Werner Streitberger

Catallaxy-based Grid markets

Grid computing has recently become an important paradigm for managing computationally demanding applications, composed of a collection of services. The dynamic discovery of services, and the selection of a particular service instance providing the best value out of the discovered alternatives, poses a complex multi-attribute n:m allocation decision problem, which is often solved using a centralized resource broker. To manage complexity, this article proposes a two-layer architecture for service discovery in such Application Layer Networks (ALN). The first layer consists of a service market in which complex services are translated to a set of basic services, which are distinguished by price and availability. The second layer provides an allocation of services to appropriate resources in order to enact the specified services. This framework comprises the foundations for a later comparison of centralized and decentralized market mechanisms for allocation of services and resources in ALNs and Grids.

A simulation of an economic, self-organising resource allocation approach for application layer networks

Application Layer Networks like Utility Grid Networks or Cloud Computing Systems, often depicted as large scale self-organising architectures using a shared infrastructure, will need innovative management approaches for controlling and matching services demand and supply. This article shows a self-organising resource allocation approach based on principles of the Catallaxy, an economic concept of F.A. von Hayek. The implementation uses a coevolutionary learning algorithm for adaptation of the agent strategy to the dynamic Application Layer Network environment. The simulation study performs a sensitivity analysis of a large scale Application Layer Network scenario with 2000 agents. A set of representative agents is selected and their evolutionary behaviour analysed.

On the Simulation of Grid Market Coordination Approaches

Grid computing has recently become an important paradigm for managing computationally demanding applications, composed of a collection of services. The dynamic discovery of services, and the selection of a particular service instance providing the best value out of the discovered alternatives, poses a complex multi-attribute n:m allocation decision problem, which is often solved using a central resource broker. However, decentralized approaches to this service allocation problem represent a much more flexible alternative, thus promising improvements in the efficiency of the resulting negotiations and service allocations. This paper compares centralized and decentralized service allocation mechanisms in Grid market scenarios according to a defined set of metrics.

Enabling the simulation of service-oriented computing and provisioning policies for autonomic utility grids

There are key challenges in utility computing environments such as the provisioning, orchestration and allocation of resources to services. In these environments, providers need to decide how resources are allocated to service applications according to their workloads in order to guarantee the Quality of Service (QoS) required by customers. Autonomic computing inspired mechanisms are appealing to enable self-organising resource allocation and provisioning. However, these mechanisms are difficult to evaluate in practice either because of the lack of a real test bed or the difficulty in replicating experimental results. This work thus describes a service framework for a Grid simulator. This framework allows the modelling and evaluation of the provisioning and negotiation of services and resources. We also discuss experimental results that demonstrate the usefulness of this framework for the simulation of a decentralised and self-organising economic model for service and resource negotiation termed Catallaxy.

CATNETS - open market approaches for self-organizing grid resource allocation

Grid computing has recently become an important paradigm for managing computationally demanding applications, composed of a collection of services. The dynamic discovery of services, and the selection of a particular service instance providing the best value out of the discovered alternatives, poses a complex multi-attribute n:m allocation decision problem. Decentralized approaches to this service allocation problem represent a flexible alternative to central resource brokers, thus promising improvements in the efficiency of the resulting negotiations and service allocations. This paper analyses the impact of the service density on the profit and market price estimation using a decentralized service allocation mechanism in a grid market scenario.

Framework for the negotiation of electronic contracts in e-business on demand

This paper presents a realized framework for the negotiation of electronic contracts for e-business on demand, in which resource demands occur unpredictably and have to be fulfilled immediately. Proposals for services are exchanged between clients and service providers, and offers are negotiated and contracted. On demand service contracts mostly have a smaller trading volume as ordinary offline contracts, but their frequency of occurrence increases. Due to the high number of contracts they should be negotiated electronically, which lowers costs of the contract process notably. The software framework presented in this paper can support this process. The core components of the framework have generic interfaces. These interfaces can be used as a base for developing domain specific applications to implement negotiations of electronic contracts. The framework is part of a contract management system extending its functionality of negotiating contracts.

Facing Price Risks in Internet-of-Services Markets

Internet-of-Services markets allow companies to procure computational resources and application services externally and thus to save both internal capital expenditures and operational costs. Despite the advantages of this new paradigm only few work has been done in the field of risk management concerning Internet-of-Services markets. We simulate such a market using a Grid simulator. The results show that market participants are exposed to price risk. Based on our results we identify and assess technical failures which could lead to loss on service consumer’s side. We also show that technical failures influence service prices which lead to volatile prices. Both, service provider and service consumer are exposed to this uncertainty and need a way to face it. Therefore we apply a financial option model to overcome price risk.

Global grids - making a case for self-organization in large-scale overlay networks

Grid computing has recently become an important paradigm for managing computationally demanding applications, composed of a collection of services. The dynamic discovery of services, and the selection of a particular service instance providing the best value out of the discovered alternatives, poses a complex multi-attribute n:m allocation decision problem, which is often solved using a centralized resource broker. To manage complexity, this article proposes a two-layered architecture for decentralized service discovery in such Application Layer Networks (ALN). The first layer consists of a service market in which complex services are translated to a set of basic services, which are distinguished by price and availability. The second layer provides an allocation of services to appropriate resources in order to enact the specified services. This framework comprises the foundations for a later comparison of centralized and decentralized market mechanisms for allocation of services and resources in ALNs and Grids in general.

Economic Evaluation Framework of Resource Allocation Methods in Service-Oriented Architectures

Economic resource allocation in application layer networks (such as grids) is critical to allow applications and users to effectively exploit computational and data infrastructures like service-oriented computing as a utility. Thus, the evaluation of resource allocation strategies plays a major part in the selection of a resource allocation method. This paper presents an evaluation framework for resource allocation methods in application layer networks, that aims at supporting both a technical and an economic evaluation. The presented evaluation model shows a layered metrics pyramid with different aggregation levels. Statistical methods are used to describe this pyramid. On top of the pyramid, one single number, the social utility, is able to characterize an economic resource allocation method. This number may serve to compare different resource allocation strategies

Ökonomische Bewertung der Dienstauswahl in Service-Netzen

Einfache IT-Dienstleistungen werden im Konzept des On-Demand bzw. Grid Computing an externe Anbieter ausgelagert. Dieser bedarfsabhangige ezug von Rechenleistung wird als Chance zur Beseitigung von Ineffizienzen und zur Kostenreduktion gesehen. Der vorliegende Beitrag beschaftigt sich insbesondere mit Verfahren zur Dienstauswahl, uber die in den vorliegenden Konzepten solcher serviceorientierter Architekturen keine Aussagen bezuglich Anwendbarkeit und Performanz gefallt werden. Er vergleicht die existierenden Ansatze mit einem koordinatorfreien - auf okonomischen Prinzipien beruhenden - Ansatz anhand okonomischer Metriken. Mittels einer Simulation werden unterschiedliche Verfahren in verschiedenen Netzwerk-szenarien untersucht und die bessere Anpassungsfahigkeit der koordinatorfreien Variante an Dynamik und Knotendichte des Netzwerks gezeigt.