Blocks, Objects, & Files oh my..
Where do we put this stuff anyway?
All organizations have a necessity to store their valuable data. A typical first question or confusion amongst those interested in the cloud is knowing the difference between types of storage. We know storage to be one of the fundamental pillars of architecture, but what types are there:
Blocks, Objects, and Files?
Files:
This type storage is useful for when multiple people need to access the same files.
File storage is a type of hierarchical structure where files are organized by the user in folders and subfolders by application type.
This tree-like structure is what differentiates file storage from the others. Making it good for spreadsheets, reports, presentations, backups, archives, anything that may need to maintain a hierarchy when utilized.
This makes it easy to collaborate on projects with and has the ability to lock files. The Files are stored as Network Attached Storage.
Blocks:
Block storage is a type of SAN — Storage Area Network, that presents data as a collection of blocks, which are the basic unit of data in this storage system. Not to be confused with blockchain. (Which are mostly append-only ledgers)
You can think of it as similar to how data is stored on your hard drive or SSD.
There’s another word for SAN as well, it’s called Network Attached Storage — NAS, and the terms are *almost* interchangeable. (For the inspector collectors, A SAN typically uses Fibre Channel connectivity, while NAS typically ties into to the network through a standard Ethernet connection.)
Block storage makes data easier to be updated without needing the entire dataset to be re-written.
Only the block with the information needed is retrieved. Useful for applications that require low latency access to the data so it is typically used to store data that is accessed and modified frequently, such as databases and application data.
Block storage is typically associated with traditional storage (RAID) arrays, and it is accessed through a block-level protocol such as SCSI or Fibre Channel. (Connects directly to your virtual machine.)
Objects:
Object storage is a type of storage that presents data as a collection of objects in a flat environment, each of which contains the data itself, along with a set of metadata that describes the data.
This metadata is a key aspect of what differentiates storage types. Data can be tagged, queried and/or categorized. Essentially giving the data more use-cases.
Allowing better data searchability, scalability, data management and discovery of new data.
Object storage has been increasingly used for IoT devices and environments due to its scalability. You can connect to object storage, like Amazon AWS’ S3 Simple Object storage via an API — Application Programming Interface.
Object storage is also more scalable than the others but there are drawbacks.
It almost seemed too good to be true. Why not use this type of storage for everything.
Due to the data being stored as objects and needing to be completely re-created after each “edit.” Along with being slower, following a BASE model (basically available, just give it a minute) rather than ACID model (instantly available across all databases.) However, Object storage is generally cheaper than Block and File, so as with all technology there are tradeoffs to consider.
Additionally,
There is no stopping the floodgates of new information pouring into data lakes today and the data ocean marketplaces of tomorrow. Data will be transferred at every step of the way.
Organizations must prioritize how this data is compartmentalized and use the best tool for each particular purpose.
The trade-offs between the types of storage will determine the cost and overall efficiency of your enterprise. As new data types arise, organizations will need to be nimble and deterministic to evaluate and optimize their storage needs in an ever evolving environment.
