Blocks, Files and Objects, Oh My! Let's talk about Storage Formats

So now that you #ITGirls know what DAS, NAS, and SAN systems are…how is data actually being saved and retrieved? Let's dive into the world of storage formats.

What is a storage format?

Storage Formats are just different ways that storage devices can hold, organize, and present data.

File storage is easy: When you save your work, data is stored as a single piece of information and saved inside a folder. The folders are saved in a hierarchy, and there is a limited amount of metadata (or descriptive tags) that you can use to describe the files folders. This is how NAS Systems store data and is the one you're probably most familiar with!

Block storage is great for larger enterprise companies who need to manage structured data (like big databases or virtual machine files or mission critical applications like email servers!) When something is saved, the data is broken up into fix-sized blocks and given a unique identifier. The ID is helpful because these blocks can be saved across multiple storage devices in your SAN System. This method of storing data can be expensive, but has low latency (aka, retrieves files pretty quickly) and has great data protection. Block storage requires a server to manage how everything is stored.

Object Storage is a little more complex, BUT…you've actually interacted with this all the time if you've ever saved pictures on Facebook or own a Dropbox account! It's cheap, scalable and has incredibly fast performance. Why? When you save your work to an object-based system, the data is broken down into tiny little "objects" along with additional metadata (which includes unique identifiers). All those objects and its metadata are storage pools in a specific order, data retrieval can be managed by a simple HTTP Application Programming Interface (or, API) by searching the metadata. This is a really cost effective type of storage that is typically sold by public cloud providers. This is perfect for unstructured data (like pictures, videos, other random media).

So to retrieve work in a file system, you just need to follow the file's address path (like C:/me/folder/file) in the file system hierarchy. To retrieve work in a block-based system, your server's operating system will look for the unique identifier to the blocks that make up the file. To retrieve work in an object-based system, an API will search through the metadata associated with that work.