Failure Classification and Storage

Failure Classification and Storage

Published by: Sareena Kumari Basnet

Published date: 30 Jul 2024

Failure Classification and Storage

In terms of a database, failure is defined as the inability to execute the given transaction or the loss of data. A DBMS is subject to a variety of failures, each of which requires a unique management approach. Database failures can occur for a variety of causes, including network failure, system crashes, natural disasters, carelessness, sabotage (intended data corruption), software flaws, and so on.

Failure Classification in DBMS

A failure in DBMS can be classified as:

1. Transaction Failures

  • Logical errors: Occur when logical conditions in the transaction are not met, such as division by zero or constraint violations.
  • System errors: Caused by hardware or software faults that result in transaction abortion. This may include deadlocks or a lack of resources.

2. System Failures

  • Power failures: A sudden loss of electrical power causes all volatile memory contents to be lost.
  • Software crashes: Failures caused by bugs or system crashes that prevent the database management system from working properly.

3. Media Failures

  • Disk failures: Physical damage to storage media, such as hard drives or SSDs, causes data corruption or loss.

Storage 

Storage in a Database Management System (DBMS) is critical for data management and retention. It can be divided into two sorts based on volatility and permanence.

1. Volatile Storage

  • Definition: Temporary storage that loses content when the power is switched off.
  • Example: Random access memory (RAM).
  • Use:
    • Buffers and caches: Holds interim results, current transactions, and frequently accessible data to accelerate operations.
    • Working Storage: Used to do temporary computations during query processing and execution.

2. Non-Volatile Storage

  • Definition: Permanent storage that keeps data even when the system is turned off.
  • Examples:
    • Hard Disk Drives (HDD): Traditional magnetic storage with moving components.
    • Solid State Drives (SSD): Using flash memory, you can store data more quickly and without moving parts.
    • Tape Drives: Used for archival and long-term backup storage.\

Storage Hierarchy in DBMS

1. Primary storage (main memory)

  • Characteristics: Fast access, limited capacity, and volatile.
  • Role: Stores data that is currently being processed and accessed by the CPU.

2. Secondary Storage (Disk Storage)

  • Characteristics: Larger capacity, slower access than main memory, and nonvolatile.
  • Role: Stores database files, indices, logs, and other permanent data.

3. Tertiary Storage (Archive Storage)

  • Characteristics: Very large capacity, slower access, and non-volatile.
  • Role: Used for backups, data archiving, and disaster recovery.

Storage Management Techniques

1. Data files:

  • Organization: Data is arranged into files, which are then separated into pages or blocks.
  • Structure: Tables are kept in rows on these pages.

2. Indexing:

  • Purpose: Creates indexes on one or more columns in a table to improve data retrieval speed.
  • Types: B-tree, hash, bitmap, etc.

3. Buffer Management:

  • Buffer Pools: A location in main memory where data pages are temporarily stored.
  • Buffer Replacement Policy: LRU (Least Recently Used) algorithms determine which pages should be kept in memory.

4. Logging and journaling:

  • Write Ahead Logging (WAL): Ensures that changes are recorded in a log before being applied to the database, hence facilitating recovery operations.
  • Transaction logs: Keep track of all transactions and modifications for recovery purposes.

5. Data Compression:

  • Purpose: Reduces storage space requirements by employing techniques such as row- and column-level compression.

 

FAQs About Topic
Transaction Failures System Failures Media Failures
Volatile storage (such as RAM) loses data when power is turned off, but non-volatile storage (such as HDDs and SSDs) preserves data even when shut off.
Logical errors include constraint violations. System problems include deadlocks and resource unavailability.