agent-based federated hybrid cloud prof. yue-shan chang distributed & mobile computing lab....

46
Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National Taipei University

Upload: helena-rose-miller

Post on 12-Jan-2016

214 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Agent-based Federated Hybrid Cloud

Prof. Yue-Shan ChangDistributed & Mobile Computing Lab.

Dept. of Computer Science & Information EngineeringNational Taipei University

Page 2: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Cloud computing: – The evolution and convergence of computing trends– Layers

• SaaS: Software As A Service• PaaS: Platform As A Service• IaaS: Infrastructure As A Service

Introduction

112/04/212

Page 3: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Introduction

• Types– Private Cloud

• each enterprise’s IT platform has their own network, servers and storage hardware (Data Centers)

– Public Cloud• User can obtain any service and resource from service

provider• pay-per-use charging model

Page 4: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Introduction

112/04/214

• Which one is suitable? – Considering Issues

• Cost (Construction, Operation, Maintenance, Tax …)• Security (Data, Network, …)• Flexibility & Convenience (Operation, Maintenance, Management, …)• Reliability & Availability• Performance

Page 5: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Main benefits of using a public cloud service: – Easy and inexpensive set-up because hardware,

application and bandwidth costs are covered by the provider.

– Scalability to meet needs. – No wasted resources because you pay for what

you use.

Introduction

112/04/215

Page 6: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Introduction

112/04/216 112/04/216

What kind of cloud do I

need?Private?

Public?

Page 7: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• A hybrid cloud – is a cloud computing

environment in which an organization provides and manages some resources in-house and has others provided externally.

Introduction

112/04/217

HiCloudHiCloud

Amazon Amazon

GoogleGoogle

Public CloudPrivate Cloud

Page 8: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Effectively utilize public cloud resource is an important issue while adopting hybrid cloud– what kind of jobs need to be dispatched or be

migrated to public cloud? – When does a job be dispatched to public cloud? – And how will a job be dispatched to public cloud?

Introduction

112/04/218

Page 9: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Hybrid Cloud Project– ITRI Cloud Center– f5 Hybrid Cloud Architecture

• http://www.f5.com/pdf/solution-center/vmware-vcloud-director.pdf

– Fujitsu Hybrid Cloud• Mikio Funahashi, Shigeo Yoshikawa “Fujitsu’s Approach to

Hybrid Cloud Systems,” Fujitsu Sci. Tech. J., Jul. 2011, Vol. 47, No.3, pp. 285-292

– IBM Hybrid Cloud • IBM Service Management Extensions for Hybrid Cloud• http://public.dhe.ibm.com/common/ssi/ecm/en/

ibd03004usen/IBD03004USEN.PDF

Introduction

112/04/219

Page 10: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• ITRI Hybrid Cloud Architecture

Introduction

112/04/2110

Public CloudPrivate Cloud

Page 11: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• f5 Hybrid Cloud Architecture – http://www.f5.com/pdf/solution-center/vmware-vcloud-director.pdf

Introduction

112/04/2111

Page 12: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Fujitsu’s Approach to Hybrid Cloud Systems– Mikio Funahashi, Shigeo Yoshikawa “Fujitsu’s Approach to Hybrid Cloud Systems,” Fujitsu Sci. Tech. J.,

Jul. 2011, Vol. 47, No.3, pp. 285-292

Introduction

112/04/2112

Page 13: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Agent & Grid Computing– Ian Foster addressed that agent technology and

grid computing need each other because agent technology can enhance the ability of problem solving of grid.

• Agent & Cloud computing– More and more research adopting agent

technology to solve problems faced in the cloud

Introduction

112/04/2113

Page 14: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Propose an automatic, intelligent framework based on agent technology.

• A federated layer to tie private and public cloud.

• Mobile agent technique is exploited – manage all resources, – monitor system behaviour, – negotiate all actions

Introduction

112/04/2114

Page 15: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Objective– For performance issue

• Load balance

– For cost issue• utilize private cloud as much as possible• if private cloud cannot complete user’s job before

deadline (Deadline-constraint Job)– dispatch the job to public cloud

» minimize the required resource of the VM

Introduction

112/04/2115

Page 16: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Agent-based Federated Broker

112/04/2116

Page 17: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Agent-based Federated Broker

• Five major components– System Monitoring Agent (SMA)

• Collects the system information

– Reconfiguration Decision Agent (RDA)• Reconfigure and adjust the cloud environment.

– Service Dispatching Agent (SDA) • assign a location in the cloud that allows the job to be

executed on. • if some clusters are overloading, SDA will notify some JAs to

migrate to some other cluster, to balance the load.

112/04/2117

Page 18: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Agent-based Federated Broker

– Cluster Management Agent (CMA) • schedules jobs locally in a FCFS fashion, so that there is only

one job is executing on the cluster. • reports the status of the cluster• collects the information and send it via heartbeats to SeMA.

– Job Agent (JA) • encapsulates a job, the job can be migrated along with the

JA. • executes and monitors the job on the cluster. • reports the job status to the CMA periodically. • brings the results back to the private cloud.

112/04/2118

Page 19: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Job dispatching – Pack a job into Job Agent(JA)– dispatching JA to destination– Unpack the JA

Agent-based Federated Broker

112/04/2119

Page 20: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Job Count (JC), –

• Tc: Job count threshold

• the SMA will pick up the (JCPU +TC+1)th job from job queue of private cloud, and trigger it to be migrated.

• For example, if the JCPR is equal to 10, the JCPU is equal to 4, and the TC is equal to 2. Therefore, the 7th job will be migrated to public cloud.

Policy of Job Dispatching to Public Cloud

112/04/2120

CPUPR T-JCJC

Page 21: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• total Size of Job (SJ),–

• the SMA will pick up the th job from the job queue of private cloud, and trigger it to be migrated.

• For example, if the total size of job in public cloud is 10Mbytes, the TS is equal to 2Mbytes, and the size of jobs in private cloud are 3, 4, 3, 3, 2, 3, 4 Mbytes respectively. The 5th job (2Mbytes) will be migrated to public cloud because the (3+3+4+3); so that the 5th job will be migrated.

Policy of Job Dispatching to Public Cloud

112/04/2121

n

i

m

kki TSJSJ

1S

1

11

S

m

kk TSJ

Ts: the threshold of SJ

Page 22: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Estimated Finish Time (EFT)–

• the SMA will pick the th job in the queue of private cloud, and trigger it to be migrated.

• For example, if the total finish time of jobs in public cloud is 100s, the TT is equal to 20s, and the finish time of jobs in private cloud are 33, 24, 45, 43, 22, 37, 24 second respectively. The 5th job (22s of finish time) will be migrated to public cloud

– Rough Set Theory

Policy of Job Migration

112/04/2122

n

iT

m

kki TTT

1 1

11

T

m

kk TT

Page 23: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Agent Platform for the hybrid cloud

Prototyping and evaluation

112/04/2123

Page 24: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Job migrated

Prototyping and evaluation

112/04/2124

Page 25: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Comparison between with migration and without migration

Evaluation

112/04/2125

Page 26: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Comparison between job count and total size of job

Evaluation

112/04/2126

Page 27: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• an agent-based automatic intelligent job migration framework on a hybrid cloud is proposed.– built a prototype that integrating our private cloud

with public cloud.

• We demonstrate the job migration mechanism on Hadoop platform– it shows that the framework can be applied to

hybrid cloud and work well.

Summary

112/04/2127

Page 28: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Execution Time Prediction Using Rough Set Theory in

Hybrid Cloud

Page 29: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Introduction

• Resource utilization is important issue in cloud computing– Could the remaining resource in private cloud serve

the incoming task and complete the task before deadline?

– If not, the incoming task need to be dispatched to public cloud.

• How much resource we need to preserve to serve the deadline-constraint task in public cloud?

• For the remaining resource , the execution time prediction of a task becomes an important issue in hybrid cloud.

112/04/2129

Page 30: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Introduction

• Exploit Rough Set Theory (RST) to predict job's execution time in the hybrid cloud environment. – RST is a well-known prediction technique that uses

the historical data to predict the attribute value of an object.

– We propose an execution time prediction algorithm based on RST to schedule jobs

• The evaluation show that the RST can be utilized to accurately predict the execution time while historical data is increasingly.

112/04/2130

Page 31: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

RST-based Prediction

• Rough Set Theory (RST)– have been witnessed that is a useful prediction technique

based on historical data in a variety of applications,• such as quantitative structure–activity relationship in the

Chemistry and data mining . – It provides an appropriate theory for identifying good

“similarity templates”. • The primary objective of similarity templates is to identify

characteristics of applications that define similarity.

• Two prediction phases– Inference rule deducing phase – Estimation phase

112/04/2131

Page 32: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

RST-based Prediction

• Inference rule deducing phase – Steps (detailed methodology of RST can refer to [2])

• Define all attributes; including condition attributes (CA) and decision attributes (DA).

• Discretize the properties of historical records for diversified attributes.

• Calculate D-Reducts– Utilize discernibility matrix to list all properties, – apply discernibility function to formulate the relation of the

properties, – and then simplify the formulation using boolean algebra.

• Derive the inference rule of DA. .

112/04/2132

Page 33: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

RST-based Prediction

• Define all attributes

112/04/21 33

Decision AttributeConditional Attributes

Page 34: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

RST-based Prediction

• Discretize the properties of historical records

112/04/2134

Page 35: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

RST-based Prediction

• Calculate D-Reducts and D-Core– Generate discernibility matrix

112/04/2135

Page 36: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

RST-based Prediction

• Calculate D-Reducts and D-Core– Formulate discernibility function: fA(D)

112/04/2136

Both {a1, a3} and {a2,a3 } are D-Reducts, {a3} is D-core

Page 37: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

RST-based Prediction

• Calculate D-Reducts and D-Core– formulate the relation of the properties, and

simplify the formulation

112/04/2137

f2(D)=a1, f3(D)=a1+a3 , f4(D)=a1+a3 , …

Page 38: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

RST-based Prediction

• Deduce Inference Rule (- : means don’t care)– a1=2 -> d=2

– a1=3-> d=1

– a3=4 -> d=4

– a1=1 and a3=2

-> d=2

112/04/2138

Page 39: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

RST-based Prediction

• Estimation Phase– Apply simple mathematical operation, such as arithmetic

average of the value of DA, to obtain the final value of the DA.» Estimated time = (job3+job5+job6)/3

112/04/2139

Element Processor Speed

Input size Execution time

3 5 2 480

5 5 2 500

6 5 2 505

The new job 5 2 ?

Page 40: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Prototyping and Evaluation

• Prototype the system using the agent platform JADE v4.0

112/04/2140

Page 41: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Prototyping and Evaluation

• two jobs are submitted to the system– Compute π– Area Approximation

112/04/2141

Page 42: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Prototyping and Evaluation

• The Error Rate– Positive->over

prediction, – Negative->under

predicted. • Vibration during the

first 25 jobs. • lack of the historical data

that can be used to predict the job.

– The more the historical data are stored, the more accurate the prediction will be.

112/04/2142

1 57 1139 17 25 33 41 49 65 73 81 89 97 105 121129137145153161169177185193-1

-0.5

0

0.5

1

1.5

2

2.5

3

3.5

Compute π

Area Approximation

Job #

Err

or R

ate

Page 43: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Prototyping and Evaluation

• Absolute Error Rate. – shows how much

improvement has the prediction made.

• The higher the absolute error is, the more improvement is needed.

– for 2 kinds of jobs with 200 submissions are 0.2008 and 0.0615.

– the accuracy is very impressive if remove the first 25 predictions

112/04/2143

1 57 1139 17 25 33 41 49 65 73 81 89 97 105 1211291371451531611691771851930

0.5

1

1.5

2

2.5

3

3.5

Compute π

Area Approximation

Job #

Abs

olut

e E

rror

Rat

e

Page 44: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Prototyping and Evaluation

• The largest prediction latency is 642.91 ms with 190 jobs– is acceptable.

• no new record to be updated, the prediction time taken can be less than 1 ms.

• generating the decision rule needs much more time than just predicting the value.

• To reduce the time of predicting, – periodically updating the

decision rules can be considered.

112/04/2144

3 102960 117 174 231 288 345 402 459 516 573 630 687 744 801 858 915 9720

100

200

300

400

500

600

700

0

20

40

60

80

100

120

140

160

180

200

No. of job in history

estimated time

Estimation #

Mill

isec

ond

Job

Num

ber

Page 45: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

Summary

• we utilized the RST to predict the execution time in hybrid cloud.

• The result shows that RST-based predictor can predict the execution time of a job – error rate under 0.1 when the number of historical job is

over 50. – When more records available, the error rate can drop

under 0.03. • Latency is reasonable,

– less than 1 second with 190 historical records to perform a full prediction. The system can aid users to schedule their jobs faster and more accurate.

112/04/2145

Page 46: Agent-based Federated Hybrid Cloud Prof. Yue-Shan Chang Distributed & Mobile Computing Lab. Dept. of Computer Science & Information Engineering National

• Please refer to – http://youtu.be/4w6YohBJ8mo

Demo

112/04/2146