SLA Design and Service Provisioning for Outsourced Services
TL;DR: A main conclusion is that the SLA established using the business-driven perspective is superior to the one based on a conventional approach since both service provider and client can simultaneously obtain higher profit.
Abstract: In this work, a business-driven approach to designing Service Level Agreements in an e-commerce environment is proposed. In contrast to conventional SLA design approaches, the one proposed better captures the relationship between service provider and service client by considering the negative business impact (business loss) originated from IT infrastructure failures and performance degradation and introduces such knowledge into the SLA itself. A complete example scenario shows the value of the proposed approach. A main conclusion is that the SLA established using the business-driven perspective is superior to the one based on a conventional approach since both service provider and client can simultaneously obtain higher profit.
Citations
More filters
01 Jan 2013
TL;DR: A qualitative study interviewing industry experts is conducted to understand the extent to which SLA specifications in traditional IT services outsourcing can be applied in cloud computing, identifying elements becoming redundant or not applicable, and new elements capturing the specificity of the business relationships entailed by cloud computing.
Abstract: Issues characterizing cloud computing such as security, interoperability, and vendor lock-in are likely to act as inhibitors of the widespread adoption of cloud computing in organizations. These issues call for relational and contractual mechanisms to articulate the desired outcomes of service provisioning and acceptable behavior of service recipients and providers. In this paper we focus on the contractual mechanisms for this purpose and, more specifically, Service Level Agreements (SLA). We have conducted a qualitative study interviewing industry experts to understand the extent to which SLA specifications in traditional IT services outsourcing can be applied in cloud computing, identifying elements becoming redundant or not applicable, and new elements capturing the specificity of the business relationships entailed by cloud computing. Our study results in a detailed SLA specification for different cloud computing deployment models, i.e. private and public cloud, and in a framework investigating the role of SLA elements in the decision making process for cloud computing outsourcing.
31 citations
Cites background from "SLA Design and Service Provisioning..."
...…SLAs only as one of the constructs, besides for instance security and performance (Hofmann and Woods 2010, Patel et al. 2009), size of outsourced IT (Misra and Mondal 2011), or IT failure and performance degradation (Marques et al. 2009), 7 determining the success of outsourcing relationships....
[...]
TL;DR: It is concluded that the log-normal distribution offers the best fit of IT service time to recovery, and using this distribution in simulation and decision-support tools offers the prospect of better predictions of downtime and downtime costs to the practitioner community.
Abstract: The context of this article is the availability of enterprise IT services, a key concern for many enterprises. While there is a plethora of literature concerned with service availability, there is no previous systematic empirical study on IT service time to recovery following outages. The existing literature typically assumes a distribution, or builds on analogies to related areas such as software engineering. Therefore, our objective is to find the statistical distribution of IT service time to recovery. Method-wise, this investigation is based on logs of more than 1 $\thinspace$ 800 incidents in a large Nordic bank, corresponding to more than 11 $\thinspace$ 000 hours of recorded downtime. Five possible distributions of time to recovery from the literature were investigated using the Akaike Information Criterion to find the distribution offering the best fit. The results show that the log-normal distribution outperformed the others for all tested service channels (collections of IT services). It is concluded that the log-normal distribution offers the best fit of IT service time to recovery. Using this distribution in simulation and decision-support tools offers the prospect of better predictions of downtime and downtime costs to the practitioner community.
30 citations
Cites background from "SLA Design and Service Provisioning..."
...While means suffice for some agreements, agreements such as having 95 percent of requests served within a certain time, require knowledge about the full distribution [37]....
[...]
TL;DR: The main result is that enterprise IT professionals fail to maximize expected value when procuring availability SLAs, and more extensive use of decision-support systems might be called for in order to facilitate proper risk management.
Abstract: As more enterprises buy information technology services, studying their underpinning contracts becomes more important. With cloud computing and outsourcing, these service level agreements (SLAs) are now often the only link between the business and the supporting IT services. This paper presents an experimental economics investigation of decision-making with regard to availability SLAs, among enterprise IT professionals. The method and the ecologically valid subjects make the study unique to date among IT service SLA studies. The experiment consisted of pairwise choices under uncertainty, and subjects ( $N=46$ ) were incentivized by payments based on one of their choices, randomly selected. The research question investigated in this paper is: Do enterprise IT professionals maximize expected value when procuring availability SLAs, as would be optimal from the business point of view? The main result is that enterprise IT professionals fail to maximize expected value. Whereas some subjects do maximize expected value, others are risk-seeking, risk-averse, or exhibit nonmonotonic preferences. The nonmonotonic behavior in particular is an interesting observation, which has no obvious explanation in the literature. For a subset of the subjects ( $N=29$ ), a few further hypotheses related to associations between general attitude to risk or professional experience on the one hand, and behavior in SLAs on the other hand, were investigated. No support for these associations was found. The results should be interpreted with caution, due to the limited number of subjects. However, given the prominence of SLAs in modern IT service management, the results are interesting and call for further research, as they indicate that current professional decision-making regarding SLAs can be improved. In particular, if general attitude to risk and professional experience do not impact decision-making with regard to SLAs, more extensive use of decision-support systems might be called for in order to facilitate proper risk management.
28 citations
04 Jan 2011
TL;DR: This work argues that the trade-off between service cost and benefit incurred by both parties is not sufficiently considered when service quality is stipulated, and lays the foundation for the field of Service Level Engineering discussing its constitutional concepts and elements.
Abstract: Today, the management of service quality poses a major challenge for many service providers and their business customers. We argue that the trade-off between service cost and benefit incurred by both parties is not sufficiently considered when service quality is stipulated. Up to now, neither in practice nor in academia a commonly accepted engineering approach exists to determine business-relevant performance metrics and associated cost-efficient target values to precisely identify efficient service quality. It can be expected, though, that such a systematically developed and economically well-founded Service Level Engineering approach - based on the individual economic conditions of customer and provider - can enhance the value generation in service systems formed by both these parties. In this work we lay the foundation for the field of Service Level Engineering discussing its constitutional concepts and elements. After developing a generic framework for different Service Level Engineering scenarios, we analyze one scenario - the system view - in detail as a quantitative optimization problem: In particular, we aim at the derivation of cost-efficient target values given business-relevant performance-metrics.
19 citations
Cites background from "SLA Design and Service Provisioning..."
...External SLE, which we consider a main area of application, is only briefly discussed in [13]....
[...]
...Only one recent paper [13] adapts the model to IT outsourcing relationships....
[...]
TL;DR: Results of the analysis indicate that a properly selected SLA-controlled interconnection charging policy should encourage providers to offer services with different QoS levels, and to achieve the contracted QoS level.
12 citations
Cites background from "SLA Design and Service Provisioning..."
...A business-driven approach to designing SLA in an e-commerce environment has been proposed in [22]....
[...]
References
More filters
Book•
02 Jan 1975
TL;DR: The purpose of this document is to summarize the main points of the book written by Leonard Kleinrock, titled, ‘Queueing Systems’, which is about queueing systems.
Abstract: The purpose of this document is to summarize the main points of the book written by Leonard Kleinrock, titled, ‘Queueing Systems”.
2,885 citations
"SLA Design and Service Provisioning..." refers methods in this paper
...In order to find RespTimeDist(x), IT services are modeled by means of a multiclass open queuing model; each CBMG class represents a queuing model class [14]....
[...]
TL;DR: Probability and Statistics with Reliability, Queuing and Computer Science Applications, Second Edition, offers a comprehensive introduction to probabiliby, stochastic processes, and statistics for students of computer science, electrical and computer engineering, and applied mathematics.
Abstract: Probability and Statistics with Reliability, Queuing and Computer Science Applications, Second Edition, offers a comprehensive introduction to probabiliby, stochastic processes, and statistics for students of computer science, electrical and computer engineering, and applied mathematics. Its wealth of practical examples and up-to-date information makes it an excellent resource for practitioners as well.
2,738 citations
"SLA Design and Service Provisioning..." refers methods in this paper
...Since the clusters being used in an IT resource class will be available and ready to handle the imposed load when at least mj resources are available for load-balancing, resource class availability, Aj, uses ‘‘m-out-of-n reliability’’ [12]....
[...]
...Standard reliability theory [12] uses mean-time-to-repair (MTTR) and mean-time-betweenfailures (MTBF) values to compute the availability, Aj , of individual IT resources: Aj 1⁄4 Y...
[...]
Book•
01 Jan 1982TL;DR: Probability and Statistics with Reliability, Queuing and Computer Science Applications, Second Edition as discussed by the authors is a comprehensive introduction to probabiliby, stochastic processes, and statistics for students of computer science, electrical and computer engineering, and applied mathematics.
Abstract: Probability and Statistics with Reliability, Queuing and Computer Science Applications, Second Edition, offers a comprehensive introduction to probabiliby, stochastic processes, and statistics for students of computer science, electrical and computer engineering, and applied mathematics. Its wealth of practical examples and up-to-date information makes it an excellent resource for practitioners as well.
2,629 citations
Book•
05 Jan 2004
TL;DR: Practical systems modeling: learning exactly how to map real-life systems to accurate performance models, and use those models to make better decisions--both up front and throughout the entire system lifecycle.
Abstract: Practical systems modeling: planning performance, availability, security, and moreComputing systems must meet increasingly strict Quality of Service (QoS) requirements for performance, availability, security, and maintainability To achieve these goals, designers, analysts, and capacity planners need a far more thorough understanding of QoS issues, and the implications of their decisions Now, three leading experts present a complete, application-driven framework for understanding and estimating performance You'll learn exactly how to map real-life systems to accurate performance models, and use those models to make better decisions--both up front and throughout the entire system lifecycle Coverage includes: State-of-the-art quantitative analysis techniques, supported by extensive numerical examples and exercises QoS issues in requirements analysis, specification, design, development, testing, deployment, operation, and system evolution Specific scenarios, including e-Business and database services, servers, clusters, and data centers Techniques for identifying potential congestion at both software and hardware levels Performance Engineering concepts and tools Detailed solution techniques including exact and approximate MVA and Markov Chains Modeling of software contention, fork-and-join, service rate variability, and priorityAbout the Web SiteThe accompanying Web site provides companion Excel workbooks that implement many of the book's algorithms and numerical examples
411 citations
"SLA Design and Service Provisioning..." refers background or methods in this paper
...The parameters used in the model to capture the example scenario are shown in Table 4 and are typical for recent technology [2, 10, 15]....
[...]
...commission—in fact, such an ‘‘autonomic self-optimization according to business objectives’’ web site scenario is the focus in [10]....
[...]
...Let us couch the argument in an e-commerce context: a customer accessing the service site will defect—that is, not conclude his purchases—when response time is greater than a given upper bound, T; the value of 8 s is frequently mentioned in the literature as an appropriate threshold value [10]....
[...]
Book•
07 May 2000
TL;DR: Scaling for E-Business presents analysis techniques for quantifying and projecting every element of your e-business site's performance-and planning for the capacity you need, no matter what!
Abstract: From the Publisher:
Foreword by Jim Gray
Microsoft Research and 1998 ACM Turing Award Recipient
The Complete Guide to Techniques and Models to Evaluate and Plan E-Business Sites!
Avoid losing customers due to site crashes
Performance modeling and capacity planning for e-business infrastructure
Build and analyze customer behavior models
Plan your e-business site to avoid frequent upgrades and migrations
"E-business is transforming every element of business in many unknown ways. This book will give its readers a headstart in understanding the quantitative underpinnings of the new environment."
Howard Frank, Dean, Robert H. Smith School of Business
University of Maryland at College Park
Solid analysis of performance modeling and capacity planning for your e-business site! S low e-commerce sites cost their owners billions and embarrass and degrade their owners' brands. Don't let it happen to you!
Scaling for E-Business presents analysis techniques for quantifying and projecting every element of your e-business site's performance-and planning for the capacity you need, no matter what! Discover how to...
Characterize e-commerce workloads more accurately
Analyze the performance of authentication and payment services
Model contention for software servers, and ensure scalability
Model and plan for communications infrastructure
Forecast and cope with peak demand
Project the impact of agent technologies and non-PC devices
You can't turn to your vendors for these state-of-the-art techniques. But you canturn to Scaling for E-Business-and if you plan to succeed, you'd better!
281 citations
"SLA Design and Service Provisioning..." refers methods in this paper
...Some CBMG states generate revenue for the client; this is the case, for example, of an e-commerce site page where the customer pays for items in a shopping cart....
[...]
...In order to find RespTimeDist(x), IT services are modeled by means of a multiclass open queuing model; each CBMG class represents a queuing model class [14]....
[...]
...A CBMG includes a set of states and a probability matrix of transitions between states; to exemplify, one can view a CBMG state as a page being visited on an e-commerce site....
[...]
...The site has two CBMGs: one to model RG sessions and another to model non-revenue-generating sessions....
[...]
...The load model adopted is based on the Customer Behavior Model Graph (CBMG) [13] that permits a representation of how a given session initiated by a customer applies load on the IT infrastructure....
[...]