Foundations of Memory Management

These notes explain the early and important ideas of memory management: basic bare machine, resident monitor, multiprogramming with fixed partitions, multiprogramming with variable partitions, protection schemes, and internal and external fragmentation. These topics form the base for understanding modern memory-management methods such as paging and virtual memory.

1. Basic Bare Machine

1.1 Meaning

A basic bare machine refers to an early computer system with almost no sophisticated operating-system support for memory management. A user program is loaded directly into memory and executed there with very little abstraction. The uploaded note describes memory here as almost one raw physical space, with weak or no memory protection and usually only one job running at a time.

1.2 Characteristics

• Simple hardware and simple OS support
• No virtual memory
• No relocation flexibility
• Very weak or no memory protection
• Usually only one job runs at a time

1.3 Limitation

The main drawback is poor CPU utilization. If the only running program waits for I/O, the CPU remains idle. That makes the system inefficient and unsuitable for multiprogramming.

2. Resident Monitor

2.1 Meaning

A resident monitor is an early operating-system component that stays permanently in memory and controls the execution of jobs. It was introduced to automate job sequencing and reduce manual intervention.

2.2 Why it was introduced

In bare-machine systems, loading and running jobs required almost manual control. The resident monitor improved this by remaining in memory and taking charge of loading jobs, passing control to them, and regaining control when they finished or requested I/O.

2.3 Main functions

• Stays in a fixed region of memory
• Loads jobs one after another
• Transfers control to the user program
• Regains control after program termination or I/O request
• Maintains simple job sequencing

3. Multiprogramming with Fixed Partitions

3.1 Meaning

In fixed partitioning, main memory is divided into a fixed number of partitions before execution begins. Each partition can hold one process at a time. Several jobs may be loaded into memory together, and if one waits for I/O, the CPU can switch to another.

3.2 Types

• Equal-size partitions – all partitions have the same size
• Unequal-size partitions – partitions have different sizes so larger jobs may fit into larger partitions

3.3 Advantages

• Simple to implement
• Supports multiprogramming
• Improves CPU utilization compared with single-job systems

3.4 Disadvantages

• Internal fragmentation occurs
• Degree of multiprogramming is limited by the number of partitions
• A process larger than the biggest partition cannot be loaded

4. Multiprogramming with Variable Partitions

4.1 Meaning

In variable partitioning, memory is not predivided into fixed blocks. Instead, partitions are created dynamically according to the size of the arriving process.

4.2 Working

1. Initially, available memory is one large free block.
2. When a process arrives, the OS allocates a suitable chunk.
3. The remaining free memory forms one or more holes.
4. When a process finishes, its memory is released and becomes a hole.

A hole is a free block of memory available for allocation.

4.3 Allocation strategies

4.4 Advantages

• Better memory utilization than fixed partitioning
• Less internal fragmentation
• More flexible for different process sizes

4.5 Disadvantages

• External fragmentation occurs
• Memory management becomes more complex
• Compaction may be required

5. Fixed vs Variable Partitioning

6. Protection Schemes

6.1 Why protection is needed

Memory protection ensures that:

• one process does not overwrite another process,
• user programs do not access OS memory,
• illegal memory access is detected and trapped.

6.2 Fence Register

A fence register stores a boundary address. Memory below the fence may be reserved for the OS, and user programs may access only memory above the fence.

This is simple but too limited for complex multiprogramming.

6.3 Base Register and Limit Register

This is the classic contiguous-memory protection scheme.

• Base register – starting physical address of the process
• Limit register – size of the process memory space

6.4 Protection bits

Memory blocks, pages, or segments may have permission bits such as:

• Read
• Write
• Execute

For example:

• code segment → read + execute
• data segment → read + write
• some pages → read only

6.5 Key / lock mechanism

In some schemes, a memory block is tagged with a key or lock value, and a process can access it only if its key matches.

7. Internal and External Fragmentation

7.1 Internal Fragmentation

Internal fragmentation means unused memory exists inside an allocated block.

Internal fragmentation is common in:

• fixed partitioning
• paging (especially the last page)

7.2 External Fragmentation

External fragmentation means total free memory exists, but it is scattered into small holes, so a large required block cannot be allocated contiguously.

External fragmentation is common in:

• variable partitioning
• segmentation

8. Quick Revision Summary

9. Exam-Style Short Questions

1. What is a resident monitor?
2. Define internal fragmentation.
3. Define external fragmentation.
4. What is the difference between fixed and variable partitioning?
5. What is the role of base and limit registers?