File, Block, and Object Storage
In this section, we explore the different concepts and building blocks of the storage software stack. There are several different parts to consider in building a stack. Starting with the most basic, there are file systems and volume managers. Then there are more advanced file systems like XFS. Finally, there are fully featured file systems that incorporate storage functionality and act as software defined storage platform like Ceph.
Sometimes the functions from these products overlap, while others play nice together. Understanding the different tools, capabilities, and formats should ease the process of selecting the ideal stack. We’ll start by explaining the storage formats, then file systems, and conclude with distributed storage software.
The three types of storage available for on-premise use or the cloud are file, block, and object. Each one is a format that stores, organizes, and presents data in a different way.
- Structured and unstructured data
- Used in DAS and NAS
- Hierarchical model and folder structure
- Similar to File Explorer
- The entire file is stored in a folder
- Users access files over Ethernet LAN
- Store CAD designs, videos, images, files, documents, etc.
- Limited metadata
- Structured data
- Used in SAN (storage area network)
- Datastores in different chunks (blocks) and disks
- Ideal for databases (RDBMS) and transactional data stores
- Store VMs (Virtual Machines) on blocks
- Similar to AWS EBS (Elastic Block Storage)
- Communicate via fiber channel in SAN
- Limited metadata
- Unstructured data
- Files store in objects
- Communicates via HTTP
- Ideal for websites and web applications (static content)
- Excellent for images, videos, and social media data
- AWS S3 like functionality
- Not ideal for databases and dynamic content
- Metadata rich with descriptive traits