Segmentation is a memory-management scheme that supports the user’s view of memory, dividing a process’s logical address space into variable-sized segments—each corresponding to a logical unit such as code, data, stack, or shared libraries. Unlike paging (fixed-size blocks), segmentation preserves the program’s modular structure.

1. Segmentation Concepts

1. Logical Address
   • Represented as a pair ⟨s, d⟩, where
     • s = segment number
     • d = offset within that segment
2. Segment Table
   • Each process has its own table of n entries, one per segment.
   • Entry i contains:
     • Base\[i] = physical start address of segment i in RAM
     • Limit\[i] = length (size) of segment i
3. Address Translation
   • On a memory reference ⟨s, d⟩:
     1. Bounds Check:
        • If d ≥ Limit\[s], raise a segmentation fault.
     2. Physical Address:
        • PA = Base\[s] + d

2. Example

Consider a process with 4 segments:

• Reference 1: ⟨0, 12 000⟩
  • Check: 12 000 < 30 000 ✔
  • Physical = 0x12000 + 12 000 = 0x15000
• Reference 2: ⟨3, 5 000⟩
  • Check: 5 000 ≥ 4 096 ✖ → Segmentation fault

This illustrates how each logical reference maps to a unique, variable-length region in physical memory.

3. Advantages of Segmentation

• User-View Alignment: Segments reflect logical program structure (modules, routines), easing protection and sharing.
• Protection & Sharing:
  • Each segment can have distinct access rights (read-only code, read/write data).
  • Individual segments (e.g., libraries) can be shared among processes by mapping their tables to the same frames.
• Dynamic Growth: Segments like the heap or stack can grow independently (by updating Limit and Base).

4. Disadvantages of Segmentation