ASM Striping and Mirroring

When managing databases, two critical challenges always come up: performance and data protection. Oracle ASM (Automatic Storage Management) directly addresses these through striping and mirroring. Let’s break these concepts down in practical terms without going into DBA-only jargon.

What is Striping?

Striping is about spreading data across multiple disks instead of writing all data into a single disk.

  • How it works in ASM:
    ASM divides files into small chunks called allocation units (AUs). These AUs are evenly distributed across all available disks in the disk group.
  • Why it matters:
    Multiple disks can be read or written to in parallel, which improves I/O performance.

Example:
Suppose you have three disks (Disk1, Disk2, Disk3). A database file isn’t placed entirely on Disk1; instead, its chunks are distributed like this:

AU1 → Disk1 AU2 → Disk2 AU3 → Disk3 AU4 → Disk1 AU5 → Disk2 AU6 → Disk3

So when the database queries data, it can fetch from all disks simultaneously instead of hitting one disk repeatedly.

Result: Better throughput and faster performance.

What is Mirroring?

Mirroring is about storing multiple copies of the same data across different disks to protect against disk failures.

  • How it works in ASM:
    ASM supports different redundancy levels:
    • Normal redundancy → 2 copies of data (like RAID 1)
    • High redundancy → 3 copies of data
    • External redundancy → No mirroring (relies on hardware RAID)

Example:
In normal redundancy, if AU1 is placed on Disk1, its mirror copy (AU1’) is placed on Disk2.

AU1 → Disk1 AU1’ → Disk2 AU2 → Disk2 AU2’ → Disk3 AU3 → Disk3 AU3’ → Disk1

If Disk1 fails, ASM automatically uses the mirrored copy from Disk2 or Disk3 without downtime.

Result: Continuous availability and protection against disk failure.

Why ASM’s Striping + Mirroring is Powerful

  • Performance → Parallel I/O from multiple disks (striping).
  • Redundancy → Automatic protection against disk failures (mirroring).
  • Simplification → No need to manually manage RAID or complex storage setups. ASM abstracts this for DBAs and ensures optimal data placement automatically.

Real-World Simplified Case

Imagine an IT system where:

  • You have 6 disks.
  • ASM is configured with normal redundancy and striping.

When a database writes a table:

  • The table’s blocks are striped across all 6 disks for performance.
  • Each block also has a mirror copy on another disk for redundancy.

Even if one disk fails, your database continues running, and performance doesn’t bottleneck on a single device.

Key Takeaway

Alt text

ASM’s striping and mirroring give databases:

  • High performance (parallel disk usage)
  • High availability (automatic data redundancy)
  • Low admin overhead (ASM handles placement and recovery)

In short, ASM makes storage both faster and safer without requiring deep storage configuration expertise.