dynamic request allocation and scheduling for context aware applications subject to a percentile...

7/27/2019 Dynamic Request Allocation and Scheduling for Context Aware Applications Subject to a Percentile Response Time SLA in a Distributed Cloud

http://slidepdf.com/reader/full/dynamic-request-allocation-and-scheduling-for-context-aware-applications-subject 1/19

Dynamic request allocation and scheduling forcontext aware applications subject to a percentile

response time SLA in a distributed cloud

Keerthana Boloor∗, Rada Chirkova, Tiia Salo

and Yannis Viniotis∗

∗Department of Electrical and Computer EngineeringDepartment of Computer Science

North Carolina State University

IBM Software GroupResearch Triangle Park

1 / 17Cloudcom 2010, Indianapolis, Indiana, USA

http://find/

http://goback/



Agenda

Agenda

Problem description

Dynamic request allocation and scheduling scheme

Comparison with static allocation and FIFO/WeightedRound Robin scheduling scheme

Conclusion


http://find/



Problem description

Problem description

More web applications are designed to be context aware.

Most context aware applications are built on SOA

principles.

Cloud computing systems - the most preferred platform

for deployment.

Service Level Agreements (SLA) - terms of service andpricing model.

What is this presentation about?


http://find/



Problem description Geographically distributed cloud computing system

Geographically distributed cloud computing system

Clients

Data center hosting

K context-aware

applications

Data center hosting

K context-aware

applications

Data center hosting

K context-aware

applications

Data center hosting

K context-aware

applications


http://find/



Problem description Context aware applications

SOA based context aware application

Contextaware SOA

applications

End servers

Contextdata

stores

Gateway

2. Client request allocated to

and scheduled at end-server

3. Load requiredservice-endpoint

4. Load required

contextdata

DATA CENTER

1. Client request with

context-id

InternetUpdates to contexts at

contextdata stores


P bl d i i M d l f d

http://find/

http://goback/



Problem description Model of an end-server

An end-server serving multiple user classes

Server ‘j’ at data

center ‘i’

Class 1

Class 2

Class K

Each context aware application services multiple classes of users

Each user class is guaranteed different quality of service based on

economic considerations

SLA specifies different service levels and service charges for the

different user classes


P bl d i ti P til S i L l A t

http://find/



Problem description Percentile Service Level Agreements

Percentile Service Level Agreements

P

X 100

Profit

Conformance(%)

0

X % - the fraction of requests of a particular user class which need to have a response

time less than r seconds

$P - The profit charged by the cloud, if the percentile of requests that have response

time less than r seconds is greater than or equal to X %


Problem description Problem statement

http://find/




Problem statement

Allocate and schedule service requests locally at theend-servers so as to globally:

max

1≤ j ≤K

profit j (1)

where profit j is the profit charged for conformance of the

requests from users of class j .



http://find/




Problem statement

Allocate and schedule service requests locally at theend-servers so as to globally:

max

1≤ j ≤K

profit j (1)

where profit j is the profit charged for conformance of the

requests from users of class j .

This problem is NP-hard!!


Solution Management scheme description

http://find/

http://goback/




Heuristic-based data-oriented request management scheme

Periodic allocation and adaptation at each datacenter.

Al locat ion

phase

Al locat ion

phase

Adaptation

phase

Adaptation

phase

Observation interval (T)

subinterval

Al locat ion

phase

Al locat ion

phase

Al locat ion

phase

Adaptation

phase

Adaptation

phase

Adaptation

phase



http://find/




Heuristic-based data-oriented request management scheme

Periodic allocation and adaptation at each datacenter.

Al locat ion

phase

Al locat ion

phase

Adaptation

phase

Adaptation

phase

Observation interval (T)

subinterval

Al locat ion

phase

Al locat ion

phase

Al locat ion

phase

Adaptation

phase

Adaptation

phase

Adaptation

phase

Adaptation phase

Datacenters exchange conformance levels.

Allocation phase

Rank-based request allocation and gi-FIFO scheduling.

Aim at increasing global profit.


Solution Rank-based allocation and gi-FIFO scheduling

http://find/



Solution Rank based allocation and gi FIFO scheduling

Rank-based allocation and gi-FIFO scheduling

Profit-score calculation

Profit: p k

Required global conformance: c k

Current global conformance: cc k

If cc k < c k

Profit-score = p k /(c k −

cc k )

Else

Profit-score = 00 10 20 30 40 50 60 70 80 90 100

0

500

1000

1500

2000

Current conformance of class 1 (%)

P r o f i t − s c o r e

a s s i g n e d

t o

e a c h

a r r i v i n g

r e q u e s t o f c l a s s 1

( $ )

Class 1 SLA − Profit of 2000$ on conformance of 75%



http://find/



g g

Rank-based request allocation

1 Query hash-based lookup table ([context-id,machine-id] or [service-id,machine-id])

2 Rank-based compatibility test

1

The arriving request is assigned a rank based on its profit-score and deadline.

2 Does the arriving request meet its deadline? - Machine compatible!!!

3 Compatible machine not found? - Choose least loaded closest to context DB



http://find/



g g

gi-FIFO scheduling

Choose the request of user class with the highest current profit-score

Choose one with maximum waiting time but which results in a response time less than

or equal to r

If no such request exists, choose the request with higher waiting time resulting in a response time greater than r

gi-FIFO has been proven to be the most suitable for percentile SLAs for a single server

serving multiple classes.


Evaluation

http://find/



Evaluation

Dynamic scheme vs static schemes

5 10 15 20 25 30 35 40 45 500

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

11000

Request rate

P r o f i t i n c u r r e d

( $ )

Dynamic rank based allocation with gi−FIFO scheduling

Static allocation with WRR scheduling

Static allocation with FIFO scheduling


Evaluation

http://find/



Dynamic rank based allocation vs static allocation scheme

0 50 100 1500

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

11000

Request rate

P r o f i t i n c u r r e d

( $ )

Static allocation with gi−FIFO scheduling

Dynamic rank based allocation with gi−FIFO scheduling


Evaluation

http://find/



Variation in subinterval length

0 50 100 150 200 250 300 350 400 450 5000

2000

4000

6000

8000

10000

12000

14000

16000

18000

Subinterval period

P r o f i t o b

t a i n e d ( $ )

Uniform distribution of classes, stringent SLA

Uniform distribution of classes, relaxed SLA

Non−uniform distribution of classes, stringent SLA

Non−uniform distribution of classes, relaxed SLA

Variation in context update interval

0 20 40 60 80 100 120 140 160 180 2000

2000

4000

6000

8000

10000

12000

14000

16000

18000

Contextdata update interval

P r o f i t o b

t a i n e d

( $ )

Low contextdata load times

High contextdata load times

Medium Contextdata load times


Conclusion

http://find/

http://goback/



Conclusion

Identified the need for dynamic request scheduling and allocation for context aware

applications in a distributed cloud.

Proposed a novel rank-based request allocation and gi-FIFO scheduling scheme for

managing percentile SLAs with an aim to maximize profit obtained by the cloud.


http://find/



Questions??


http://find/

dynamic request allocation and scheduling for context aware applications subject to a percentile...

Documents