,Process Migration and
Load Balancing
, Introduction
This lecture establishes the necessity for, and
fundamental concepts of, Process Migration and Load
Balancing within the study of modern Operating
Systems.
The emergence of distributed computing—a collection
of independent computers that appears to users as a
single coherent system—presents significant resource
management challenges not found in centralized
systems.
The Problem of Uneven Workload
In a distributed system, resources (CPU time, memory,
I/O bandwidth) are naturally utilized unevenly due to
random task arrivals, user behavior, and application
demands.
SPC2201: Operating Systems II 3
, Introduction
Overloaded Nodes (Hotspots): Some nodes might be
executing many processes, leading to high CPU utilization,
long waiting queues, excessive swapping/paging
(thrashing), and very poor response times.
Underloaded Nodes (Idle Hosts): Simultaneously, other
nodes might have few or no processes running, leading to
wasted computational power.
This imbalance results in inefficient resource
utilization and degrades overall system performance,
defeating the purpose of having a powerful network of
machines
The Solution: Dynamic Resource Allocation
To address this imbalance, the operating system must employ
advanced Resource Management techniques capable of
moving work to available resources.
SPC2201: Operating Systems II 4
Load Balancing
, Introduction
This lecture establishes the necessity for, and
fundamental concepts of, Process Migration and Load
Balancing within the study of modern Operating
Systems.
The emergence of distributed computing—a collection
of independent computers that appears to users as a
single coherent system—presents significant resource
management challenges not found in centralized
systems.
The Problem of Uneven Workload
In a distributed system, resources (CPU time, memory,
I/O bandwidth) are naturally utilized unevenly due to
random task arrivals, user behavior, and application
demands.
SPC2201: Operating Systems II 3
, Introduction
Overloaded Nodes (Hotspots): Some nodes might be
executing many processes, leading to high CPU utilization,
long waiting queues, excessive swapping/paging
(thrashing), and very poor response times.
Underloaded Nodes (Idle Hosts): Simultaneously, other
nodes might have few or no processes running, leading to
wasted computational power.
This imbalance results in inefficient resource
utilization and degrades overall system performance,
defeating the purpose of having a powerful network of
machines
The Solution: Dynamic Resource Allocation
To address this imbalance, the operating system must employ
advanced Resource Management techniques capable of
moving work to available resources.
SPC2201: Operating Systems II 4
