8/6/2019 L33 Clusters

1/13

Clusters

Alternative to symmetric

multiprocessing (SMP)

Group of interconnected, wholecomputers working together as a unified

computing resource

illusion is one machine

system can run on its own

8/6/2019 L33 Clusters

2/13

Benefits of Clusters

Absolute scalability

Can have dozens of machines each of which is a

multiprocessor

Incremental scalability Add new systems in small increments

High availability

Failure of one node does not mean loss of service

Superior price/performance

Cluster can be equal or greater computing power

than a single large machine at a much lower cost

8/6/2019 L33 Clusters

3/13

Cluster Configurations - Standby

Server, No Shared Disk

8/6/2019 L33 Clusters

4/13

Cluster Configurations -

Shared Disk

8/6/2019 L33 Clusters

5/13

Clustering Methods

Passive standby

Active secondary

Separate serversServers connected to disks

Servers share disks

8/6/2019 L33 Clusters

6/13

Clusters

Separate server

Each computer is a separate server

No shared disks

Need management or scheduling software

Data must be constantly copied among

systems so each is current

8/6/2019 L33 Clusters

7/13

Clusters

Shared nothing

Reduces communication overhead

Several servers connected to common disks

Disks partitioned into volumes

Each volume owned by a computer

If computer fails another computer gets

ownership of the volume

8/6/2019 L33 Clusters

8/13

Clusters

Shared disk

Multiple computers share the same disks at

the same time

Each computer has access to all of the

volumes on all of the disks

8/6/2019 L33 Clusters

9/13

Operating System Design

Issues Failure managementHighly available cluster offers a high

probability that all resources will be in

service

No guarantee about the state of partially

executed transactions if failure occurs

Fault-tolerant cluster ensures that all

resources are always available

8/6/2019 L33 Clusters

10/13

8/6/2019 L33 Clusters

11/13

Cluster Computer

Architecture Cluster middleware services and functionsSingle entry point : User logs onto cluster rather

than to an individual computer

Single file hierarchy : user sees a single hierarchyof file directories under same root directory

Single control point : A default workstation is usedfor cluster management and control

Single virtual networking : Any node can accessany other point in the cluster, even though the actualcluster config. consists of multiple interconnected n/ws

Single memory space

Single job-management system : cluster jobscheduler

8/6/2019 L33 Clusters

12/13

Cluster Computer

Architecture Cluster middleware services and functionsSingle user interface

Single I/O space: any node can access remote I/Operipheral without knowledge of its physical location

Single process space : a uniform processidentification scheme is used

Checkpointing : this function periodically saves theprocess state & intermediate computing results, toallow rollback recovery after a failure

Process migration: Enables load balancing

8/6/2019 L33 Clusters

13/13

Clusters Compared to SMP

SMP is easier to manage and configure

SMP takes up less space and draws less

power Clusters are better for incremental and

absolute scalability

Clusters are superior in terms of

availability