Deadlock-in-Parallel-and-Distributed-Computing.pptx
Download as PPTX, PDF0 likes15 views
Deadlock-in-Parallel-and-Distributed-Computing.pptx
- 1. Deadlock in Parallel and Distributed Computing
- 2. Course Name : Parallel and distributed computing Course Code : CSE21856 Submitted To : Debdutta Pal Professor, Adamas University. Submitted By : Anupam Das PG/SOET/210/24/003
- 3. Defining Deadlock: Circular Wait The Core Concept In a distributed system, deadlock refers to a situation where a set of processes or threads are unable to proceed because each one is waiting for a resource that is held by another process in the set. Conditions for Deadlock • Mutual exclusion: non-sharable resources. • Hold and wait: holding resources while requesting more. • No preemption: resources cannot be forcibly taken. • Circular wait: a circular chain of requests.
- 4. Real-World Deadlock Examples Database Systems Transactions wait for locks held by others. This results in rollbacks, impacting performance. Operating Systems Processes compete for memory, files, and I/O. Resource contention can lead to system stalls. Distributed Systems Services wait for responses in a loop. This creates dependency issues and downtime.
- 5. Deadlock Detection Methods Resource Deadlock Visual cycles show deadlock. Communication Deadlock Long waits for message Line Lock Cycle detection pinpoints deadlocks.
- 6. Deadlock Prevention Strategies Eliminate Mutual Exclusion Make resources sharable. Ordered Request Requesting resource should be less than acquired process. Process requests for all the possible resources. Collective Request
- 7. Deadlock Avoidance Techniques 1 Banker's Algorithm Ensures a "safe state." 2 Resource Allocation Graphs Processes declare needs upfront. 3 Optimistic Approaches Recover if deadlocks happen.
- 8. Strategies for Deadlock Recovery 1 Process Termination Abort processes to free resources. 2 Resource Preemption Take resources away from processes. 3 Rollback Restore processes to a safe state.
- 9. Best Practices for Handling Deadlock 1 Design Minimize contention with system and code review. 2 Lock Order Prevent circular wait. 3 Timeouts Detect deadlocks and trigger recovery. 4 Monitor Track resource allocation for scenarios.
- 10. Thank You