Cloud Storage Architectures

---

• various ways to store data in the cloud

Storage Types

• 3 storages types in CCSP exam
  • Long-term storage
    • is specifically designed to be used for extended periods of time
    • e.g.,
      • Amazon Glacier
      • Azure Archive Storage
      • Google Coldline and Archive tiers
  • Ephemeral storage
    • used for data that often exists only as long as an instance does
    • /tmp directory in AWS Lambda
      • data that is created and used can be stored there
      • meant to be ephemeral scratch resource
      • deleted when new execution environment is created
  • Raw storage
    • is storage that you have direct access to
    • e.g., access to hard drive, SSD, or storage volume
    • have access to the underlying storage rather than storage service

Volume Storage: File-based Storage and Block Storage

• with volume storage, customer is allocated a storage space within the cloud
  • represented as a drive attached to the virtual machine
• behaves in same manner as physical drive
  • locations and memory addresses are transparent to user
• volume storage architecture can take different forms:
  • File storage
    • aka file-level or file-based
    • data is stored and displayed as files and folders like a traditional file structure
    • popular with big data analytical tools and processes
  • Block storage
    • is a blank volume that the user can put anything into
    • might allow more flexible and higher performance
      • but requires a greater amount of administration
    • may require installation of OS or other apps to manage data
    • better suited for a volume and purpose that includes data of multiple types and kinds
• can include underlying data protection such as data replication
• can be offered in any cloud service model
  • most common in IaaS

Object-based Storage

• data is stored as objects, not files or blocks
• objects include
  • actual production content
  • metadata
  • unique address identifier for the location of a specific object
• allow for a significant level of description
  • marking, labels, classification, categorization
  • enhances indexing and data policy enforcement
• any cloud storage model can offer
  • most common in IaaS

Databases

• databases in the cloud provide structure for stored data
• data is arranged according to characteristics and elements in the data itself
  • including a primary key used to file the data
• in the cloud, a database is usually a back-end storage in the data center
  • accessed by users using online app or APIs
• can provide multiple types of databases
  • relational database
  • non-relational database (NoSQL)
    • key-value database
    • document-oriented database
• can be implemented in any cloud service model
  • most common in PaaS and SaaS