Deadlock Ignorance

The Deadlock Ignorance strategy simply assumes:

• That deadlocks are so rare that it is not worth the cost of preventing or detecting them.
• If a deadlock does occur, the operating system may take drastic measures such as rebooting to recover.
• This approach is called the Ostrich Algorithm, because it resembles the ostrich burying its head in the sand and pretending the problem doesn’t exist.

Why Use Deadlock Ignorance?

1. Rarity of Deadlocks:

• Deadlocks occur very rarely in well-designed programs.
• Other failures (hardware crashes, software bugs, compiler errors) are much more common.
• Engineers prefer to spend effort on frequent problems rather than rare ones.

2. High Overhead of Handling:

• Deadlock prevention/avoidance requires keeping track of resource allocation graphs, running detection cycles or restricting process behavior.
• These add significant time and space overhead.
• For most users, ignoring deadlock makes the OS faster and simpler.

3. Practical Philosophy:

• For personal computers or single-user systems, rebooting after a rare deadlock is acceptable.
• For mission-critical systems (e.g., banking, aviation, medical devices), this strategy is not suitable.

Note: many widely used systems, including UNIX and Windows, often adopt deadlock ignorance.

Real-World Examples

Process Table Limit:

• OS maintains a fixed-size process table.
• If all entries are full, fork() fails.
• Instead of detecting a deadlock, the system simply waits until an entry is freed.

File System (i-node Table):

• Limited number of open files per process.
• If the table is full, new file requests are blocked until a slot is released.

Swap Space Limit:

• Memory allocation is limited.
• If multiple processes demand more than available space, OS does not detect a “deadlock” but lets processes fail gracefully or retry later.

Advantages of Deadlock Ignorance

1. Simplicity: The OS design remains simple without complex prevention or detection algorithms.
2. Better Performance: Avoids the runtime overhead of tracking wait-for graphs or running detection cycles.
3. Low Development Cost: No need to implement prevention/recovery modules.
4. Practical Feasibility: Deadlocks occur so infrequently that ignoring them is often cheaper than handling them.

Disadvantages of Deadlock Ignorance

1. Unpredictable Behavior: System may suddenly freeze if deadlock occurs.
2. System Crashes: May require reboot, leading to data loss.
3. Reduced Reliability: Not acceptable for critical systems.
4. No Control: Users cannot resolve deadlocks explicitly.
5. Resource Starvation Risk: Processes may remain blocked indefinitely.

Note: Scientists all over the world believe that the most efficient method to deal with deadlock is deadlock prevention. But the Engineers that deal with the system believe that deadlock prevention should be paid less attention as there are very less chances for deadlock occurrence.