Parallel and Distributed Computing Chapter 13

Parallel and Distributed Computing Chapter 13

• 1.
  PARALLEL AND DISTRIBUTEDCOMPUTING SYNCHRONOUS/ASYNCHRONOUS COMPUTATION
• 2.
  SYNCHRONOUS  Coordination ofactions between processes  Usually asynchronous i.e. process are independent  but sometime we need to cooperate/synchronize for  Mutual exclusion or for event ordering.  Synchronization in distributed system is harder because clock system of every system having own clock and no shared memory (different)
• 3.
  COMPUTER CLOCK  Eachcomputer in distributed system has its own internal clock and can result in different time read values hen two processes running on different system.  Clock synchronization is a mechanism to sync.The time of all computers in distributed system.
• 4.
  SYNCHRONOUS DISTRIBUTED SYSTEM Process execution time lower and upper bound are set.This time include transmission, transportation and receiving time of messages  This time is known to all involved processes.  Drift rates between local clocks have known bound  Transmitted messages are received within known bound time  Can be used for hard real time applications  Timeouts are used to detect failure  Difficult and costly to implement
• 5.
  ASYNCHRONOUS DISTRIBUTED SYSTEM No Bound on process execution time, message transmission delay and drift rates  So, a distributed system is asynchronous if there is no bounds on message delivery time, speed of involved processes and clock drift.  Unpredictable in terms of timing  No timeouts can be used.  Process can crash or be too slow.
• 6.
  EVENT ORDERING  Helpsin knowing wheatear an event at one process occurred before, after or concurrently with another event at another process.