scispace - formally typeset

Journal ArticleDOI

A 2 thOS: availability analysis and optimisation in SLAs

01 Mar 2012-International Journal of Network Management (Wiley)-Vol. 22, Iss: 2, pp 104-130

TL;DR: This paper presents A2thOS, a framework to calculate the availability of partially outsourcing IT services in the presence of SLAs and to achieve a cost-optimal choice of availability levels for outsourced IT components while guaranteeing a target availability level for the service.

AbstractInformation technology (IT) service availability is at the core of customer satisfaction and business success for today's organisations. Many medium- to large-size organisations outsource part of their IT services to external providers, with service-level agreements describing the agreed availability of outsourced service components. Availability management of partially outsourced IT services is a non-trivial task since classic approaches for calculating availability are not applicable, and IT managers can only rely on their expertise to fulfil it. This often leads to the adoption of non-optimal solutions. In this paper we present A2thOS, a framework to calculate the availability of partially outsourced IT services in the presence of SLAs and to achieve a cost-optimal choice of availability levels for outsourced IT components while guaranteeing a target availability level for the service. Copyright © 2011 John Wiley & Sons, Ltd.

Topics: Service (business) (55%)

Summary (1 min read)

Jump to:  and [Introduction]

Introduction

  • A framework to calculate the availability of partially outsourced IT services in the presence of SLAs and to achieve a cost-optimal choice of availability levels for outsourced IT components while guaranteeing a target availability level for the service.the authors.
  • Figure 4 shows one possible scheduling for the failure of the components on which Service1 depends on, resulting in Service1 having an availability of αService1 (0.984).
  • To this end the authors distinguish among three types of nodes in a dependency graph: target availability nodes, variable availability nodes and given availability nodes.
  • The analysis engine solves the availability analysis problem, described in Section 3.

Did you find this useful? Give us your feedback

...read more

Content maybe subject to copyright    Report

A
2
THOS: Availability Analysis and Optimisation in SLAs
Emmanuele Zambon
1
, Sandro Etalle
1,2
and Roel J. Wieringa
1
1
University of Twente
Enschede, The Netherlands
Email: {emmanuele.zambon, sandro.etalle, r.j.wieringa}@utwente.nl
2
Technical University of Eindhoven
Eindhoven, The Netherlands
Email: s.etalle@tue.nl
SUMMARY
IT service availability is at the core of customer satisfaction and business success for today’s organisations. Many medium-large
size organisations outsource part of their IT services to external providers, with Service Level Agreements describing the agreed
availability of outsourced service components. Availability management of partially outsourced IT services is a non trivial task since
classic approaches for calculating availability are not applicable, and IT managers can only rely on their expertise to fulfil it. This
often leads to the adoption of non optimal solutions. In this paper we present A
2
THOS, a framework to calculate the availability of
partially outsourced IT services in the presence of SLAs and to achieve a cost-optimal choice of availability levels for outsourced
IT components while guaranteeing a target availability level for the service. Copyright
c
2010 John Wiley & Sons, Ltd.
KEY WORDS: SLA Management, Availability, Optimisation, Modelling
1. Introduction
Having a functional, cost effective and and properly managed IT infrastructure has become one of the main key success
factors for all kinds of organisations. Nowadays, the IT infrastructure of most large organisations is so complex that it is
often organised in terms of services that are offered as part of an internal market in which different business units offer
and buy IT services to and from each other. In some cases, services are acquired from an external organisation rather
than from an internal business unit (outsourcing). Typically, services offered by an internal provider are customised and
tailored to support the business goals of the organisation, while those offered by external providers are standardised and
large-scale, and therefore are less specific but potentially cheaper than those implemented internally. In some cases,
internal providers outsource some sub-services to external ones, for instance when it lacks specific competencies (e.g.,
SAP configuration). This is a so-called mixed sourcing strategy.
Regardless of whether the service is bought internally or externally, the terms and conditions of the contract are
determined in the so-called Service Level Agreement (SLA). (Figure 1 summarises the concept of mixed-sourced IT
services regulated by SLAs.) For instance, ITIL [15] is one of the most popular frameworks providing guidelines and
best practice for a correct IT service management and it describes this process in detail in [17].
In this paper we focus on IT service availability, which is at the core of customer satisfaction and business success
for organisations [16], and indeed it is one of the main topics in a SLA. In fact a typical SLA includes hard clauses on
the minimal availability of the service offered (for example, it may include that the service should not be “down” for
more than two hours per week, and a penalty fee for each week in which this is not satisfied).
Now, the two concerns we focus on (and at the same time the two questions to which we provide an answer within
the limits of the settings of this paper) are:
1. how can a business unit check and/or guarantee that a given (offered) service will respect some given minimal
availability levels;
2. as (1) while minimising costs.

Figure 1: Mixed-sourced IT service provision regulated by SLAs.
Let us elaborate on these two points and explain why they are not only relevant, but also non-trivial problems.
An IT service is usually offered by a system consisting of several components. These components can interact in
non-trivial ways: for instance a component could be crucial to the service in a way that if the component is unavailable
then the service becomes unavailable as well; other components my be organised in such a way (e.g., exploiting
redundancy) that only if a number of them fails the service will be affected. In addition, a component may depend in
a non-trivial way on sub-services which are in turn regulated by other SLAs.
To ensure that the minimal service availability remains within the agreed margins, IT managers can take reactive
(e.g., monitoring, measuring) and/or proactive measures. A key proactive measure is planning and designing service
availability when services are created or changed. At the business level, planning service availability allows the service
provider to set availability figures on the SLAs that both satisfy the customer needs and can be guaranteed by the
technical infrastructure providing the service. To achieve this at the technical level the service provider needs to
(a) calculate the availability of the IT system providing the service(s) based on the information available on system
components, and (b) make appropriate system design choices to support a specific availability level by selecting the
system components based on their contribution to the availability of the system.
Reliability studies have introduced a number of by now standard techniques (e.g., Continuous Time Markov Chains
(CTMC) [19] and Petri Nets [9]) which allow one to compute system availability when the mean time between
component failures and the mean time to repair a component is known. However, in the context of mixed-sourced
IT services, this information is usually not available. Instead, SLAs between the external and the internal provider
typically only include the minimal guaranteed availability of the component. Therefore, it is not possible to apply
these standard techniques to calculate the system availability (see Section 2 for details).
Regarding the second point, the service catalogue of most IT outsourcing companies include different availability
levels (e.g., gold, silver and bronze) with different associated prices (same service, only different availability levels, at
different costs). Service providers need to minimise the cost of outsourced (sub)services while guaranteeing that their
own service achieves the desired minimal availability level. Given the interactions mentioned above, this is a nontrivial
optimisation problem: one needs to determine the combination of minimal availability levels for the sub-services in
such a way that the total cost is minimal while ensuring that the resulting service achieves the availability specified in
the SLAs. This cannot be solved without the use of specific optimisation algorithms and typically IT managers choose
non-optimal, conservative solutions.
Contribution We present A
2
THOS, a framework for the analysis and optimisation of the availability of mixed-
sourced IT services. The framework consists of (1) a modelling technique to represent partially-outsourced IT systems,
their components and the services they provide, based on dependency graphs, (2) a procedure to calculate (a lower
bound of) the system availability given the (lower bounds of) components availability, and (3) a procedure to select
the optimum availability level for outsourced components in order to guarantee a desired target availability level for
the service(s) and to minimise costs.
1

A dependency graph is an AND/OR graph in which nodes represent system components and services, and edges
between nodes represent the functional dependency of one node with the other. We use the graph in order to calculate
a state function describing the availability of each service based on the state of the components (operational or not
operational). We then use the state function and the information about components availability to determine a lower
bound for the availability of the service, by setting up a linear programming problem. Based on this procedure, we
finally present the procedure to set up an integer programming problem which allows one to determine the cost-optimal
combination of availability levels for outsourced components in order to guarantee a target service availability. We
show the practical use of A
2
THOS by implementing it in a tool which we apply to the service availability planning of
an industrial case.
Limitation of the approach A
2
THOS uses an AND/OR graph to represent the system, thus it is unable to explicitly
represent failure recovery mechanisms such as spare parts. Spare parts are used to implement warm and cold standby
mechanisms. For example, to shorten the downtime caused by a server breakdown, the system administrators can keep
another server ready to replace the broken one. This second server is the spare part. When it is always running (but not
operating) and the workload of the broken server is automatically routed to the spare server, this mechanism is called
hot standby. When the workload of the broken server needs to be manually routed to the spare server, this mechanism
is called warm standby. When the spare server is not readily available, but it needs a setup phase before the workload
of the broken server can be redirected to it, the mechanism is called cold standby. Our representation allows us to
explicitly model hot standby mechanisms by using OR nodes, but it is not applicable in case of warm and cold standby
mechanisms. We share this limitation with other well-known modelling techniques, such as traditional Fault Trees and
Reliability Block Diagrams.
Organisation The rest of the paper is organised as follows. In Section 2 we present the related work in the fields of
reliability and IT service composition. In Section 3 we present dependency graphs and we provide the mathematical
foundation for using them to calculate service availability. In Section 4 we present the procedure to find the optimal
choice of availability level for outsourced components. In Section 5 we describe the tool we created to implement the
A
2
THOS framework and the benchmarks we conducted to test its scalability performances. Finally, in Section 6 we
show how we applied A
2
THOS to a practical case of service availability planning in an industrial context.
2. Related Works
In this section we discuss related works in four relevant areas for our problem: (1) the general approach to calculate
system availability, (2) modelling techniques to represent the system under analysis, (3) existing tools and (4) other
approaches taking into account availability to optimise IT service composition.
The general approach Referring to a classic formulation [2] taken from the reliability theory, a repairable system
is a system which can be repaired after a failure.
In the simplest case, the system m for which availability must be determined is represented by the state function
χ(m, t) which assumes value 1 if m is operating within tolerances at time t, 0 otherwise. The general way of calculating
the availability of a repairable system is to assume it has an independent, exponential distribution of failure and repair
time (a so-called stationary alternating renewal process [14]). However, to do so one must know at least two properties
of the system: its failure rate λ, and its repair rate µ. The first property specifies how often the system will fail on
average, i.e., its Mean Time Between Failure (MTBF): λ =
1
MTBF
. The second one specifies its Mean Time To Repair
(MTTR): µ =
1
MTTR
. Under this assumption the limiting availability is then obtained by the formula
¯
A =
µ
µ+λ
.
In the general case, the system can assume more than two states. Such a system is called complex. A complex
system is a system which is made of interconnected components that as a whole exhibit one or more properties
depending on the properties of the individual component. For example, a complex system can be made of two “simple”
components (i.e., two components that can independently be either in operative or in repairing state). The state of the
system depends on the state of the two components: the system may work properly even if one component only is
operative, or it may need both components to be operative. To model the state of the system, a state formula is used.
2

Components can have more than two states (e.g., operative, planned maintenance, emergency repair, etc.). To compute
the availability of complex systems, Continuous Time Markov Chains (CTMC) [19], or Petri Nets [9] are used. To
employ such techniques, one has to (1) define a state formula of the system based on the component’s state, and (2)
know the transaction probability of each component from one state to the other.
In our case, the information available in the SLAs for outsourced components concerns only a minimal availability
in a given time frame (e.g., one month). Therefore, classic techniques are not applicable to this problem, as the internal
states of each component and the probability of state transition (i.e., failure and repair rate) are only known by the
outsourcing company.
System modelling Several approaches have been proposed in the literature for system reliability modelling. Fault
trees (FTs) and Reliability Block Diagrams (RDBs) are the most used ones. However, we should mention that also
other approaches have been proposed, e.g., Torres-Toledano and Sucar [22] use bayesian networks, and Leangsuksun
et al. [13] use an UML representation (although in this second case the authors do not provide the mathematical
support for reliability analysis). In FTs, a number of components (called basic events) are linked together to make up a
system according to AND/OR relationships. The same behaviour is achieved in RBDs through SERIES/PARALLEL
compositions. According to [9], FTs are easy to use, as they do not require very skilled modellers, and relatively fast
to evaluate, as it is possible to use very efficient combinatorial solving techniques to obtain most of the reliability
indexes.
In FTs, the system state is represented by the top event, i.e., the root of the tree. It is possible to build a boolean
equation from the FT, and to reduce it to the minimal cut set, i.e., the smallest set of combinations of basic events
(component failures) which all need to occur for the top event to take place (system failure) [23]. Based on the
minimal cut set, a combination of combinatorial techniques and CTMC or PetriNets is then used to calculate the
system (limiting) availability.
According to Flamini et al. [9], the main limitation of FTs and RBDs consists in the lack of modelling power, as they
do not allow to model maintenance-related issues explicitly. To solve this problem, FTs and RDBs have been extended
into Dynamic Fault Trees [6] and Dynamic Reliability Block Diagrams [5], allowing one to model maintenance-related
issues.
The modelling notation we use in this paper (dependency graphs) can be seen as a condensed form of fault trees.
With a single dependency graph we are able to model a forest of fault trees sharing (some of) the basic events (i.e.,
the failure of a component), but with different top events. A single dependency graph can thus model separately the
failure of all the business services which the IT system provides, and for which a specific availability level must be
calculated. In fact, it is possible to (automatically) transform any dependency graph into a forest of FTs, as well as
in a set of RBD, as we show in Appendix B. We share with FTs the use of minimal cut sets, which in our notation
are called Dependency Sets (see Section 3), but the availability calculation we apply to dependency graphs is different
from the one used in FTs (for the reason we mentioned above).
Tools IBM Tivoli [12] and HP Business Availability Centre [11] are two of the most popular configuration
management tools. These tools are meant to support IT managers in the configuration and maintenance of complex IT
systems. Among the many features they possess, they can be used to manage SLAs, including availability levels. One
can assign to each IT component the availability level imposed by SLAs, and keep track of the actual availability levels
to check for SLA compliancy. However, to the best of our knowledge there is no support for the analytical calculation
of the service availability.
Galileo [21], Coral [4], Relex [18] and BlockSim [3] are tools operating with Dynamic Fault Trees. Although
integrating the A
2
THOS engines in one of these tools would be useful, this was not possible: Relex and BlockSim are
commercial tools, Coral is mostly a MatLab library without a GUI, and Galileo is free software, but not open source.
For these reasons we developed our prototype as an independent Java/Prolog tool.
Availability in service composition In the field of IT service composition, several approaches have been proposed
that consider availability as one of the QoS parameters to optimise the performances of the resulting composite IT
service. Gu et al. [10] propose QUEST, a framework to schedule dynamically a composite IT service while satisfying
QoS requirements (e.g., response time and availability) imposed by SLAs. Zeng et al. [26], Yu et al. [24] and Ardagna
3

et al. [1] propose scheduling techniques to create a cost-optimal execution plan for composite web services which
respect QoS parameters (including availability) defined in SLA contracts.
In all these works, an estimation of the availability of the composite service is made by multiplying the availability
level of the components (expressed as a real number in the interval [0,1]. This is possible thanks to two simplifying
assumptions. First, all the components must be available at the same time for the system to operate (i.e., the system
is an AND-combination of its components and it becomes unavailable in the moment that any of its component is
unavailable). Secondly, the resulting availability is not a lower bound, i.e., there can be a run of the composite service
in which the resulting availability is lower than the calculated one. Differently from these approaches, A
2
THOS is
able to deal with a wider range of dependencies, namely combinations of AND and OR dependencies. In the sequel
we also argue in more detail why OR dependencies are necessary to model complex IT services correctly. A
2
THOS
also allows one to calculate an absolute the lower bound for the availability, which can be safely included in an SLA
contract.
3. Analysis of the minimal service availability
We now present the theoretical foundations of A
2
THOS. Let us first start with an intuitive explanation. We model
the system using a dependency graph, in which a node represents a component of the system that at any given time
may (or may not) be available. A directed edge from node m to node n indicates that m depends on n, i.e. that the
availability of m depends also from the availability of n in a way that we are about to explain.
In a dependency graph, a node m can be unavailable because of an internal failure, or because (some) nodes it
depends on are unavailable. To model internal failure, to each node m we associate a (virtual) internal node m
0
.
On the other hand, to model the fact that m becomes unavailable because one or more nodes it depends on are
unavailable, we then consider nodes of two types: AND and OR .
(a) AND (b) OR
Figure 2: Two simple dependency graphs, respectively with AND and OR nodes
If m is a node in a dependency graph and n
1
, . . . , n
k
are the nodes m depends on, we say that
m is unavailable at time t iff its internal node m
0
is unavailable at time t or
n
1
, . . . , n
k
are all unavailable at time t, in case m is an AND node,
at least one node in n
1
, . . . , n
k
is unavailable at time t, in case m is an OR node.
Formally,
Definition 3.1 (Dependency graph) A dependency graph hN, Ei is a directed and acyclic graph (DAG) where N is
the set of nodes, and is partitioned in AND-N and OR-N, and E is the set of edges E {hu, vi | u, v N }.
Given a graph hN, Ei, we call N
0
the set of the internal nodes of g; N
0
= {n
0
internal of n | n N}.
Running example - Part 1. In this example we analyse the availability of an IT system providing two IT services
(Service1 and Service2), and implemented by means of three applications (App1, App2 and App3) running
on five different servers (Srv1, Srv2, Srv3, Srv4, Srv5). Service1 is implemented by App1 and App2 in
such a way that the service goes off-line only when both applications are off-line (OR dependency). Service2 is
4

Citations
More filters

DissertationDOI
20 Jan 2011
TL;DR: A graph-based framework for modelling the availability dependencies of the components of an IT infrastructure is proposed and techniques based on this framework are developed to support availability planning.
Abstract: The availability of an organisation’s IT infrastructure is of vital importance for supporting business activities. IT outages are a cause of competitive liability, chipping away at a company financial performance and reputation. To achieve the maximum possible IT availability within the available budget, organisations need to carry out a set of analysis activities to prioritise efforts and take decisions based on the business needs. This set of analysis activities is called IT availability planning. Most (large) organisations address IT availability planning from one or more of the three main angles: information risk management, business continuity and service level management. Information risk management consists of identifying, analysing, evaluating and mitigating the risks that can affect the information processed by an organisation and the information-processing (IT) systems. Business continuity consists of creating a logistic plan, called business continuity plan, which contains the procedures and all the useful information needed to recover an organisations’ critical processes after major disruption. Service level management mainly consists of organising, documenting and ensuring a certain quality level (e.g. the availability level) for the services offered by IT systems to the business units of an organisation. There exist several standard documents that provide the guidelines to set up the processes of risk, business continuity and service level management. However, to be as generally applicable as possible, these standards do not include implementation details. Consequently, to do IT availability planning each organisation needs to develop the concrete techniques that suit its needs. To be of practical use, these techniques must be accurate enough to deal with the increasing complexity of IT infrastructures, but remain feasible within the budget available to organisations. As we argue in this dissertation, basic approaches currently adopted by organisations are feasible but often lack of accuracy. In this thesis we propose a graph-based framework for modelling the availability dependencies of the components of an IT infrastructure and we develop techniques based on this framework to support availability planning.

65 citations


Journal ArticleDOI
TL;DR: A Petri net Monte Carlo simulation is developed that estimates the availability and costs of a specific design of an IT service redundancy allocation problem and two meta-heuristics, namely a genetic algorithm and tabu search, are adapted.
Abstract: The continuous provision of highly available IT services is a crucial task for IT service providers in order to fulfill service level agreements with customers. Although the introduction of redundant components increases availability, the associated cost may be very high. Therefore, decision makers in the IT service design stage face a trade-off between cost and availability in order to define suitable service level objectives. Although this task can be seen as a redundancy allocation problem, the existing definitions in this area are not transferable to IT service design due to the assumption of independent component failures, which has been identified as unrealistic in IT systems. In this paper, a multi-objective redundancy allocation problem for IT service design is defined. Therefore, a Petri net Monte Carlo simulation is developed that estimates the availability and costs of a specific design. In order to provide (sub)optimal solutions to an IT service redundancy allocation problem, two meta-heuristics, namely a genetic algorithm and tabu search, are adapted. The approach is utilized to optimize the IT service design of an application service provider in terms of availability and cost to demonstrate its feasibility and suitability.

27 citations


Journal ArticleDOI
01 Jan 2015
TL;DR: The approach is based on model-driven principles and uses both UML and Bayesian Networks to capture, analyse and optimise cloud deployment configurations and is extensible to the operational phases of the life-cycle.
Abstract: This paper proposes an approach to support cloud brokers finding optimal configurations in the deployment of dependability and security sensitive cloud applications. The approach is based on model-driven principles and uses both UML and Bayesian Networks to capture, analyse and optimise cloud deployment configurations. While the paper is most focused on the initial allocation phase, the approach is extensible to the operational phases of the life-cycle. In such a way, a continuous improvement of cloud applications may be realised by monitoring, enforcing and re-negotiating cloud resources following detected anomalies and failures.

18 citations


Journal ArticleDOI
TL;DR: The use of this language for allocating cloud resources to maximise service dependability by definition of a model-driven approach able to guide the software engineering to define a cloud infrastructure using a semi-automated process using both high-level languages such as UML as well as Bayesian networks.
Abstract: Bayesian networks have demonstrated their capability in several applications spanning from reasoning under uncertainty in artificial intelligence to dependability modelling and analysis. This paper focuses on the use of this language for allocating cloud resources to maximise service dependability. This objective is accomplished by the definition of a model-driven approach able to guide the software engineering to define a cloud infrastructure (applications, services, virtual and concrete resources) using a semi-automated process. This process exploits both high-level languages such as UML as well as Bayesian networks. Using all their features (backward analysis, ease of usage, low analysis time), Bayesian networks are used in this process as a driver for the optimization, learning and estimation phases. The paper discusses all the issues that the application of Bayesian networks in the proposed process arises.

3 citations


References
More filters

Journal ArticleDOI
Abstract: Agency theory is an important, yet controversial, theory. This paper reviews agency theory, its contributions to organization theory, and the extant empirical work and develops testable propositions. The conclusions are that agency theory (a) offers unique insight into information systems, outcome uncertainty, incentives, and risk and (b) is an empirically valid perspective, particularly when coupled with complementary perspectives. The principal recommendation is to incorporate an agency perspective in studies of the many problems having a cooperative structure.

10,675 citations


"A 2 thOS: availability analysis and..." refers background in this paper

  • ...If promise theory can be used to model existing SLAs, the problem of contract selection can be addressed by means of the principal–agent model [7,8]....

    [...]


Journal ArticleDOI
TL;DR: There is a comprehensive introduction to the applied models of probability that stresses intuition, and both professionals, researchers, and the interested reader will agree that this is the most solid and widely used book for probability theory.
Abstract: The Seventh Edition of the successful Introduction to Probability Models introduces elementary probability theory and stochastic processes. This book is particularly well-suited to those applying probability theory to the study of phenomena in engineering, management science, the physical and social sciences, and operations research. Skillfully organized, Introduction to Probability Models covers all essential topics. Sheldon Ross, a talented and prolific textbook author, distinguishes this book by his effort to develop in students an intuitive, and therefore lasting, grasp of probability theory. Ross' classic and best-selling text has been carefully and substantially revised. The Seventh Edition includes many new examples and exercises, with the majority of the new exercises being of the easier type. Also, the book introduces stochastic processes, stressing applications, in an easily understood manner. There is a comprehensive introduction to the applied models of probability that stresses intuition. Both professionals, researchers, and the interested reader will agree that this is the most solid and widely used book for probability theory. Features: * Provides a detailed coverage of the Markov Chain Monte Carlo methods and Markov Chain covertimes * Gives a thorough presentation of k-record values and the surprising Ignatov's * theorem * Includes examples relating to: "Random walks to circles," "The matching rounds problem," "The best prize problem," and many more * Contains a comprehensive appendix with the answers to approximately 100 exercises from throughout the text * Accompanied by a complete instructor's solutions manual with step-by-step solutions to all exercises New to this edition: * Includes many new and easier examples and exercises * Offers new material on utilizing probabilistic method in combinatorial optimization problems * Includes new material on suspended animation reliability models * Contains new material on random algorithms and cycles of random permutations

4,938 citations


Book
01 Jan 2001
TL;DR: This chapter discusses Decision-Theoretic Foundations, Game Theory, Rationality, and Intelligence, and the Decision-Analytic Approach to Games, which aims to clarify the role of rationality in decision-making.
Abstract: Preface 1. Decision-Theoretic Foundations 1.1 Game Theory, Rationality, and Intelligence 1.2 Basic Concepts of Decision Theory 1.3 Axioms 1.4 The Expected-Utility Maximization Theorem 1.5 Equivalent Representations 1.6 Bayesian Conditional-Probability Systems 1.7 Limitations of the Bayesian Model 1.8 Domination 1.9 Proofs of the Domination Theorems Exercises 2. Basic Models 2.1 Games in Extensive Form 2.2 Strategic Form and the Normal Representation 2.3 Equivalence of Strategic-Form Games 2.4 Reduced Normal Representations 2.5 Elimination of Dominated Strategies 2.6 Multiagent Representations 2.7 Common Knowledge 2.8 Bayesian Games 2.9 Modeling Games with Incomplete Information Exercises 3. Equilibria of Strategic-Form Games 3.1 Domination and Ratonalizability 3.2 Nash Equilibrium 3.3 Computing Nash Equilibria 3.4 Significance of Nash Equilibria 3.5 The Focal-Point Effect 3.6 The Decision-Analytic Approach to Games 3.7 Evolution. Resistance. and Risk Dominance 3.8 Two-Person Zero-Sum Games 3.9 Bayesian Equilibria 3.10 Purification of Randomized Strategies in Equilibria 3.11 Auctions 3.12 Proof of Existence of Equilibrium 3.13 Infinite Strategy Sets Exercises 4. Sequential Equilibria of Extensive-Form Games 4.1 Mixed Strategies and Behavioral Strategies 4.2 Equilibria in Behavioral Strategies 4.3 Sequential Rationality at Information States with Positive Probability 4.4 Consistent Beliefs and Sequential Rationality at All Information States 4.5 Computing Sequential Equilibria 4.6 Subgame-Perfect Equilibria 4.7 Games with Perfect Information 4.8 Adding Chance Events with Small Probability 4.9 Forward Induction 4.10 Voting and Binary Agendas 4.11 Technical Proofs Exercises 5. Refinements of Equilibrium in Strategic Form 5.1 Introduction 5.2 Perfect Equilibria 5.3 Existence of Perfect and Sequential Equilibria 5.4 Proper Equilibria 5.5 Persistent Equilibria 5.6 Stable Sets 01 Equilibria 5.7 Generic Properties 5.8 Conclusions Exercises 6. Games with Communication 6.1 Contracts and Correlated Strategies 6.2 Correlated Equilibria 6.3 Bayesian Games with Communication 6.4 Bayesian Collective-Choice Problems and Bayesian Bargaining Problems 6.5 Trading Problems with Linear Utility 6.6 General Participation Constraints for Bayesian Games with Contracts 6.7 Sender-Receiver Games 6.8 Acceptable and Predominant Correlated Equilibria 6.9 Communication in Extensive-Form and Multistage Games Exercises Bibliographic Note 7. Repeated Games 7.1 The Repeated Prisoners Dilemma 7.2 A General Model of Repeated Garnet 7.3 Stationary Equilibria of Repeated Games with Complete State Information and Discounting 7.4 Repeated Games with Standard Information: Examples 7.5 General Feasibility Theorems for Standard Repeated Games 7.6 Finitely Repeated Games and the Role of Initial Doubt 7.7 Imperfect Observability of Moves 7.8 Repeated Wines in Large Decentralized Groups 7.9 Repeated Games with Incomplete Information 7.10 Continuous Time 7.11 Evolutionary Simulation of Repeated Games Exercises 8. Bargaining and Cooperation in Two-Person Games 8.1 Noncooperative Foundations of Cooperative Game Theory 8.2 Two-Person Bargaining Problems and the Nash Bargaining Solution 8.3 Interpersonal Comparisons of Weighted Utility 8.4 Transferable Utility 8.5 Rational Threats 8.6 Other Bargaining Solutions 8.7 An Alternating-Offer Bargaining Game 8.8 An Alternating-Offer Game with Incomplete Information 8.9 A Discrete Alternating-Offer Game 8.10 Renegotiation Exercises 9. Coalitions in Cooperative Games 9.1 Introduction to Coalitional Analysis 9.2 Characteristic Functions with Transferable Utility 9.3 The Core 9.4 The Shapkey Value 9.5 Values with Cooperation Structures 9.6 Other Solution Concepts 9.7 Colational Games with Nontransferable Utility 9.8 Cores without Transferable Utility 9.9 Values without Transferable Utility Exercises Bibliographic Note 10. Cooperation under Uncertainty 10.1 Introduction 10.2 Concepts of Efficiency 10.3 An Example 10.4 Ex Post Inefficiency and Subsequent Oilers 10.5 Computing Incentive-Efficient Mechanisms 10.6 Inscrutability and Durability 10.7 Mechanism Selection by an Informed Principal 10.8 Neutral Bargaining Solutions 10.9 Dynamic Matching Processes with Incomplete Information Exercises Bibliography Index

3,474 citations


"A 2 thOS: availability analysis and..." refers background in this paper

  • ...This kind of conflict can be modelled using game theory [4]....

    [...]


Book
01 Jan 1991
Abstract: Preface 1. Decision-Theoretic Foundations 1.1 Game Theory, Rationality, and Intelligence 1.2 Basic Concepts of Decision Theory 1.3 Axioms 1.4 The Expected-Utility Maximization Theorem 1.5 Equivalent Representations 1.6 Bayesian Conditional-Probability Systems 1.7 Limitations of the Bayesian Model 1.8 Domination 1.9 Proofs of the Domination Theorems Exercises 2. Basic Models 2.1 Games in Extensive Form 2.2 Strategic Form and the Normal Representation 2.3 Equivalence of Strategic-Form Games 2.4 Reduced Normal Representations 2.5 Elimination of Dominated Strategies 2.6 Multiagent Representations 2.7 Common Knowledge 2.8 Bayesian Games 2.9 Modeling Games with Incomplete Information Exercises 3. Equilibria of Strategic-Form Games 3.1 Domination and Ratonalizability 3.2 Nash Equilibrium 3.3 Computing Nash Equilibria 3.4 Significance of Nash Equilibria 3.5 The Focal-Point Effect 3.6 The Decision-Analytic Approach to Games 3.7 Evolution. Resistance. and Risk Dominance 3.8 Two-Person Zero-Sum Games 3.9 Bayesian Equilibria 3.10 Purification of Randomized Strategies in Equilibria 3.11 Auctions 3.12 Proof of Existence of Equilibrium 3.13 Infinite Strategy Sets Exercises 4. Sequential Equilibria of Extensive-Form Games 4.1 Mixed Strategies and Behavioral Strategies 4.2 Equilibria in Behavioral Strategies 4.3 Sequential Rationality at Information States with Positive Probability 4.4 Consistent Beliefs and Sequential Rationality at All Information States 4.5 Computing Sequential Equilibria 4.6 Subgame-Perfect Equilibria 4.7 Games with Perfect Information 4.8 Adding Chance Events with Small Probability 4.9 Forward Induction 4.10 Voting and Binary Agendas 4.11 Technical Proofs Exercises 5. Refinements of Equilibrium in Strategic Form 5.1 Introduction 5.2 Perfect Equilibria 5.3 Existence of Perfect and Sequential Equilibria 5.4 Proper Equilibria 5.5 Persistent Equilibria 5.6 Stable Sets 01 Equilibria 5.7 Generic Properties 5.8 Conclusions Exercises 6. Games with Communication 6.1 Contracts and Correlated Strategies 6.2 Correlated Equilibria 6.3 Bayesian Games with Communication 6.4 Bayesian Collective-Choice Problems and Bayesian Bargaining Problems 6.5 Trading Problems with Linear Utility 6.6 General Participation Constraints for Bayesian Games with Contracts 6.7 Sender-Receiver Games 6.8 Acceptable and Predominant Correlated Equilibria 6.9 Communication in Extensive-Form and Multistage Games Exercises Bibliographic Note 7. Repeated Games 7.1 The Repeated Prisoners Dilemma 7.2 A General Model of Repeated Garnet 7.3 Stationary Equilibria of Repeated Games with Complete State Information and Discounting 7.4 Repeated Games with Standard Information: Examples 7.5 General Feasibility Theorems for Standard Repeated Games 7.6 Finitely Repeated Games and the Role of Initial Doubt 7.7 Imperfect Observability of Moves 7.8 Repeated Wines in Large Decentralized Groups 7.9 Repeated Games with Incomplete Information 7.10 Continuous Time 7.11 Evolutionary Simulation of Repeated Games Exercises 8. Bargaining and Cooperation in Two-Person Games 8.1 Noncooperative Foundations of Cooperative Game Theory 8.2 Two-Person Bargaining Problems and the Nash Bargaining Solution 8.3 Interpersonal Comparisons of Weighted Utility 8.4 Transferable Utility 8.5 Rational Threats 8.6 Other Bargaining Solutions 8.7 An Alternating-Offer Bargaining Game 8.8 An Alternating-Offer Game with Incomplete Information 8.9 A Discrete Alternating-Offer Game 8.10 Renegotiation Exercises 9. Coalitions in Cooperative Games 9.1 Introduction to Coalitional Analysis 9.2 Characteristic Functions with Transferable Utility 9.3 The Core 9.4 The Shapkey Value 9.5 Values with Cooperation Structures 9.6 Other Solution Concepts 9.7 Colational Games with Nontransferable Utility 9.8 Cores without Transferable Utility 9.9 Values without Transferable Utility Exercises Bibliographic Note 10. Cooperation under Uncertainty 10.1 Introduction 10.2 Concepts of Efficiency 10.3 An Example 10.4 Ex Post Inefficiency and Subsequent Oilers 10.5 Computing Incentive-Efficient Mechanisms 10.6 Inscrutability and Durability 10.7 Mechanism Selection by an Informed Principal 10.8 Neutral Bargaining Solutions 10.9 Dynamic Matching Processes with Incomplete Information Exercises Bibliography Index

3,003 citations


Journal ArticleDOI
TL;DR: This paper presents a middleware platform which addresses the issue of selecting Web services for the purpose of their composition in a way that maximizes user satisfaction expressed as utility functions over QoS attributes, while satisfying the constraints set by the user and by the structure of the composite service.
Abstract: The paradigmatic shift from a Web of manual interactions to a Web of programmatic interactions driven by Web services is creating unprecedented opportunities for the formation of online business-to-business (B2B) collaborations. In particular, the creation of value-added services by composition of existing ones is gaining a significant momentum. Since many available Web services provide overlapping or identical functionality, albeit with different quality of service (QoS), a choice needs to be made to determine which services are to participate in a given composite service. This paper presents a middleware platform which addresses the issue of selecting Web services for the purpose of their composition in a way that maximizes user satisfaction expressed as utility functions over QoS attributes, while satisfying the constraints set by the user and by the structure of the composite service. Two selection approaches are described and compared: one based on local (task-level) selection of services and the other based on global allocation of tasks to services using integer programming.

2,828 citations


"A 2 thOS: availability analysis and..." refers methods in this paper

  • ...[25], Yu and Lin [26] and Ardagna and Pernici [27] propose scheduling techniques to create a cost‐optimal execution plan for composite web services which respect QoS parameters (including availability) defined in SLA contracts....

    [...]


Frequently Asked Questions (1)
Q1. What are the contributions in "A2thos: availability analysis and optimisation in slas" ?

In this paper the authors present ATHOS, a framework to calculate the availability of partially outsourced IT services in the presence of SLAs and to achieve a cost-optimal choice of availability levels for outsourced IT components while guaranteeing a target availability level for the service.