modeling of event-based communication in component-based ...€¦ · palladio component model,”...

KIT – University of the State of Baden-Wuerttemberg and

National Research Center of the Helmholtz Association

DESCARTES RESEARCH GROUP, CHAIR FOR SOFTWARE DESIGN AND QUALITY

INSTITUTE FOR PROGRAM STRUCTURES AND DATA ORGANIZATION, FACULTY OF INFORMATICS

www.kit.edu

Christoph Rathfelder, Benjamin Klatt and Samuel Kounev

Keynote talk presented by Samuel Kounev at FESCA @ ETAPS 2012

Tallinn, Estonia, March 31, 2012

Modeling of Event-based Communication in

Component-based Architectures

Descartes Research Group

Institute for Program Structures and Data Organization 2 31.03.2012

Agenda

Introduction to Event-based Communication

Design-time Modeling and Analysis

Palladio Component Model (PCM)

Meta-Model Extensions for Event-based Communication

Case Studies

Run-Time Quality-of-Service Management

Descartes Meta-Model (DMM)

Outlook

© Christoph Rathfelder, Benjamin Klatt and Samuel Kounev



References (1)

Christoph Rathfelder. Modelling Event-Based Interactions in

Component-Based Architectures for Quantitative System Evaluation.

PhD Thesis, KIT, In preparation, March 2012.




References (2)

[1] B. Klatt, C. Rathfelder and S. Kounev, “Integration of event-based communication in the palladio software quality prediction

framework,” in Proceedings of the joint ACM SIGSOFT conference - QoSA and ACM SIGSOFT symposium - ISARCS on Quality

of software architectures - QoSA and architecting critical systems - ISARCS, QoSA-ISARCS '11, Seiten 43-52, New York, NY,

USA; Best Paper Nomination, 2011.

[2] C. Rathfelder and B. Klatt, “A quality-prediction tool for component-based architectures,” in Proceedings of the 2011 Ninth Working

IEEE/IFIP Conference on Software Architecture, WICSA '11, Seiten 347-350, Washington, DC, USA, 2011.

[3] C. Rathfelder and S. Kounev, "Model-based performance prediction for event-driven systems," in Proceedings of the Third ACM

International Conference on Distributed Event-Based Systems, DEBS '09, Seiten 33:1-33:2, 2009.

[4] C. Rathfelder and S. Kounev, “Modeling Event-Driven Service-Oriented Systems using the Palladio Component Model,” in

Proceedings of the International Workshop on the Quality of Service-Oriented Software Systems (QUASOSS), Seiten 33-38.,

2009.

[5] C. Rathfelder, D. Evans and S. Kounev, “Predictive Modelling of Peer-to-Peer Event-driven Communication in Component-based

Systems,” in Proceedings of the 7th European Performance Engineering Workshop (EPEW'10), volume 6342 of Lecture Notes in

Computer Science, Seiten 219-235., University Residential Center of Bertinoro, Italy, 2010.

[6] C. Rathfelder, B. Klatt, S. Kounev and D. Evans, “Towards middleware-aware integration of event-based communication into the

palladio component model,” in Proceedings of the Fourth ACM International Conference on Distributed Event-Based Systems,

DEBS '10, Seiten 97-98, 2010.

[7] C. Rathfelder, S. Kounev and D. Evans, “Capacity Planning for Event-based Systems using Automated Performance Predictions,”

in 26th IEEE/ACM International Conference On Automated Software Engineering (ASE 2011), Seiten 352-361, Oread, Lawrence,

Kansas; Annahmequote: 14.7% (37/252), 2011.

[8] C. Rathfelder, B. Klatt, F. Brosch and S. Kounev, “Performance Modeling for Quality of Service Prediction in Service-Oriented

Systems,” in Handbook of Research on Service-Oriented Systems and Non-Functional Properties: Future Directions, Seiten 258-

279, IGI Global, 2012.

[9] C. Rathfelder, B. Klatt, K. Sachs und S. Kounev, „Modeling Event-based Communication in Component-based Software,“ Journal

on Software and Systems Modeling; Theme Issue on Models for Quality of Software Architecture, 2012; Invited paper under

review.




References (3)

Kai Sachs. Performance Modeling and Benchmarking of Event-Based

Systems. PhD Thesis, TU Darmstadt, ISBN-13: 9783868443530,

Sierke Verlag, August 30, 2011.




Agenda





Case Studies



Outlook




Event-based Communication


Decoupling in the 3 dimensions [Eugster]

Communicating parties:

1. Do not need to be active at the same time (Time)

2. Do not need to know each other (Space)

3. Are not blocked when exchanging messages

(Synchronization)



Event-based Communication (2)

Event A significant change in state

Source Producer, publisher, sender, generator, or monitoring component.

Transmission System

Notification service, event service, event-based middleware,

channel or event bus.

Sink Reactive components, consumers, subscribers, or receivers




Event Delivery Model




Motivating Example

Store Scenario

Event Bus

Logging Service

Sink UpdateStockData

Source UpdateStockData

RFID

Scanner


Cashdesk Service


Prov . Interface CreateOrder

Order

Managment Service

Req Interface CreateOrder

Inventory

Management

Service



Institute for Program Structures and Data Organization 11 31.03.2012 © Christoph Rathfelder, Benjamin Klatt and Samuel Kounev

Event Bus

Logging Service



RFID

Scanner


Cashdesk Service


Prov . Interface CreateOrder

Order

Managment Service

Req Interface CreateOrder

Inventory

Management

Service

Sizing

vs.

Changing Usage

vs. vs.

Prediction

• Max. Throughput

• Processing Time

Analyses

• Bottleneck

• Resource Utilization

Motivating Example

Store Scenario



Agenda





Case Studies



Outlook




Palladio Software Quality Prediction Framework

• Domain specific modelling language

• Aligned with UML 2 design models

• Component-based architectures

• http://www.palladio-approach.net

Palladio Component Model

• Design-time quality prediction

• http://www.palladio-simulator.com

Eclipse-based modelling and prediction tool

• Simulation code

• Layered queueing networks

• Queueing Petri nets

Transformations into prediction models




Performance Analyses:

Component

Repository

System Model

Deployment

Model

Usage Model

Palladio Modelling Approach



Component Model

Architecture Model

Deployment Model

Usage Model

Service-Level Prediction

Resource Utilization

Response Time Palladio

Approach

Simulation Code

Layered Queueing

Networks

Queueing

Petri Nets

Stochastic

Regular Expr.

Model Solution Techniques




Agenda





Case Studies



Outlook




Integration of Event-Based Communication in the

Palladio Software Quality Prediction Framework

Benjamin Klatt, FZI

Christoph Rathfelder, FZI

Samuel Kounev, KIT

Event

B. Klatt, C. Rathfelder and S. Kounev. Integration of Event-Based Communication

in the Palladio Software Quality Prediction Framework. In 7th ACM SIGSOFT

International Conference on the Quality of Software Architectures (QoSA 2011),

pages 43-52, Boulder, Colorado, USA, June 20-24, 2011.




Events and Components




Direct Point-to-Point Connections




Component Component

ComponentInterface

Event

Sink

Event

Source

Component

ComponentComponent

Graphical Notation




Publish/Subscribe Communication




PCM Meta-Model Extensions

Sources, Sinks & Connectors

Emit Event Action Components & Roles

Explicit modelling of

Events

Source and sink ports

Many-to-many connections

Event producer (Source)

and handler (Sink)




Event

Extended PCM Model

Transformation

Original PCM Model

Middleware

Weaving

Middleware

Platform SpecificModel

Prediction

PredictionResult

Quality-Prediction of component-based architectures with event-based communication

integrated into Palladio

Explicit event modeling

with reduced effort

Considering event communication middleware influences

Reuse existing prediction techniques

Approach




1. Meta-Model Extension

Extension of the PCM

meta-model

Semantically correct

modelling of • Events

• Source and sink ports

• 1-many connections

• Event handlers

Extended PCM platform-independent




Extensions are substituted

with existing elements Performance equivalent model

Allows reuse of existing prediction

techniques

Does not consider platform-

specific resource demands


Classical PCM platform-independent

Transformation

2. Model-to-Model Transformation




3. Weaving of Platform Specific Components

• Middleware repository • Platform-specific components

• Behaviour

• Resource demands

• Weaving with platform-

independent model

• Input for quality predictions


Classical PCM platform-independent

Transformation

Middleware Repository platform-specific

Final Model platform-specific

Weaving

Quality Prediction




Transformation

Example

Ext.PCM

Class. PCM Middleware Repository

Final Model

Prediction

Source A

Sink B

Sink C

A

B

C

Platform independent

C

Sink C

Event

Receiver

IMiddleware

Receiver

SinkPort

IMiddleware

SinkPort

IEvent

Sink

A

Source A

IEvent

Source

IEvent Receiver

IEvent

Receiver

SourcePort

Event

Distribution

IMiddleware

SourcePort

IMiddleware

EventDistribution

IMiddleware

Sender

Event

Sender Distribution

Preparation

IMiddleware

DistributionPreparation

Event

Sender

Platform specific

Middleware

SourcePort Hub

Middleware Middleware SinkPort




Agenda





Case Studies



Outlook




Predictive Modelling of Peer-to-Peer Event-driven

Communication in Component-based Systems

Christoph Rathfelder FZI, Germany

David Evans University of Cambridge

Samuel Kounev KIT, Germany

Event

C. Rathfelder, D. Evans, and S. Kounev. Predictive Modelling of Peer-to-Peer Event-driven

Communication in Component-based Systems. In Proceedings of the 7th European

Performance Engineering Workshop (EPEW'10), University Residential Center of Bertinoro,

Italy, volume 6342 of Lecture Notes in Computer Science, pages 219-235. Springer-Verlag

Berlin Heidelberg, 2010.




Traffic Monitoring in Cambridge

TrafficLight

Sensors Acis

Location Storage

RedlightBus Proximity

DB

Acis Location

Sensors

Motivating Example

vs





TrafficLight

Sensors Acis

Location Storage

RedlightBus Proximity

DB

Acis Location

Sensors

Motivating Example

Changing Usage

vs

Deployment and Sizing

vs

Prediction

• Processing Time

• Resource

Utilisation

Analyses

• Max. Throughput

• Bottleneck

• Event delivery latency

Design Alternatives

• One component per

• Traffic light

• Intersection

• District

• New components

• Speeding detection

vs




Preliminary Case Study - Application


GPS

Location

Data

SCOOT Traffic

Light

Status Proximity

Detection

ACIS Location

Storage

Output of TIME-EACM research project

Real world scenario

SBUS (Stream Bus) middleware

Supports RPC as well as event streams

Developed in Cambridge

Project website: http://www.cl.cam.ac.uk/research/time/




SBUS Middleware




Preliminary Case Study – System Model




Prediction of

Event processing time

CPU utilisation

Different scenarios

Several instances

Different deployment

Up to 4 quad-core machines

Variation of event rates

Prediction error

Mostly < 10%

Never exceeded 20%

CPU utilisation prediction

Processing time prediction

Preliminary Case Study – Evaluation




Preliminary Case Study – Evaluation (2)


Initial Modelling Effort Model Adaptation Effort



Capacity Planning for Event-based Systems

using Automated Performance Predictions

Christoph Rathfelder FZI, Germany

Samuel Kounev KIT, Germany

David Evans University of Cambridge

Event

C. Rathfelder, S. Kounev, and D. Evans. Capacity Planning for Event-based Systems using

Automated Performance Predictions. In 26th IEEE/ACM International Conference On

Automated Software Engineering (ASE 2011), pages 352-361, Oread, Lawrence, Kansas,

November 2011.




Event Bus

Bus

Sensors

Traffic

Control

License

Plate

RecognitionCamCam

Speeding

Toll

Location

Bus

Proximity

Motivating Example




Event Bus

Bus

Sensors

Traffic

Control

License

Plate

RecognitionCamCam

Speeding

Toll

Location

Bus

Proximity

Changing Load

vs.

System Evolution

• New components

• New algorithms

Vs.

Sizing and Capacity Planning

• Max. throughput

• Resource utilization

• Latency

• Bottlenecks

Over-provisioning of hardware resources

Motivating Example




Capacity Planning

Performance

Predictions

Result Evaluation

Model Adaptation

QoS requirements

fullfilled?

No

Yes

Variation of

Architecture/Deployment

Capacity Planning

System Evolution/

Workload Changes

System Deployment/

ReconfigurationResources used

efficiently?

Yes

System Modeling/

Resource Demand Estimation

End of life?

Yes

Capacity Planning using

Performance Predictions




3

2 1

Automated Performance Prediction Process

1. Systematic variation of simulated workload intensity

2. Automated model transformations

3. Simulation-based performance prediction using PCM

Middleware-

Weaving

M2M-Event-

Transformation

Parameter

Variation

Solving/

Simulation

Transformation

Prediction ModelEnd of

Parameter Range?

NoYes




Transformation

Example

A

B

C

Platform independent

C

Sink C

Event

Receiver

IMiddleware

Receiver

SinkPort

IMiddleware

SinkPort

IEvent

Sink

A

Source A

IEvent

Source

IEvent Receiver

IEvent

Receiver

SourcePort

Event

Distribution

IMiddleware

SourcePort

IMiddleware

EventDistribution

IMiddleware

Sender

Event

Sender Distribution

Preparation

IMiddleware

DistributionPreparation

Event

Sender

Platform specific

SBUS

SourcePort

SBUS

Middleware SBUS

SinkPort

Middleware-

Weaving

M2M-Event-

Transformation

Parameter

Variation

Solving/

Simulation

Transformation

Prediction ModelEnd of

Parameter Range?

NoYes




Case Study

Traffic Monitoring System

Based on output of TIME-EACM research project

Real world scenario with data from the City of Cambridge

SBUS (Stream Bus) middleware

Supports RPC as well as event streams

Developed in Cambridge

Scenarios

System variations

Evaluation of deployment options

Maximal throughput

Hardware utilization

Latency

Bottlenecks

ACIS

SCOOT

License Plate

Recognition Cam Cam Cam

Speeding Toll

Location

Bus Proximity




Specification of Architecture-level

Prediction Model

Component Repository

• Controlled experiments for each component

• Resource demand estimation based on time measurements

• Probabilistic and parameter dependent

System Model

• Instantiation and connection of components

• Variations depending on scenario (e.g., load balancing)

Deployment and Hardware

• Specification of hardware resources

• Deployment of components depending on scenario

Usage Model

• Automated variation of workload

1 Variant

3 Variants

7 Variants

> 100

Variants




Scenario 2 - Growing Workload

Adding additional cameras causes

additional workload

License Plate Recognition (LPR)

is bottleneck

Result: Small improvement by

deploying LPR on dedicated server

AllOnOne Distributed

Cam Cam Cam

ACIS

SCOOT

License

Plate


Speeding

Location

Bus

Proximity

Cam Cam Cam

ACIS

SCOOT

License

Plate


Speeding

Location

Bus

Proximity

10 8 6 4 2

020

40

60

80

100

Timespan between two images [s]

CP

U u

tiliz

ation [%

]

AllOnOne

Distrib. LPD

Distrib. Other

LPR

0.1 0.125 0,167 0.25 0.5

Image/event rate [1/s]




Scenario 3 - Additional Component

New Toll component

LPR is the bottleneck

2 additional servers

Load balancing on 3 LPR instances

Result: Centralized deployment

of Speeding and Toll

1 LPR 3 LPR cent. 3 LPR decent.

2.0 1.5 1.0 0.5

02

04

06

08

01

00


CP

U u

tiliz

atio

n [

%]

1 LPR: LPR3 LPR cent.: LPR3 LPR: decent.: LPRProc. decent.central

Cam Cam Cam

ACIS

SCOOT

License Plate


Speeding

Location

Bus Proximity

Toll

Cam Cam Cam

ACIS

SCOOT

License Plate


Speeding

Location

Bus Proximity

Toll

Cam Cam Cam

ACIS

SCOOT

License Plate


Speeding

Location

Bus Proximity

Toll

0.5 0.67 1.0 2.0





Scenario 4 – Improved Cameras

New cameras with higher resolution

Improved LPR success rate

Higher overall CPU demand for processing

Result: Decentralized deployment of Speeding and Toll

2.5 2.0 1.5 1.0 0.5

02

04

06

08

01

00


CP

U u

tiliz

atio

n [

%]

Cent. LPR

Decent. LPRCent. Proc.

2.5 2.0 1.5 1.0 0.5

02

04

06

08

01

00


CP

U u

tiliz

atio

n [

%]

Cent. LPRDecent. LPRCent. Proc.

Old cameras New cameras

0.4 0.5 0.67 1.0 2.0


0.4 0.5 0.67 1.0 2.0





Evaluation of Prediction Accuracy

Deployment of each scenario in our testbed

Load drivers with configurable event rate using real world data

Compare measured and predicted values in different load situations

S3

Experiment

Controller

S2S1

S9S8S7

S4 S5 S6

S10 S11 S12

Gigabit

Switch

Each machine equipped with:

Intel Core 2 Quad Q6600 2,4GHz,

8GB RAM, Ubuntu 8.04




Evaluation Results

Prediction error:

Utilization always underestimated

Mean error < 20%, Max error < 25%

0.5 1.0 1.5 2.0 2.5 3.0

02

04

06

08

01

00

Scenario3

CPU utilization

Frequency of images per Cam [1/s]

CP

U u

tiliz

atio

n

LPD Meas. (decent.)

LPD Pred. (decent.)LPD Meas. (cent.)LPD Pred. (cent.)Proc. Meas. (cent.)Proc. Pred. (cent.)

Scenario 3

1 2 3 4 5

02

04

06

08

01

00

Scenario 4

CPU utilization

Frequency of images per Cam [1/s]C

PU

utiliz

atio

n

Meas. (cent., old)Pred (cent., old)Meas. (decent., old)Pred. (decent., old)Meas. (cent., new)Pred. (cent., new)Meas. (decent., new)Pred. (decent., new)

Scenario 4

LPR LPR LPR LPR

Image/event rate [1/s] Image/event rate [1/s]

[%]

[%]




Effort Reduction

Architecture-level prediction models

• Eased variation of architecture and deployment

Automated model transformation for events

• 80% less manual element creations compared to manual modeling

Automation of performance predictions

• Time saving

• Prediction time: 3 min

• Experiment run time: 2.7 hours

• Automated load-variation




Conclusion

Capacity Planning

• Based on automated performance predictions

• Prediction error < 25%

• Always underestimated

• Improves the system‘s efficiency

• Often over-provisioning by factor 2 and more

Effort Reduction

• Modeling effort for event-based systems reduced by 80%

• Significant time saving by using prediction techniques




Agenda





Case Studies



Outlook





Architecture-level modeling language for self-aware

run-time systems management of modern IT systems,

infrastructures and services

Main Goal: Provide Quality-of-Service (QoS) guarantees

Performance (current focus)

Response time, throughput, scalability and efficiency

Or more generally, dependability

Including also availability, reliability and security aspects




1) Self-Reflective Aware of their software architecture, execution environment and

hardware infrastructure, as well as of their operational goals

2) Self-Predictive

Able to anticipate and predict the effect of dynamic changes in the environment, as well as the effect of possible adaptation actions

3) Self-Adaptive

Proactively adapting as the environment evolves to ensure that their operational goals are continuously met

http://www.descartes-research.net





Collection of several meta-models each focusing on different system aspects




PCM and DMM

Design-time aspects Run-time aspects

Palladio Component Model (PCM) Descartes Meta-Model (DMM)




Design-time vs. Run-time Models

Two orthogonal dimensions

Modeling of design-time vs. run-time aspects

Use of models at design-time vs. run-time

Fine granular differentiating factors

1. Model purpose

2. Model target users / consumers

3. Degrees of freedom in model use case scenarios

4. Model structure & parameterization

5. Possibilities for model calibration

6. Required model flexibility




Descartes vs. Palladio

PCM

DMM




Example Usage Scenarios

Example:

Design-time

QoS analysis

Example:

Run-time QoS

management

Example:

Elasticity

evaluation at

design-time

Example:

Design-time QoS

analysis in a DMM

resource landscape




DMM Technical Report

http://descartes.ipd.kit.edu/research_and_profile/descartes_meta_model/




Discussion

http://www.descartes-research.net/


modeling of event-based communication in component-based ...€¦ · palladio component model,”...

Documents