Amro Najjar

Survey of Elasticity Management Solutions in Cloud Computing

Application Service Providers (ASPs) are increasingly adopting the cloud computing paradigm to provision remotely available resources for their applications. In this context, the ability of cloud computing to provision resources on-demand in an elastic manner is of the utmost practical interest for them. As a consequence, the field of cloud computing has witnessed the development of a large amount of elasticity management solutions deeply rooted in works from distributed systems and grid computing research communities. This chapter presents some solutions that differ in their goals, in the actions they are able to perform and in their architectures. In this chapter, we provide an overview of the concept of cloud elasticity and propose a classification of the mechanisms and techniques employed to manage elasticity. We also use this classification as a common ground to study and compare elasticity management solutions.

Modeling User Expectations & Satisfaction for SaaS Applications Using Multi-agent Negotiation

As more personal and interactive applications are moving to the cloud, modeling the end-user expectations and satisfaction is becoming necessary for any SaaS provider to survive and thrive in todays competitive market. However, most of existing works addressing cloud elasticity management adopt a centralized approach where user preferences are mostly overlooked. Based on evidence from the fields of customer expectation management and psychophysics, in this article we propose a personal user model to represent end-user satisfaction and her expectations. To integrate the end-user into the decision loop we develop multi-agent negotiation architecture in which the end-user model is embodied by a personal agent who negotiates on her behalf. The results of the evaluation process show that automated negotiation provides a useful platform to empower the user choices, fulfill her expectations, and maximize her satisfaction hereby outperforming centralized approaches where the provider acts in a unilateral manner.

Multi-agent Systems for Personalized QoE-Management

User satisfaction is becoming a key factor to secure the success of any online service. Quality of Experience is a subjective measure of the service quality as perceived by the user. QoE has been introduced to bridge the gap between the purely technical characteristics of QoS and user satisfaction. Recent research on QoE has shown that QoE is highly personal and influenced by multiple interrelated factors including the user expectations, preferences and cultural background. However, most existing QoE management solutions overlook the personal aspect of QoE and ignore inter-user differences despite the promise of adopting a user-centric approach. In this paper, we propose multi-agent technology as means to achieve personalized QoE-management. In particular, we propose a multi-agent architecture called EMan where each end-user is embodied by an autonomous agent that represents her personal preferences and expectations and seeks to maximize her QoE. To evaluate our approach, we use Repast, a multi-agent simulation platform. The preliminary results proves that such a decentralized multi-agent QoE-management outperforms an equivalent centralized approach both in terms of end-user satisfaction and service acceptability.

Multi-agent Negotiation for User-centric Elasticity Management in the Cloud

Application Service Providers (ASP) are increasingly adopting the cloud computing paradigm to provision remote resources for their applications. Elasticity management solutions have been proposed in the domain to enable flexible and autonomic cloud resource provisioning. However, only few solutions consider the key issue of the conflict between ASP business goals and end-users preferences, that is to say Quality of Experience of the end-user. This paper proposes EMan, a decentralized approach for user-centric Elasticity Management. We use multi-agent automated negotiation strategies and tactics that allow the ASP to address the tradeoff between its customers satisfaction and its business objectives and achieve conflict-free solutions. We use Repast simulation environment to simulate EMan. The results show that the ASP can achieve the desired balance between its end-users satisfaction and its business objectives by combining different negotiation strategies and tactics.

Elastic & Load-Spike Proof One-to-Many Negotiation to Improve the Service Acceptability of an Open SaaS Provider

Service acceptability rate and user satisfaction are becoming key factors to avoid client churn and secure the success of any Software as a Service (SaaS) provider. Nevertheless, the provider must also accommodate fluctuating workloads and minimize the cost it pays to rent resources from the cloud. To address these contradicting concerns, most of existing works carry out resource management unilaterally by the provider. Consequently, end-user preferences and her subjective acceptability of the service are mostly ignored. In order to assess user satisfaction and service acceptability recent studies in the domain of Quality of Experience (QoE) recommend providers to use quantiles and percentile to gauge user service acceptability precisely. In this article we propose an elastic, load-spike proof, and adaptive one-to-many negotiation mechanism to improve the service acceptability of an open SaaS provider. Based on quantile estimation of service acceptability rate and a learned model of the user negotiation strategy, this mechanism adjusts the provider negotiation process in order to guarantee the desired service acceptability rate while meeting the budget limits of the provider and accommodating workload fluctuations. The proposed mechanism is implemented and its results are examined and analyzed.