marin silic, goran delac and sinisa srbljic prediction of atomic web services reliability based on...

Marin Silic, Goran Delac and Sinisa Srbljic

Prediction of Atomic Web Services Reliability Based on K-means Clustering

Consumer Computing Laboratory

Faculty of Electrical

Engineering and Computing

University of Zagreb, Croatia

http://ccl.fer.hr/

http://www.fer.hr/

http://www.unizg.hr/

ESEC/FSE, Saint Petersburg, Russia, 2013.

http://ccl.fer.hr/

http://www.fer.hr/

http://www.unizg.hr/

Outline

Motivation

Reliability in SOA

State-of-the-art

CLUS Approach

Evaluation

Conclusion


Motivation

Contemporary web applications - SOA


A 2

A 1A 3

A 4

A 5

Web Application

A 2

A 1

A 3

A 4

Composite service

Process of candidates selection


A 2A 1

A 3A 4

A 5 A 6

Functional properties Nonfunctional properties

Ensure the desired functionality Reliability Availability…

Impact Qos & QoE

Repository

A 2

A 1

A 3

A 4

A 5

A 6

Service Oriented System

“Reliability on demand” definition


REQ

RES

𝑡𝑅𝐸𝑄𝑡𝑅𝐸𝑆

𝑡𝑅𝐸𝑆−𝑡𝑅𝐸𝑄<∆ 𝑡

𝑝𝑟=𝑁 𝑆

𝑁𝑇

The ratio of successful against

total number of invocations

Application

PastInvocation

Sample

Drawbacks/Obstacles

Client’s vs. provider’s perspective Service invocation context Depends on the quality of the sample Acquiring a sample proves to be a difficult task


A 2A 1 A 3

A 4A 5 A 6

QoS QoS QoS

QoS QoS QoS

A 1

QoS1

QoS2

QoS QoS1 QoS2≠≠

Service Provider

Client

Client

Insight to the Solution

To overcome the drawbacks and obstacles

Collect partial, but relevant past invocation sample

Utilize prediction methods to estimate the reliability for the missing records


State-of-the-art

Collaborative filtering


p1n?…p11 ? … p1i ??…? p22 … ? ………… … … … pun?…pu1 ? … pui ………… … … … pmn?…pm1 ? … pmi

m users

n services ??…?…?

ui matrix m,n >>matrix is extremely sparsenumber of values to predict

Collaborative filtering

Computes the similarity using PCC Matrix can be employed in two different ways


p1n?…p11 ? … p1i ??…? p22 … ? ………… … … … pun?…pu1 ? … pui ………… … … … pmn?…pm1 ? … pmi

UPCC approach

IPCC approachHybrid approach

Disadvantages of Collaborative Filtering

Scalability

Having millions of users and services – these approaches do not scale

Accuracy in dynamic environments

Internet is a highly dynamic systemDo not consider environment conditions


CLUStering

To address scalabilityApplies the principle of aggregationReduces the redundant data by clustering users

and services using K-means

To improve the accuracy Introduces environment-specific parametersDisperses the collected data across the

additional dimension


CLUS Overview


(1c)

(2c)

(5c)

Data Clustering Phase

r(u, s, t) p(r)

RawData

ClusteredData

Environment Clustering

UsersClustering

ServicesClustering

Creationof D

(3c)

(4c)

Prediction

Prediction Phase

Environment-specific Clustering

Set of environment conditions


𝐸={𝑒1 ,𝑒2 ,…,𝑒𝑖 ,…,𝑒𝑛}

t0 tct1 ti-1 ti tc-1w1 w2 wi wc

e1 e2 ei …… en

… …𝑝𝑤 𝑖

=1

¿𝑊 𝑖∨¿ ∑𝑟 ∈𝑊 𝑖

𝑝𝑟 ¿

𝑝𝑤1𝑝𝑤2

𝑝𝑤 𝑖𝑝𝑤𝑐

K-meansclusteringA day

User-specific Clustering

Set of users clusters


𝑈={𝑢1 ,𝑢2 ,…,𝑢𝑖 ,…,𝑢𝑚}

u1 u2 ui …… um

……

𝑝𝑟={𝑝𝑒1 ,𝑝𝑒2 ,…,𝑝𝑒𝑖 ,…,𝑝𝑒𝑛 }

e1 e2 ei …… en𝑝𝑒1𝑝𝑒2

𝑝𝑒𝑖𝑝𝑒𝑛

𝑝𝑟 𝑝𝑟 𝑝𝑟𝑝𝑟

𝑝𝑟 𝑝𝑟 𝑝𝑟 𝑝𝑟 𝑝𝑟

K-meansclustering

Service-specific Clustering

Set of services clusters


𝑆={𝑠1 ,𝑠2 ,…, 𝑠𝑖 ,…,𝑠𝑙 }

s1 s2 si …… sl

……e1 e2 ei …… en𝑝𝑒1𝑝𝑒2

𝑝𝑒𝑖𝑝𝑒𝑛 K-meansclustering

𝑝𝑟 𝑝𝑟 𝑝𝑟𝑝𝑟

𝑝𝑟 𝑝𝑟 𝑝𝑟 𝑝𝑟 𝑝𝑟

𝑝𝑟={𝑝𝑒1 ,𝑝𝑒2 ,…,𝑝𝑒𝑖 ,…,𝑝𝑒𝑛 }

Creation of Space D Each record, r(u, s, t), is associated to the

belonging clusters uk , sj , ei Each entry in D is computed as follows:

R contains all the records that belong to clusters uk , sj , ei


𝐷 [𝑢𝑘 ,𝑠 𝑗 ,𝑒𝑖 ]=1

¿𝑅∨¿ ∑𝑟 ∈𝑅

𝑝𝑟 ¿

Prediction

Assuming an ongoing rc=(uc, sc, tc) First, it checks the collected sample:

If H is not empty

Otherwise,


𝐻={𝑟 h∨𝑢𝑐=𝑢h⋀ 𝑠𝑐=𝑠h⋀ 𝑡𝑐 , 𝑡h∈𝑤𝑖 }

𝑝𝑐=1

¿𝐻∨¿ ∑𝑟 ∈𝐻

𝑝𝑟 ¿

𝑝𝑐=𝐷 [𝑢𝑘 , 𝑠 𝑗 ,𝑒𝑖 ] ,𝑢𝑐∈𝑢𝑘∧𝑠𝑐∈𝑠 𝑗∧𝑡𝑐∈𝑒𝑖

Evaluation

Comparison with the state-of-the-artUPCC IPCCHybrid

Evaluation measuresPrediction accuracy

oMAE, RMSEPrediction performance

oAggregated prediction time


Evaluation

Experiment setupAmazon EC2 Cloud


Data

Evaluation

Results – Impact of data densityPrediction accuracy – with load intensity


Evaluation

Results – Impact of data densityPrediction performance – with load intensity


Evaluation

Results – Impact of number of clustersPrediction accuracy, Data density = 20%


Evaluation

Results – Impact of number of clustersPrediction performance, Data density = 20%


Conclusion

Proposed a CLUS approach Improved the prediction accuracy

By introducing environment-specific parameters At least 56% lower RMSE value than the state-of-the-art

Improved the prediction performance By applying principle of aggregation Execution time reduced for two orders of magnitude when

compared to the state-of-the-art Flexibility of approach

Trade-off between accuracy and scalability Can be applied in different environments


Q&A

Thanks the audience for listening.


marin silic, goran delac and sinisa srbljic prediction of atomic web services reliability based on...

Documents

computed service reliability

saint petersburg

composite services

service oriented systems

adopted reliability

atomic web services

atomic reusable services

overall reliability