Distributed deadlocks

Views:
 
Category: Education
     
 

Presentation Description

No description available.

Comments

Presentation Transcript

Distributed deadlocks : 

Distributed deadlocks Sindoora

CONTENT : 

CONTENT Definitions Strategies for handling deadlocks Deadlock Detection Approaches for deadlock detection Existing algorithms Deadlock Resolution OR model AND model

Definition : 

Definition Deadlock: A deadlock is a situation in which two computer programs sharing the same resource are effectively preventing each other from accessing the resource, resulting in both programs ceasing to function. Distributed deadlocks: This event occurs when two programs residing on separate computers are deadlocked. Deadlocks in distributed systems. are harder to avoid, prevent or even detect

Handling Deadlocks : 

Handling Deadlocks There are four strategies for handling deadlocks Deadlock prevention Deadlock avoidance Deadlock detection Deadlock resolution

Distributed Deadlock Detection : 

Distributed Deadlock Detection The AND model: A process or transaction can simultaneously request multiple resources. Remains blocked until all the requested resources are allocated. - Cycle is sufficient to declare deadlock The OR model: A process or transaction can simultaneously request multiple resources. Remains blocked until any one of the requested resource is allocated. - Knot is sufficient to declare deadlock

Distributed Deadlock Detection : 

Distributed Deadlock Detection Approaches fall under three categories Centralized Control Distributed Control Hierarchal Control

Algorithms for deadlock detection : 

Algorithms for deadlock detection Ho-Ramamoorthy Algorithm Path-pushing algorithm Edge-chasing algorithm Diffusion Computation Global state detection The Menasce-Muntz Algorithm

Deadlock Resolution Algorithms : 

Deadlock Resolution Algorithms OR model: Never causes false deadlock resolutions Each process issues request to multiple resources Remains blocked until any one of the resource is allocated Request, reply, cancel are the three requests that flow throughout the execution AND model: Only true deadlocks are resolved. The node specific variables used are status, version, create, transmitted, to_reply. Three types of messages are handled by the algorithm probes, information messages, reply messages.

Distributed deadlock resolution algorithms for AND & OR models : 

Distributed deadlock resolution algorithms for AND & OR models OR model algorithm execution: (7, t_init7, 7, t_block7, 1, path) (5, t_init5, notify)

Distributed deadlock resolution algorithms for AND & OR models : 

Distributed deadlock resolution algorithms for AND & OR models AND model algorithm execution:

Queries ? : 

Queries ?

Thank you : 

Thank you