33
Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems Group

Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Embed Size (px)

Citation preview

Page 1: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Performance-responsive Middleware for Grid Computing

Dr Stephen JarvisHigh Performance Systems Group

University of Warwick, UK

High Performance Systems Group

Page 2: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Context

• Funded by / collaborating with – UK e-Science Core Programme– IBM (Watson, Hursley)– NASA (Ames)– NEC Europe– Los Alamos National Laboratory

• Integrate established performance tools into emerging grid middleware

High Performance Systems Group

Page 3: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

High Performance Systems Group

Grid Resource Management

How do we enable and regulate the resource

sharing between users?

While…

providing vision of access to full resources

hiding detail & unnecessary complexity

providing acceptable levels of service

Page 4: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

High Performance Systems Group

Workload Generation, Visualisation…

Discovery, Mapping, Scheduling, Security, Accounting…

Computing, Storage, Instrumentation…

Managing through Middleware

Key interface between applications & resources

Page 5: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

High Performance Systems Group

Key Middleware Activities

• Determine what resources are required (advertise)

•Determine what resources are available (discovery)

•Map requirements to available resources (scheduling)

•Maintain contract of performance (service level agreement)

Page 6: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Performance Services

• Intra-domain– Lab- / department-based

– Shared resources under local administration

• Multi-domain– Campus- / country-based

– Wide-area resource and task management

– Cross domain

High Performance Systems Group

Page 7: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Performance Services

High Performance Systems Group

• Intra-domain– Lab- / department-based

– Shared resources under local administration

• Multi-domain– Campus- / country-based

– Wide-area resource and task management

– Cross domain

Page 8: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Performance Services

High Performance Systems Group

• Intra-domain– Lab- / department-based

– Shared resources under local administration

• Multi-domain– Campus- / country-based

– Wide-area resource and task management

– Cross domain

Page 9: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Performance Prediction

• Performance prediction tools• Aim to predict

– Execution time– Communication usage– Data and resource requirements

• Provides best guess as to how an application will execute on a given resource

High Performance Systems Group

Page 10: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

High Performance Systems Group

PACE User

Application

Resource

Page 11: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

High Performance Systems Group

PACE User

Application

Resource

ApplicationModel

Resource Model

Page 12: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Application

ApplicationModel

Resource

Resource Model

PACE User

Evaluation Engine

Model parameters

Resource config.

High Performance Systems Group

Page 13: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Application

ApplicationModel

Resource

Resource Model

PACE User

Evaluation Engine

Model parameters

Resource config.

High Performance Systems Group

Page 14: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Why is prediction useful?

• Scaling properties

• Compare runtime options with– deadline

– available resources

– priority / other jobs

– etc.

High Performance Systems Group

0

5

10

15

20

25

30

35

40

45

50

1 4 7 10 13 16

The Number of Processors

Run

ning

Tim

e on

SG

IOri

gin2

000

(sec

)

sweep3d

fft

improc

closure

jacobi

memsort

cpi

Allows runtime scenarios to be explored before deployment

Page 15: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

1. Intra-Domain Co-Scheduling

High Performance Systems Group

• Augment emerging middleware with

additional performance information

• Handle predictive and non-predictive tasks

• Use predictive data for system improvement

– Time to complete tasks / utilisation of resources

– QoS – ability to meet deadlines

• Scheduler driver, or co-scheduler (called

Titan)

Page 16: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Intra-Domain Co-Scheduling

High Performance Systems Group

• Non-predictive tasks

PORTALPRE-

EXECUTIONENGINE MATCHMAKER

SCHEDULEQUEUE

PACE

GA CLUSTERCONNECTOR

CONDORCONDOR

REQUESTS FROM USERS OR OTHERDOMAIN SCHEDULERS

RESOURCES

CLASSADS

Titan

Page 17: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Intra-Domain Co-Scheduling

High Performance Systems Group

• Non-predictive tasks

PORTALPRE-

EXECUTIONENGINE MATCHMAKER

SCHEDULEQUEUE

PACE

GA CLUSTERCONNECTOR

CONDORCONDOR

REQUESTS FROM USERS OR OTHERDOMAIN SCHEDULERS

RESOURCES

CLASSADS

Titan

Page 18: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Intra-Domain Co-Scheduling

High Performance Systems Group

• Non-predictive tasks• Tasks with prediction

data

PORTALPRE-

EXECUTIONENGINE MATCHMAKER

SCHEDULEQUEUE

PACE

GA CLUSTERCONNECTOR

CONDORCONDOR

REQUESTS FROM USERS OR OTHERDOMAIN SCHEDULERS

RESOURCES

CLASSADS

Titan

Page 19: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Intra-Domain Co-Scheduling

High Performance Systems Group

• Non-predictive tasks• Tasks with prediction

data

PORTALPRE-

EXECUTIONENGINE MATCHMAKER

SCHEDULEQUEUE

PACE

GA CLUSTERCONNECTOR

CONDORCONDOR

REQUESTS FROM USERS OR OTHERDOMAIN SCHEDULERS

RESOURCES

CLASSADS

Titan

Page 20: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Intra-Domain Co-Scheduling

High Performance Systems Group

• Non-predictive tasks• Tasks with prediction

data

PORTALPRE-

EXECUTIONENGINE MATCHMAKER

SCHEDULEQUEUE

PACE

GA CLUSTERCONNECTOR

CONDORCONDOR

REQUESTS FROM USERS OR OTHERDOMAIN SCHEDULERS

RESOURCES

CLASSADS

Titan

Page 21: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Intra-Domain Co-Scheduling

High Performance Systems Group

• Non-predictive tasks• Tasks with prediction

data

PORTALPRE-

EXECUTIONENGINE MATCHMAKER

SCHEDULEQUEUE

PACE

GA CLUSTERCONNECTOR

CONDORCONDOR

REQUESTS FROM USERS OR OTHERDOMAIN SCHEDULERS

RESOURCES

CLASSADS

Titan

Page 22: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Intra-Domain Co-Scheduling

High Performance Systems Group

• Non-predictive tasks• Tasks with prediction

data

PORTALPRE-

EXECUTIONENGINE MATCHMAKER

SCHEDULEQUEUE

PACE

GA CLUSTERCONNECTOR

CONDORCONDOR

REQUESTS FROM USERS OR OTHERDOMAIN SCHEDULERS

RESOURCES

CLASSADS

Titan

Page 23: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Intra-Domain Deployment

Without co-scheduler With co-scheduler

Time to complete = 70.08m Time to complete = 35.19m

High Performance Systems Group

Page 24: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

• Publish intra-domain perf. data through

global information services (MDS)

• Augment service with agent system

– One agent per domain / VO

• When a task is submitted

– Agents query IS, and negotiate to discover best

domain to run task

• Scheme is tested on a 256-node exp. Grid

– 16 resource domains; 6 arch. types

High Performance Systems Group

2. Multi-Domain Management

Page 25: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

High Performance Systems Group

Multi-Domain Management

time

Page 26: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

High Performance Systems Group

Multi-Domain Management

time

Page 27: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

High Performance Systems Group

Multi-Domain Management

time

Page 28: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

High Performance Systems Group

Multi-Domain Management

Time to complete = 2752s

Page 29: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Multi-Domain Management

High Performance Systems Group

Time to complete = 467s; an improvement of 83%

Page 30: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Multi-Domain Management

High Performance Systems Group

Time to complete = 467s; an improvement of 83%

Page 31: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

QoS: Ability to Meet Deadline

High Performance Systems Group

active inactive

Page 32: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Resource usage

High Performance Systems Group

active inactive

Page 33: Performance-responsive Middleware for Grid Computing Dr Stephen Jarvis High Performance Systems Group University of Warwick, UK High Performance Systems

Many Issues Remain• Identification of meaningful QoS metrics

– User-orientated– Contract-based

• Honouring of SLA – End-to-end service management– Resolving conflicts

• Managing Workflow (CCGrid 2003)– See poster & demo

• But…version 1.0, Condor/GT2-based, available for download– See www.dcs.warwick.ac.uk/~hpsg

High Performance Systems Group