game theoretic approaches for cloud computing

86
Game-theoretic Approaches for Modeling Cloud Environments Presented by: Ganesh Neelakanta Iyer CCWS -2012, Coimbatore Institute of Technology, Coimbatore, 9-August-2012 Wednesday, August 8, 12

Upload: ganeshn9

Post on 24-Jan-2015

1.305 views

Category:

Education


0 download

DESCRIPTION

Slides of the talk in Indo-US workshop on CCWS 2012

TRANSCRIPT

Page 1: Game theoretic approaches for Cloud Computing

Game-theoretic Approaches for Modeling Cloud EnvironmentsPresented by:Ganesh Neelakanta Iyer

CCWS -2012, Coimbatore Institute of Technology,Coimbatore, 9-August-2012

Wednesday, August 8, 12

Page 2: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

About Me

2

Three years of Industry work experience in Bangalore, India

Finished masters from National University of Singapore in 2008.

Submitted PhD thesis under the guidance of A/Prof. Bharadwaj Veeravalli: August 2012

Research interests: Cloud computing, Game theory, Wireless Networks, Pricing

Personal Interests: Kathakali, Teaching, Traveling, Photography, Cooking

Website: http://ganeshniyer.com

Wednesday, August 8, 12

Page 3: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

About Me

2

Three years of Industry work experience in Bangalore, India

Finished masters from National University of Singapore in 2008.

Submitted PhD thesis under the guidance of A/Prof. Bharadwaj Veeravalli: August 2012

Research interests: Cloud computing, Game theory, Wireless Networks, Pricing

Personal Interests: Kathakali, Teaching, Traveling, Photography, Cooking

Website: http://ganeshniyer.com

Wednesday, August 8, 12

Page 4: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Outline

• Overview of Cloud Computing and Major Challenges

• Overview of Game Theory

• Resource Allocation in Cloud - Bargaining theory

• Multiple Cloud Orchestration - Continuous Double Auctions

• Revenue Maximization on Mobile Clouds - Coalitional game theory

• Cloud Infrastructure Robustness and Security - Non-cooperative games

• Conclusions

3

Wednesday, August 8, 12

Page 5: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 20124

A quarter century ago, John Gage (Sun Microsystems) made the prophetic statement that:

“The network is the computer.”

Twenty-five years later, the advent of Cloud Computing has finally made this a reality.

http://www.tmforum.org/CloudServicesBrokerage/10617/home.html

http://blog.industrysoftware.automation.siemens.com/blog/tag/john-gage/http://historyofinformation.com/images/eniac.png

http://cloudcomputingcompaniesnow.com

Cloud Computing - A vision to reality

Wednesday, August 8, 12

Page 6: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 20124

A quarter century ago, John Gage (Sun Microsystems) made the prophetic statement that:

“The network is the computer.”

Twenty-five years later, the advent of Cloud Computing has finally made this a reality.

http://www.tmforum.org/CloudServicesBrokerage/10617/home.html

http://blog.industrysoftware.automation.siemens.com/blog/tag/john-gage/http://historyofinformation.com/images/eniac.png

http://cloudcomputingcompaniesnow.com

Cloud Computing - A vision to reality

Wednesday, August 8, 12

Page 7: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 20125

Definition of Cloud Computing

NIST defines Cloud Computing as1:

“Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.”

[1] P. Mell and T. Grance. The NIST definition of cloud computing. NIST Special Publication 800-145, 2011.

http://cloudcomputingcompaniesnow.com/

Wednesday, August 8, 12

Page 8: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 20125

Definition of Cloud Computing

NIST defines Cloud Computing as1:

“Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.”

[1] P. Mell and T. Grance. The NIST definition of cloud computing. NIST Special Publication 800-145, 2011.

http://cloudcomputingcompaniesnow.com/

Wednesday, August 8, 12

Page 9: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 20126

Gartner Hype cycle for Emerging Technologies: 2009-2011

http://www.gartner.com

Wednesday, August 8, 12

Page 10: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 20126

Gartner Hype cycle for Emerging Technologies: 2009-2011

http://www.gartner.com

Wednesday, August 8, 12

Page 11: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 20126

Gartner Hype cycle for Emerging Technologies: 2009-2011

http://www.gartner.com

Wednesday, August 8, 12

Page 12: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 20127

Gartner Hype cycle for Emerging Technologies:

2009-2011

2009

2011

2010

http://www.gartner.com

Wednesday, August 8, 12

Page 13: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Characteristics of Cloud...

8

Wednesday, August 8, 12

Page 14: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Characteristics of Cloud...

8

Elastic Computing

Wednesday, August 8, 12

Page 15: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Characteristics of Cloud...

8

Elastic Computing

On-demand availability

Wednesday, August 8, 12

Page 16: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Characteristics of Cloud...

8

Elastic ComputingPay-as-you-go

On-demand availability

Wednesday, August 8, 12

Page 17: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Characteristics of Cloud...

8

Elastic ComputingPay-as-you-go

On-demand availability

Do your business

Wednesday, August 8, 12

Page 18: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Characteristics of Cloud...

8

Elastic ComputingPay-as-you-go

Different Services

On-demand availability

Do your business

Wednesday, August 8, 12

Page 19: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Challenges in moving into the Cloud

9

http://www.accenture.com/us-en/outlook/Pages/outlook-online-2011-challenges-cloud-computing.aspx

Wednesday, August 8, 12

Page 20: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Challenges in moving into the Cloud

• Which CSP best matches my requirement?

9

http://www.accenture.com/us-en/outlook/Pages/outlook-online-2011-challenges-cloud-computing.aspx

Wednesday, August 8, 12

Page 21: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Challenges in moving into the Cloud

• Which CSP best matches my requirement?

• How secure is to move my data/job into a Cloud?

9

http://www.accenture.com/us-en/outlook/Pages/outlook-online-2011-challenges-cloud-computing.aspx

Wednesday, August 8, 12

Page 22: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Challenges in moving into the Cloud

• Which CSP best matches my requirement?

• How secure is to move my data/job into a Cloud?

• How trust worthy are the CSPs?

9

http://www.accenture.com/us-en/outlook/Pages/outlook-online-2011-challenges-cloud-computing.aspx

Wednesday, August 8, 12

Page 23: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Challenges in moving into the Cloud

• Which CSP best matches my requirement?

• How secure is to move my data/job into a Cloud?

• How trust worthy are the CSPs?

• How easy is to deal with lock-in?

9

http://www.accenture.com/us-en/outlook/Pages/outlook-online-2011-challenges-cloud-computing.aspx

Wednesday, August 8, 12

Page 24: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Challenges faced by the providers

10

Wednesday, August 8, 12

Page 25: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Challenges faced by the providers

• How to offer the right price to increase the revenue?

10

Wednesday, August 8, 12

Page 26: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Challenges faced by the providers

• How to offer the right price to increase the revenue?

• How to manage the resources efficiently?

10

Wednesday, August 8, 12

Page 27: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Challenges faced by the providers

• How to offer the right price to increase the revenue?

• How to manage the resources efficiently?

• How do I know the behavior of my competitors?

10

Wednesday, August 8, 12

Page 28: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Challenges faced by the providers

• How to offer the right price to increase the revenue?

• How to manage the resources efficiently?

• How do I know the behavior of my competitors?

• How to manage mobile applications on Mobile Clouds?

10

Wednesday, August 8, 12

Page 29: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Overview of Game Theory 11

Raffles Place, Singapore

Wednesday, August 8, 12

Page 30: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Game Theory

• Study of how people interact and make decisions

• “…Game Theory is designed to address situations in which the outcome of a person’s decision depends not just on how they choose among several options, but also on the choices made by the people they are interacting with…”

• The study of strategic interactions among economic (rational) agents and the outcomes with respect to the preferences (or utilities) of those agents

12

Wednesday, August 8, 12

Page 31: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

What is a game?

13

Wednesday, August 8, 12

Page 32: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

What is a game?

A Game consists ofat least two players a set of strategies for each playera preference relation over possible outcomes

13

Wednesday, August 8, 12

Page 33: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

What is a game?

A Game consists ofat least two players a set of strategies for each playera preference relation over possible outcomes

Player is general entityindividual, company, nation, protocol, animal, etc

13

Wednesday, August 8, 12

Page 34: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

What is a game?

A Game consists ofat least two players a set of strategies for each playera preference relation over possible outcomes

Player is general entityindividual, company, nation, protocol, animal, etc

Strategiesactions which a player chooses to follow

13

Wednesday, August 8, 12

Page 35: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

What is a game?

A Game consists ofat least two players a set of strategies for each playera preference relation over possible outcomes

Player is general entityindividual, company, nation, protocol, animal, etc

Strategiesactions which a player chooses to follow

Outcomedetermined by mutual choice of strategies

13

Wednesday, August 8, 12

Page 36: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

What is a game?

A Game consists ofat least two players a set of strategies for each playera preference relation over possible outcomes

Player is general entityindividual, company, nation, protocol, animal, etc

Strategiesactions which a player chooses to follow

Outcomedetermined by mutual choice of strategies

Preference relationmodeled as utility (payoff) over set of outcomes

13

Wednesday, August 8, 12

Page 37: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Game Theory: Applications

• Economics: Oligopoly markets, Mergers and acquisitions pricing, auctions

• Political Science: fair division, public choice, political economy

• Biology: modeling competition between tumor and normal cells, Foraging bees

• Sports coaching staffs: run vs pass or pitch fast balls vs sliders

• Computer Science: Distributed systems, Computer Networks, AI, scheduling

14

http://customergauge.com/wordpress/wp-content/uploads/2008/10/power_retailers_oligopoly.jpghttp://cricketradius.com/wp-content/uploads/2011/11/fast-bowling.jpg

Wednesday, August 8, 12

Page 38: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Prisoner’s Dilemma

• Two suspects arrested for a crime

• Prisoners decide whether to confess or not to confess

• If both confess, both sentenced to 3 months of jail

• If both do not confess, then both will be sentenced to 1 month of jail

• If one confesses and the other does not, then the confessor gets freed (0 months of jail) and the non-confessor sentenced to 9 months of jail

• What should each prisoner do?

15

Wednesday, August 8, 12

Page 39: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Prisoner’s Dilemma: Revisited

16

Wednesday, August 8, 12

Page 40: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Prisoner’s Dilemma: Revisited

• Two suspects arrested for a crime

16

Pris

oner

2

Prisoner 1

Wednesday, August 8, 12

Page 41: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Prisoner’s Dilemma: Revisited

• Two suspects arrested for a crime

• Prisoners decide whether to confess or not to confess

16

Confess Not Confess

Confess

Not Confess

Pris

oner

2

Prisoner 1

Wednesday, August 8, 12

Page 42: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Prisoner’s Dilemma: Revisited

• Two suspects arrested for a crime

• Prisoners decide whether to confess or not to confess

• If both confess, both sentenced to 3 months of jail

16

Confess Not Confess

Confess

Not Confess

Pris

oner

2

Prisoner 1

-3,-3

Wednesday, August 8, 12

Page 43: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Prisoner’s Dilemma: Revisited

• Two suspects arrested for a crime

• Prisoners decide whether to confess or not to confess

• If both confess, both sentenced to 3 months of jail

• If both do not confess, then both will be sentenced to 1 month of jail

16

Confess Not Confess

Confess

Not Confess

Pris

oner

2

Prisoner 1

-3,-3

-1,-1

Wednesday, August 8, 12

Page 44: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Prisoner’s Dilemma: Revisited

• Two suspects arrested for a crime

• Prisoners decide whether to confess or not to confess

• If both confess, both sentenced to 3 months of jail

• If both do not confess, then both will be sentenced to 1 month of jail

• If one confesses and the other does not, then the confessor gets freed (0 months of jail) and the non-confessor sentenced to 9 months of jail

16

Confess Not Confess

Confess

Not Confess

Pris

oner

2

Prisoner 1

-3,-3

-1,-1-9,0

0,-9

Wednesday, August 8, 12

Page 45: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Prisoner’s Dilemma: Revisited

• Two suspects arrested for a crime

• Prisoners decide whether to confess or not to confess

• If both confess, both sentenced to 3 months of jail

• If both do not confess, then both will be sentenced to 1 month of jail

• If one confesses and the other does not, then the confessor gets freed (0 months of jail) and the non-confessor sentenced to 9 months of jail

• What should each prisoner do?

16

Confess Not Confess

Confess

Not Confess

Pris

oner

2

Prisoner 1

-3,-3

-1,-1-9,0

0,-9

Wednesday, August 8, 12

Page 46: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Prisoner’s Dilemma: Nash Equilibrium

• Each player’s predicted strategy is the best response to the predicted strategies of other players

• No incentive to deviate unilaterally

• Strategically stable or self-enforcing

17

Confess Not Confess

Confess

Not Confess

Pris

oner

2

Prisoner 1

-3,-3-1,-1-9,0

0,-9

Wednesday, August 8, 12

Page 47: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Rock-paper-scissors game

• A probability distribution over the pure strategies of the game

• Rock-paper-scissors game

• Each player simultaneously forms his or her hand into the shape of either a rock, a piece of paper, or a pair of scissors

• Rule: rock beats (breaks) scissors, scissors beats (cuts) paper, and paper beats (covers) rock

• No pure strategy Nash equilibrium

• One mixed strategy Nash equilibrium – each player plays rock, paper and scissors each with 1/3 probability

18

Wednesday, August 8, 12

Page 48: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Resource Allocation in Cloud Computing Envirnments

19Ulu Watu, Bali, Indonesia

Wednesday, August 8, 12

Page 49: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Resource Allocation in Cloud

Problem  under  considera0on  is  “Resource  Alloca,on  and  Pricing  Strategies  for  tasks  in  Compute  Cloud  Environments”.  

We  employ  “Axioma,c  Bargaining  Approaches  to  derive  the  op,mal  solu,on  for  alloca,ng  resources  in  a  Compute  Cloud”.  

•  Nash  Bargaining  Solu0on  (NBS)  and  Raiffa  Bargaining  Solu0on  (RBS)•    Handling  various  parameters  such  as  deadline,  budget    constraints  etc

•    Introduc0on  of  asymmetric  pricing  scheme  for  CSPs

•    Handling  auto-­‐elas0city,  fairness

20

Reference: Ganesh Neelakanta Iyer and Bharadwaj Veeravalli, “On the Resource Allocation and Pricing Strategies in Compute Clouds Using Bargaining Approaches”, IEEE International Conference on Networks (ICON 2011), Singapore, December 2011.

Wednesday, August 8, 12

Page 50: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Resource Allocation in Cloud

21

Resource(Allocator(

1 2

i

Rtot

Compute(Node(i"

Internet(Task(1(

Task(1(

Task(T(

Compute(Cloud(Environment(Suitable  for  both  independent  tasks,  Bag-­‐of-­‐Tasks  (BoT)  and  tasks  from  workflow  schemes

Assump?on:  Tasks  are  known  apriori,  but  it  can  handle  real-­‐?me  arrival  of  tasks

Coopera?ve  game  theory  framework

Reference: Ganesh Neelakanta Iyer and Bharadwaj Veeravalli, “On the Resource Allocation and Pricing Strategies in Compute Clouds Using Bargaining Approaches”, IEEE International Conference on Networks (ICON 2011), Singapore, December 2011.

Wednesday, August 8, 12

Page 51: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Axiomatic Bargaining Approaches

Good to derive fair and Pareto-optimal solution

Pareto optimal: It is impossible to increase the allocation of a connection without strictly decreasing another one.

It assumes some desirable and fair properties, defined using axioms, about the outcome of the resource bargaining process.

Two approaches:

Nash Bargaining Solution (NBS)

Raiffa-Kalai-Smorodinsky Bargaining Solution (RBS)

22

Reference: Ganesh Neelakanta Iyer and Bharadwaj Veeravalli, “On the Resource Allocation and Pricing Strategies in Compute Clouds Using Bargaining Approaches”, IEEE International Conference on Networks (ICON 2011), Singapore, December 2011.

Wednesday, August 8, 12

Page 52: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Axiomatic Bargaining Approaches

Nash Bargaining Solution (NBS)

23

Solving, we obtain

Reference: Ganesh Neelakanta Iyer and Bharadwaj Veeravalli, “On the Resource Allocation and Pricing Strategies in Compute Clouds Using Bargaining Approaches”, IEEE International Conference on Networks (ICON 2011), Singapore, December 2011.

Wednesday, August 8, 12

Page 53: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Axiomatic Bargaining Approaches

Raiffa-Kalai-Smorodinsky Bargaining Solution (RBS)

24

Solving, we obtain

Reference: Ganesh Neelakanta Iyer and Bharadwaj Veeravalli, “On the Resource Allocation and Pricing Strategies in Compute Clouds Using Bargaining Approaches”, IEEE International Conference on Networks (ICON 2011), Singapore, December 2011.

Wednesday, August 8, 12

Page 54: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Resource Allocation in Cloud

25

Performance evaluation:

Deadline based Real-time task arrival

Reference: Ganesh Neelakanta Iyer and Bharadwaj Veeravalli, “On the Resource Allocation and Pricing Strategies in Compute Clouds Using Bargaining Approaches”, IEEE International Conference on Networks (ICON 2011), Singapore, December 2011.

Wednesday, August 8, 12

Page 55: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Resource Allocation in Cloud

26

Pricing AnalysisSymmetric:

price/resource = $0.75

Asymmetric:

A value in [0.5,1.0]

Tasks specify maximum budget

Current CSPs follow symmetric pricing schemes (EC2, Azure)

Introducing asymmetric pricing approach, which would give adequate flexibility in managing the resources as well as generating more revenue.

Reference: Ganesh Neelakanta Iyer and Bharadwaj Veeravalli, “On the Resource Allocation and Pricing Strategies in Compute Clouds Using Bargaining Approaches”, IEEE International Conference on Networks (ICON 2011), Singapore, December 2011.

Wednesday, August 8, 12

Page 56: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Resource Allocation in Cloud

27

Observations:

•Allocation in NBS and RBS depends on bargaining power and is within the Pareto boundary

•When NBS maximizes the product of the gain of all players, RBS in addition considers how much other players gave up

•NBS efficiently utilizes maximum number of resources

•RBS indirectly maps to an energy efficient solution by meeting the deadline with less number of resources.

•RBS effectively handles auto-elasticity and task dynamics

•NBS is shown to be suitable for shorter deadline tasks whereas RBS is for handling tasks of longer deadline tasks.

•Asymmetric pricing scheme

Reference: Ganesh Neelakanta Iyer and Bharadwaj Veeravalli, “On the Resource Allocation and Pricing Strategies in Compute Clouds Using Bargaining Approaches”, IEEE International Conference on Networks (ICON 2011), Singapore, December 2011.

Wednesday, August 8, 12

Page 57: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Multiple Cloud Orchestration

28Melaca, MalaysiaWednesday, August 8, 12

Page 58: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Cloud Orchestration

29

• Relates to the connectivity of IT and business process levels between Cloud environments.

• As cloud emerges as a competitive sourcing strategy, a demand is clearly arising for the integration of Cloud environments to create an end-to-end managed landscape of cloud-based functions.

• Benefits include

• Helps users to choose the best service they are looking for (for example the cheapest or the best email provider)

• Helps providers to offer better services and adapt to market conditions quickly

• Ability to create a best of breed service-based environment in which a change of provider does not break the business process

http://lookout.atos.net/en-us/enabling_information_technologies/cloud_orchestration/default.htm

Reference: Ganesh Neelakanta Iyer, Bharadwaj Veeravalli and Ramkumar Chandrasekaran, “Broker-agent based Cloud Service Arbitrage Mechanisms using Sealed-bid Double Auctions and Incentives”, Journal of Network and Computer Applications (JNCA), Elsevier 2012

Wednesday, August 8, 12

Page 59: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Cloud Brokers

• Cloud Broker plays an intermediary role to help customers locate the best and the most cost-effective CSP for the customer needs

• One stop solution for Multiple Cloud Orchestration (aggregating, integrating, customizing and governing Cloud services for SMEs and large enterprises)

• Advantages are cost savings, information availability and market adaptation

• As the number of CSPs continues to grow, a single interface (Broker) for information, combined with service, could be compelling to companies that prefer to spend more time with their Clouds than doing the research.

• Some ways to implement :- Auctions, Incentives

30

Reference: Ganesh Neelakanta Iyer, Bharadwaj Veeravalli and Ramkumar Chandrasekaran, “Broker-agent based Cloud Service Arbitrage Mechanisms using Sealed-bid Double Auctions and Incentives”, Journal of Network and Computer Applications (JNCA), Elsevier 2012

Wednesday, August 8, 12

Page 60: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Typical Cloud Broker ecosystem showing the

players involved The Broker helps to connect the providers and users

31

Requirements and Architecture of a Cloud Broker

© OPTIMIS Consortium Page 3 of 27

3 Use Case Scenario The use of cloud based services in order to provide online services to customers is expected to

bring in a new era in the area of ICT infrastructure and delivery. In a simplistic scenario, a

service provider (SP) decides to host the service it wants to provide to an end customer on the

infrastructure provided by an infrastructure provider (IP). This is the current state of art in the

use of the cloud services and is considered as a simple scenario because of the limited

flexibility available to the SP to split its services into components and deploy them into

infrastructures provided by multiple IPs.

However it is expected that the growth and maturity of the cloud offerings would necessitate

the building of delivery models that will allow use of multiple IPs for the hosting of multiple

components in a single service. It is anticipated that the evolution of such a delivery model

would then lead to the formation of a service entity know as the Cloud Broker that provides

the SPs, at minimum, with a mechanism to choose a group of IPs from a list of available ones

for deploying various components of its service based on various parameters. This Work

Package considers the concept of this Cloud Broker in detail.

3.1 Storyboard

The main aim of WP 6.4 is to showcase the use of the OPTIMIS toolkit to build up a cloud

services brokerage ecosystem that allows the Service Provider (SP) to use multiple

Infrastructure services provided by respective Infrastructure Providers (IPs) by integrating

them in a way to as to implement a singular service or process. The aim is to utilize the various

components of the toolkit and thus leverage the work done in other work packages of the

project. While the initial WP’s  Description  of  work  (DoW)  describes three scenario setups with

varying levels of complexity, in this deliverable we mould the presentation to concentrate on

the scenario specific to the cloud broker and use the other two scenarios as stepping stones

and stretch goals of the WP.

The Cloud Broker (CB) can be considered as an architectural, business and IT operations model

that enables the delivery and management of different cloud services in a framework that

provides consistent provisioning, security, administration and other support. In this use case,

an SP planning to deploy a service in the cloud approaches a CB with a given set of functional

requirements and constraints (including costs, performance etc.) with the aim of selecting the

best available match of IPs in terms of the functional requirements as well as other variable

constraints like cost, SLA parameters and other non-functional requirements like audit,

compliance and security capabilities.

Programmer

Cloud Broker

IDflex

IDbt

Flexiant

Users Users

Identity Brokerage

Entitlement Mgmt.

Policy Enforcement

Usage Monitoring,

Reporting

Admin

BT

Network defense,

Platform security

Service Provider

ARSYS

IDarsys

Figure 1: Cloud Broker ecosystem showing the players involved.

http://www.optimis-project.eu/

Wednesday, August 8, 12

Page 61: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Auction Theory

• In economic theory, an auction may refer to any mechanism or set of trading rules for exchange.

• English Auction:open ascending price auction. • Dutch Auction:open descending price auction. • Vickery Auction: Sealed-bid second price auction• First Price auction: Highest bidder pays the price they submitted• Call Market: Mediator determines market clearing price based on number of bid and ask orders.• CDA: Continuous Double Auctions

32

Auc$ons(

Single( Double(

English( Dutch( First(Price( Call(Market( CDA(

Outcry( Outcry(Sealed=bid( Sealed=bid(

Vickery(

Wednesday, August 8, 12

Page 62: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Sealed-bid Continuous Double Auctions

CDA: Continuous Double Auctions

The Continuous Double Auction (CDA) is a mechanism to match buyers and sellers of a particular good, and to determine the prices at which trades are executed. Instead, in non-institutional trade-determination, buyers and sellers can choose to accept a bid or ask, and then update their allocation, at any point in time.

33Reference: Ganesh Neelakanta Iyer, Bharadwaj Veeravalli and Ramkumar Chandrasekaran, “Broker-agent based Cloud Service Arbitrage Mechanisms using Sealed-bid Double Auctions and Incentives”, Journal of Network and Computer Applications (JNCA), Elsevier 2012

Wednesday, August 8, 12

Page 63: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Sealed-bid Continuous Double Auctions

Comparison of revenue

Hit Ratio is the ratio of the number of successful auctions to the total number of auctions.

Fair revenue for all users

Lowers user expenditure at the expense of response-time for choosing appropriate CSP.

34

Reference: Ganesh Neelakanta Iyer, Bharadwaj Veeravalli and Ramkumar Chandrasekaran, “Broker-agent based Cloud Service Arbitrage Mechanisms using Sealed-bid Double Auctions and Incentives”, Journal of Network and Computer Applications (JNCA), Elsevier 2012

Wednesday, August 8, 12

Page 64: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Revenue Maximization in Mobile Clouds

35From my home, Thodupuzha, Kerala

Wednesday, August 8, 12

Page 65: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Mobile Cloud Environments

Mobile cloud computing combines wireless access service and cloud computing to improve the performance of mobile applications.

Mobile applications can offload some computing modules (such as online gaming) to be executed on a powerful server in a cloud.

A scenario where multiple CSPs cooperatively offer mobile services to users.

Coalition games

36

Reference: Dusit Niyato, Ping Wang, Ekram Hossain, Walid Saad, and Zhu Han, “Game Theoretic Modeling of Cooperation amongService Providers in Mobile Cloud Computing Environments”, IEEE Wireless Communications and Networking Conference, 2012

Wednesday, August 8, 12

Page 66: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Coalition Game: An example

Players = {1,2,3}

All nonempty subset (named as coalition) {1}, {2}, {3}, {1,2}, {1,3}, {2,3}, {1,2,3}

A cost function c related to all coalitions. c({1}) = v1, c({2}) = v2, ..., c({1,2,3}) = v7

c(S) is the amount that the players in the coalition S have to pay collectively in order to have access to a service.

37

Reference: Dusit Niyato, Ping Wang, Ekram Hossain, Walid Saad, and Zhu Han, “Game Theoretic Modeling of Cooperation amongService Providers in Mobile Cloud Computing Environments”, IEEE Wireless Communications and Networking Conference, 2012

Wednesday, August 8, 12

Page 67: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Coalition Game: Core

•The problem is to find the core of this coalition game.

•Core is a cost distribution of the grand coalition such that no other coalition can obtain an outcome better for all its members than the current assignment.

•There may not exist any core.

•Emptiness of the core.

•There may exist many cores.

•Some players would unhappy with the cost allocation.

38

Reference: Dusit Niyato, Ping Wang, Ekram Hossain, Walid Saad, and Zhu Han, “Game Theoretic Modeling of Cooperation amongService Providers in Mobile Cloud Computing Environments”, IEEE Wireless Communications and Networking Conference, 2012

Wednesday, August 8, 12

Page 68: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Coalition Game: Example

We want to find the cost allocation {x1, x2, x3} such that

x1+x2+x3 = c({1,2,3})

x1 ≦ c({1})x2 ≦ c({2})x3 ≦ c({3})x1+x2 ≦ c({1, 2})x1+x3 ≦ c({1, 3})x2+x3 ≦ c({2, 3})

Given a solution in the core, there is no incentive for a player to leave the grand coalition.

39

Reference: Dusit Niyato, Ping Wang, Ekram Hossain, Walid Saad, and Zhu Han, “Game Theoretic Modeling of Cooperation amongService Providers in Mobile Cloud Computing Environments”, IEEE Wireless Communications and Networking Conference, 2012

Wednesday, August 8, 12

Page 69: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Mobile Clouds and Coalition Game

•Mobile applications are supported by the mobile CSPs in which the radio (bandwidth) and computing (servers) resources are reserved for the users.

•To improve resource utilization and revenue, mobile CSPs cooperate to form a coalition and create a resource pool for users running mobile applications.

•Revenue sharing among the CSPs is based on a coalitional game.

•With a coalition, providers can optimize the capacity expansion, which determines the reserved bandwidth and servers for a resource pool.

•The objective of provider is to maximize the profit from supporting mobile applications through a resource pool.

40

Reference: Dusit Niyato, Ping Wang, Ekram Hossain, Walid Saad, and Zhu Han, “Game Theoretic Modeling of Cooperation amongService Providers in Mobile Cloud Computing Environments”, IEEE Wireless Communications and Networking Conference, 2012

Wednesday, August 8, 12

Page 70: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Cyber-Physical SystemsRobustness

41

Phang Nga Bay, Thailand

Wednesday, August 8, 12

Page 71: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Attack and defense in cyber-physical systems

42

Reference: Chris Y. T. Ma, Nageswara S. V. Rao and David K. Y. Yau, “A Game Theoretic Study of Attack and Defense in Cyber-Physical Systems”, The First IEEE International Workshop on Cyber-Physical Networking Systems, 2011

Wednesday, August 8, 12

Page 72: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Attack and defense in cyber-physical systems

Cyber physical systems :- Systems which need cyber and physical components to function.

42

Reference: Chris Y. T. Ma, Nageswara S. V. Rao and David K. Y. Yau, “A Game Theoretic Study of Attack and Defense in Cyber-Physical Systems”, The First IEEE International Workshop on Cyber-Physical Networking Systems, 2011

Wednesday, August 8, 12

Page 73: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Attack and defense in cyber-physical systems

Cyber physical systems :- Systems which need cyber and physical components to function.

Examples: Cloud Computing systems, Sensor network systems, Communication networks

42

Reference: Chris Y. T. Ma, Nageswara S. V. Rao and David K. Y. Yau, “A Game Theoretic Study of Attack and Defense in Cyber-Physical Systems”, The First IEEE International Workshop on Cyber-Physical Networking Systems, 2011

Wednesday, August 8, 12

Page 74: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Attack and defense in cyber-physical systems

Cyber physical systems :- Systems which need cyber and physical components to function.

Examples: Cloud Computing systems, Sensor network systems, Communication networks

Players: Defenders aim to keep the system functioning and the attacker aims to disrupt.Actions represent the resources deployed by the defender and disrupted by the attacker, respectively.

42

Reference: Chris Y. T. Ma, Nageswara S. V. Rao and David K. Y. Yau, “A Game Theoretic Study of Attack and Defense in Cyber-Physical Systems”, The First IEEE International Workshop on Cyber-Physical Networking Systems, 2011

Wednesday, August 8, 12

Page 75: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Attack and defense in cyber-physical systems

Cyber physical systems :- Systems which need cyber and physical components to function.

Examples: Cloud Computing systems, Sensor network systems, Communication networks

Players: Defenders aim to keep the system functioning and the attacker aims to disrupt.Actions represent the resources deployed by the defender and disrupted by the attacker, respectively.

Costs and benefits: Each player has a payoff function U consisting of two parts: benefit B and/or cost C. The attacker incurs a cost in launching an attack, and the defender incurs a cost in deploying the resources. In a game, either player will aim to maximize its payoff given the other player's best strategy.

42

Reference: Chris Y. T. Ma, Nageswara S. V. Rao and David K. Y. Yau, “A Game Theoretic Study of Attack and Defense in Cyber-Physical Systems”, The First IEEE International Workshop on Cyber-Physical Networking Systems, 2011

Wednesday, August 8, 12

Page 76: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Attack and defense in cyber-physical systems

Cyber physical systems :- Systems which need cyber and physical components to function.

Examples: Cloud Computing systems, Sensor network systems, Communication networks

Players: Defenders aim to keep the system functioning and the attacker aims to disrupt.Actions represent the resources deployed by the defender and disrupted by the attacker, respectively.

Costs and benefits: Each player has a payoff function U consisting of two parts: benefit B and/or cost C. The attacker incurs a cost in launching an attack, and the defender incurs a cost in deploying the resources. In a game, either player will aim to maximize its payoff given the other player's best strategy.

Existence and solutions of pure and mixed-strategy Nash Equilibria can be found

42

Reference: Chris Y. T. Ma, Nageswara S. V. Rao and David K. Y. Yau, “A Game Theoretic Study of Attack and Defense in Cyber-Physical Systems”, The First IEEE International Workshop on Cyber-Physical Networking Systems, 2011

Wednesday, August 8, 12

Page 77: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

Summary...43

Water-puppetry, Vietnam

Wednesday, August 8, 12

Page 78: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

To Summarize...

44

Wednesday, August 8, 12

Page 79: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

To Summarize...

• Bargaining theory, Auction Theory, Coalition games, Non-cooperative games etc.

44

Wednesday, August 8, 12

Page 80: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

To Summarize...

• Bargaining theory, Auction Theory, Coalition games, Non-cooperative games etc.

• Resource allocation, Cloud orchestration, Robustness, Security, Mobile Clouds etc.

44

Wednesday, August 8, 12

Page 81: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

To Summarize...

• Bargaining theory, Auction Theory, Coalition games, Non-cooperative games etc.

• Resource allocation, Cloud orchestration, Robustness, Security, Mobile Clouds etc.

• Different aspects of Game Theory can be applied for tackling various problems in Cloud Computing environments

44

Wednesday, August 8, 12

Page 82: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

To Summarize...

• Bargaining theory, Auction Theory, Coalition games, Non-cooperative games etc.

• Resource allocation, Cloud orchestration, Robustness, Security, Mobile Clouds etc.

• Different aspects of Game Theory can be applied for tackling various problems in Cloud Computing environments

• Topics not covered (much more than what is discussed)

44

Wednesday, August 8, 12

Page 83: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

To Summarize...

• Bargaining theory, Auction Theory, Coalition games, Non-cooperative games etc.

• Resource allocation, Cloud orchestration, Robustness, Security, Mobile Clouds etc.

• Different aspects of Game Theory can be applied for tackling various problems in Cloud Computing environments

• Topics not covered (much more than what is discussed)

Repeated games, Dynamic games, Bayesian games, Combinatorial auctions .......

44

Wednesday, August 8, 12

Page 84: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

To Summarize...

• Bargaining theory, Auction Theory, Coalition games, Non-cooperative games etc.

• Resource allocation, Cloud orchestration, Robustness, Security, Mobile Clouds etc.

• Different aspects of Game Theory can be applied for tackling various problems in Cloud Computing environments

• Topics not covered (much more than what is discussed)

Repeated games, Dynamic games, Bayesian games, Combinatorial auctions .......

Energy minimization, Reliability, Trust and Risk modeling in Clouds......

44

Wednesday, August 8, 12

Page 85: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012Wednesday, August 8, 12

Page 86: Game theoretic approaches for Cloud Computing

©All Rights Reserved, Ganesh Neelakanta Iyer August 2012

THANK YOU!Wednesday, August 8, 12