Distributed Resource

Management
DISTRIBUTED SCHEDULING
DISTRIBUTED SHARED MEMORY
DISTRIBUTED FILE SYSTEM
FAILURE RECOVERY & FAULT
TOLERANCE
Distributed Scheduling
Introduction

 Good resource allocation schemes are needed to fully

utilize the computing capacity of the DS
 Distributed scheduler is a resource management component
of a DOS
 It focuses on judiciously and transparently redistributing the
load of the system among the computers
 Target is to maximize the overall performance of the system
 More suitable for DS based on LANs
Motivation

 A locally distributed system consists of a collection of

autonomous computers connected by a local area
communication network
 Users submit tasks at their host computers for processing
 Load distributed is required in such environment because of
random arrival of tasks and their random CPU service time
 There is a possibility that several computers are heavily loaded
and others are idle of lightly loaded
 If the load is heavier on some systems or if some processors
execute tasks at a slower rate than others, this situation will
occur often
Distributed Systems Modeling

 Consider a system of N identical and independent

servers
 Identical means that all servers have the same task
arrival and service rates
 Let ρ be the utilization of each server, then P=1- ρ, is
the probability that a server is idle
 If the ρ=0.6, it means that P=0.4,
 If the systems have different load than load can be
transferred from highly loaded systems to lightly load
systems to increase the performance
Issues in Load Distribution

 Load
 Resource queue lengths and particularly the CPU queue length
are good indicators of load
 Measuring the CPU queue length is fairly simple and carries
little overhead
 CPU queue length does not always tell the correct situation as
the jobs may differ in types
 Another load measuring criterion is the processor utilization
 Requires a background process that monitors CPU utilization
continuously and imposes more overhead
 Used in most of the load balancing algorithms
Classification of LDA

 Basic function is to transfer load from heavily loaded

systems to idle or lightly loaded systems
 These algorithms can be classified as :
 Static
 decisions are hard-wired in the algorithm using a prior knowledge of the
system
 Dynamic
 use system state information to make load distributing decisions
 Adaptive
 special case of dynamic algorithms in that they adapt their activities by
dynamically changing the parameters of the algorithm to suit the changing
system state
Basic Terminologies

 Load Balancing vs. Load sharing

 Load sharing algorithms strive to reduce the possibility for a
system to go to a state in which it lies idle while at the same
time tasks contend service at another, by transferring tasks to
lightly loaded nodes
 Load balancing algorithms try to equalize loads at al
computers
 Because a load balancing algorithm transfers tasks at higher
rate than a load sharing algorithm, the higher overhead
incurred by the load balancing algorithm may outweigh this
potential performance improvement
Basic Terminologies (contd.)

 Preemptive vs. Non-preemptive transfer

 Preemptive task transfers involve the transfer of a task that is
partially executed
 Non-preemptive task transfers involve the transfer of the tasks
that have not begun execution and hence do not require the
transfer of the task’s state
 Preemptive transfer is an expensive operation as the collection
of a task’s state can be difficult
 What does a task’s state consist of?
 Non-preemptive task transfers are also referred to as task
placements
 Components of a Load
Balancing Algorithm
 Transfer Policy
 determines whether a node is in a suitable state to participate in a task transfer
 requires information on the local nodes’ state to make decisions
 Selection Policy
 determines which task should be transferred
 Location Policy
 determines to which node a task selected for transfer should be sent
 requires information on the states of remote nodes to make decisions
 Information policy
 responsible for triggering the collection of system state information
 Three types are: Demand-Driven, Periodic, State-Change-Driven
Stability

 The two views of stability are,

 The Queuing-Theoretic Perspective
 A system is termed as unstable if the CPU queues grow without bound
when the long term arrival rate of work to a system is greater than the
rate at which the system can perform work.
 The Algorithmic Perspective
 If an algorithm can perform fruitless actions indefinitely with finite
probability, the algorithm is said to be unstable.
Load Distributing Algorithms

 Sender-Initiated Algorithms
 Receiver-Initiated Algorithms
 Symmetrically Initiated Algorithms
 Adaptive Algorithms
Sender-Initiated Algorithms

 Activity is initiated by an overloaded node (sender)

 A task is sent to an underloaded node (receiver)
 Transfer Policy
 A node is identified as a sender if a new task originating at the node makes the queue length exceed a
threshold T.
 Selection Policy
 Only new arrived tasks are considered for transfer
 Location Policy
 Random: dynamic location policy, no prior information exchange
 Threshold: polling a node (selected at random) to find a receiver
 Shortest: a group of nodes are polled to determine their queue
 Information Policy
 A demand-driven type
 Stability
 Location policies adopted cause system instability at high loads
 Yes

Select Node “i” No

“i” is Poll-set Poll-set=Poll-set U “i” Poll Node “i”
randomly

Poll-set = Nil
Transfer task Yes QueueLength at “I”
Yes to “i” <T

Task
Arrives QueueLength+1 No
>T
No. of polls
Yes
<
No
PollLimit

No
Queue the
task locally
Receiver-Initiated Algorithms

 Initiated from an underloaded node (receiver) to obtain a task from an

overloaded node (sender)
 Transfer Policy
 Triggered when a task departs
 Selection Policy
 Same as the previous
 Location Policy
 A node selected at random is polled to determine if transferring a task from it would place its
queue length below the threshold level, if not, the polled node transfers a task.
 Information Policy
 A demand-driven type
 Stability
 Do not cause system instability in high system load, however, in low load it spare CPU
cycles
 Most transfers are preemptive and therefore expensive
 Yes

Select Node “i” No

“i” is Poll-set Poll-set=Poll-set U “i” Poll Node “i”
randomly

Poll-set = Nil
Transfer task Yes QueueLength at “I”
Yes from “i” to “j” >T

QueueLength No
<T
No. of polls
Yes
<
No
PollLimit
Wait for a
perdetermined period No

Task Departure at “j”

 Symmetrically Initiated
Algorithms
 Both senders and receivers search for receiver and senders,
respectively, for task transfer.
 The Above-Average Algorithm
 Transfer Policy
 Thresholds are equidistant from the node’s estimate of the average load across all
node.
 Location Policy
 Sender-initiated component: Timeout messages TooHigh, TooLow, Accept,
AwaitingTask, ChangeAverage
 Receiver-initiated component: Timeout messages TooLow, LooHigh, Accept,
AwaitingTask, ChangeAverage
 Selection Policy
 Similar to both the earlier algorithms
 Information Policy
 A demand-driven type but the acceptable range can be increased/decreased by each
node individually.
Adaptive Algorithms

 A Stable Symmetrically Initiated Algorithm

 Utilizes the information gathered during polling to classify the nodes in the system as either
Sender, Receiver or OK.
 The knowledge concerning the state of nodes is maintained by a data structure at each node,
comprised of a senders list, a receivers list, and an OK list.
 Initially, each node assumes that every other node is a receiver.
 Transfer Policy
 Triggers when a new task originates or when a task departs.
 Makes use of two threshold values, i.e. Lower (LT) and Upper (UT)
 Location Policy
 Sender-initiated component: Polls the node at the head of receiver’s list
 Receiver-initiated component: Polling in three order
 Head-Tail (senders list), Tail-Head (OK list), Tail-Head (receivers list)
 Selection Policy: Newly arrived task (SI), other approached (RI)
 Information Policy: A demand-driven type
Students’ Tasks

11.7 PERFORMANCE COMPARISON

11.8 SELECTING A SUITABLE
LOAD-SHARING ALGORITHM
Task Migration

 Receiver-initiated task transfers can improve system performance at high

system loads.
 Receiver-initiated transfers require preemptive task transfer.
 Task Placement refers to the transfer of a task that is yet to begin execution to a new
location and start its execution there.
 Task Migration refers to that transfer of a task that has already begun execution to a
new location and continuing its execution there.
Task Migration (contd.)

 Steps involved in Task Migration

 State Transfer
 The transfer of the task’s state including information e.g. registers, stack, ready/blocked,
virtual memory address space, file descriptors, buffered messages etc. to the new machine.
 The task is frozen at some point during the transfer so that the state does not change
further.
 Unfreeze
 The task is installed at the new machine and is put in the ready queue so that it can
continue executing.
Issues in Task Migration

 State Transfer
 Location Transparency
 Structure of a Migration Mechanism
 Performance
State Transfer

 The Cost
 To support remote execution, obtaining and transferring the state, and unfreezing the task.
 Residual Dependencies
 Refers to the amount of resources a former host of a preempted or migrated task continues to
dedicate to service requests from the migrated task.
 Implementations
 The V-System
 Attempts to reduce the freezing time of a migrating task by precopying the state.
 The bulk of the task state is copied to the new host
 It increases the number of messages that are sent to new host
 SPRITE
 Makes use of the location-transparent file access mechanism provided by its file system
 All the modified pages of the migrating task are swapped to file server
 ACCENT
 Reduction in migration is achieved by using a feature called Copy-on-Reference
 The entire virtual memory address space is not copied to the new host
Location Transparency

 Services that are provided to user processes irrespective of the location of

the processes and services.
 In distributed systems, it is essential that the location transparency by
supported.
 Location transparency in principle requires that names (e.g. process
names, file names) be independent of their location (i.e. host names).
 Any operation (such as signaling) or communication that was possible
before the migration of a task should be possible after its migration
 Example – SPRITE – Location Transparency Mechanisms
 A location-transparent distributed file system is provided
 The entire state of the migrating task is made available at the new host, and
therefore, any kernel calls made will be local at new host
 Location-dependent information such as host of a task is maintained at the home
machine of a task
Structure of a Migration
Mechanism
 Issues involved in Migration Mechanisms
 Decision whether to separate the policy-making modules from
mechanism modules
 It has implications for both performance and the ease of development
 The separation of policy and mechanism modules simplifies the development
efforts
 Decision to where the policy and mechanisms should reside
 The migration mechanism may best fit inside the kernel
 Policy modules decide whether a task transfer should occur, this can be
placed in the kernel as well
 Interplay between the task migration mechanism and various other
mechanisms
 The mechanisms can be designed to be independent of one another so that if
one mechanism’s protocol changes, the other’s need not
Performance

 Comparing the performance of task migration mechanisms implemented in

different systems is a difficult task, because of the different,
 Hardware
 SPRITE consists of a collection of SPARCSTATION 1
 CHARLOTTE consists of VAX/11-750 machines
 Operating systems
 IPC mechanism
 File systems
 Policy mechanisms