1:22there are four essential components to 1:25that object that make it usable in the 1:27computing sense right the first one is 1:30for every object we have to have an id 1:32we have to have some sort of a unique 1:34identifier that lets us know that what 1:37this object is when it comes time to 1:39retrieve it that's the first piece the 1:41second piece is obviously you've got to 1:44have some data 1:45you know what's the point of having a 1:47file and all of this information about 1:48it if you don't have the data itself so 1:50this could be an excel file it could be 1:52a video file it could be an audio file 1:55any kind of file but you've got to have 1:56the actual data itself 1:59the third piece 2:00that you have to have is 2:02metadata so what is metadata metadata is 2:06everything that you need to know 2:09about this file and about the data 2:12itself who created it when was it 2:14created 2:15what is it used for what is the file 2:17type how large is it the metadata is 2:20what makes it possible to go out and 2:22find it search it index it so that you 2:25can bring it up and use it whenever you 2:27need to because again all that's being 2:29displayed out right now is the id the id 2:31is not terribly useful and then the last 2:33thing that you have to have 2:35are your attributes now attributes are 2:37related to metadata they aren't exactly 2:39the same thing but they're close 2:41attributes can be are certain users 2:44allowed to override it are certain users 2:46allowed to download it are certain users 2:48allowed to delete it so the attributes 2:50are about the object itself rather than 2:53about the data 2:54but again those are those four essential 2:56components this is what you use 2:59when you actually create something so 3:01you've got your object here right and 3:04then what do you do with it well in the 3:06world of object storage which is 3:08unstructured you take all of your 3:10objects and you drop them into a bucket 3:13so 3:14let's do this on this side so imagine i 3:17have 3:18a bucket 3:21and my bucket can have a handle why not 3:24and so into my bucket i'm going to put 3:27all of these objects and now this is the 3:29neat part these buckets can be as big as 3:32you need them to be they can scale to 3:34hold billions of objects now it's not 3:37always a good idea to put billions of 3:38objects in it but they can in fact scale 3:41to that size 3:42and then what's really neat about object 3:44storage because remember we want to be 3:45able to use this for cold storage too 3:47right we need to put things in it that 3:49can never be deleted 3:50is whenever you the user you're going to 3:53interact with this thing by an api right 3:56so you're not ever going to interact 3:58with the buckets individually here we'll 4:00throw you in 4:01there you are 4:02you're using object storage it's great 4:05so 4:05you're going to take your objects you're 4:07going to drop them into this bucket now 4:09on the back end the magic of object 4:11storage 4:12is your object here call it o 4:16it is going to replicate out of this 4:18bucket the bucket is a virtual construct 4:21but we are going to take it and we are 4:23going to say all right well 4:25i have 4:26three physically separated devices 4:31and i'm going to take a copy of your 4:33object 4:34and i'm going to replicate it out into 4:36all three of them 4:37and my purpose in replicating them out 4:40into all three of those locations is 4:42data integrity data security that way if 4:45the physical device here right each of 4:47these buckets is going to reside on a 4:48physical device so we'll call this p1 4:51p2 4:53and p3 4:55so we're going to replicate it out your 4:57object is going to be on three physical 4:59different devices so if there's a 5:00hardware problem or a network outage or 5:03whatever in p1 that's okay you've still 5:06got copies in p2 and p3 you can still 5:08get the data that you need 5:10now when we get into the details of 5:13object storage all of a sudden you start 5:15saying man there's a lot of complexity 5:17in this right so not only do i have 5:20the object buckets but the way that it's 5:23priced and metered and billed is kind of 5:25unique and people are asking me about 5:27how many gets and how many puts and how 5:28much storage am i taking up 5:31and the metering and building around 5:33object storage everyone that sells it 5:35has their own different metrics on it 5:36but here's the important parts to 5:38consider 5:39the first is what is your responsiveness 5:42or your response time right so from a 5:44price perspective the higher performance 5:46that you want and this is true for all 5:48cloud storage but the more performance 5:50you want 5:51the more it's going to cost if you have 5:53data 5:54that you have packaged up and it's data 5:56that for 20 years your company's been 5:58taking it writing it to tape sending 6:01that tape to vaults and storing it in 6:02vaults where it will probably never be 6:04seen again that can go into the very 6:06coldest of cold tears right that is 6:08going to be the least expensive 6:11and so we're going to call that cold 6:14times 10. 6:15that is the coldest of cold tears 6:18up from that where you need to get to it 6:20every so often but you don't really need 6:22to get to it much that's gonna be a two 6:24dollar sign one right we're gonna we're 6:26gonna charge a little bit more for that 6:28we're just gonna call that cold 6:30maybe this is data that you need to get 6:31to once every six months something like 6:33that 6:35but maybe you're earning a website and 6:37you are using these object storage 6:39buckets to host all of the assets for 6:41your website obviously that is going to 6:44cost a little bit more right because you 6:46are sending data out on a regular basis 6:48so that's going to get our three dollar 6:50signs 6:52and 6:53i'm going to call this one cold-ish 6:56it's not really warm because the concept 6:59of a warm tier doesn't really exist in 7:01object storage but it's cold-ish 7:03and so this is videos it is pictures it 7:08is physical files um and what makes 7:11object storage really neat is on the 7:13back end through the magic because 7:14remember you you're interacting via an 7:16api you're not interacting with these 7:18things directly 7:20you can do all kinds of nifty front-end 7:23things for it and that's the really neat 7:25part of object storage and that's the 7:27benefits of it so let's talk about some 7:29of them right um hosting videos for 7:31streaming over the internet perfect 7:33perfect use case for object storage 7:35because like i said before we can 7:37replicate these things out right so for 7:39your video streaming 7:41rather than just replicating it into 7:42three buckets that are all in the same 7:44place let's replicate it out globally 7:46let's replicate it out so that we are 7:47hosting those videos 7:49not just in their primary location but 7:51in three different locations 7:53in seven different countries around the 7:55world decreasing the latency for your 7:57users so that their streaming experience 7:59is better 8:00let's talk about another good use case 8:02um 8:03uh file sharing so let's imagine a 8:06company where you're collaborating on 8:08spreadsheets and powerpoints and things 8:09like that and you've got collaborators 8:10that work all over the world 8:13well 8:14for versioning purposes 8:16you can have different versions that 8:18overwrite the previous version as long 8:20as those attributes allow for it and so 8:23if i've got my friend eric who is 8:25working in london can create a file 8:28i can get a copy of that file make my 8:31changes to it upload it back in and i've 8:33got version two that's the metadata 8:35and then i can have a colleague in 8:38singapore sam who downloads that file 8:40makes some more changes to it uploads it 8:42back in version three again incrementing 8:44the metadata 8:45but that allows for this cross-globe 8:47collaboration on a single file type and 8:50it makes everybody's lives easier but 8:52we're doing it at very low cost much 8:54less expensive than trying to 8:56send files around via email because 8:58emails have to get stored in a much more 9:00expensive storage tier much less 9:02expensive than fedexing thumb drives 9:04around and again 9:06being able to be done in a secure manner 9:08because we have control over who can see 9:10and access these objects 9:13um other web hosting obviously is going 9:15to be important we talked about 9:16regulatory data before the regulatory 9:18data and cold storage 9:21let's say that you're involved in a 9:23digital archiving project and you're 9:26working with a university or library and 9:28you're taking digital photographs of 9:30manuscript documents and you want to 9:32host those out for everyone in the world 9:34to be able to see object storage perfect 9:36place to store that kind of data because 9:39again it's not getting accessed very 9:40often 9:41it is 9:43right once and then read many many many 9:46times and you can host it in a secure 9:48manner 9:48so hopefully you've enjoyed this this 9:50has been a good overview of object 9:52storage as always if you have any 9:54questions please feel free to leave them 9:55in the comments i'll do my best to 9:57answer some of my colleagues at ibm well 9:59and if you have any other questions just 10:01let us know 10:03thank you so much for your time today if 10:05you have any questions please drop us a 10:07line below if you want to see more 10:09videos like this in the future please do 10:11like and subscribe and let us know 10:14and don't forget you can grow your 10:15skills and earn a badge with ibm cloud 10:17labs which are free browser-based 10:20interactive kubernetes labs that you can 10:23find more information on by looking 10:24below thanks again