Tobias Unger

Combining Different Multi-tenancy Patterns in Service-Oriented Applications

Software as a service (SaaS) providers exploiteconomies of scale by offering the same instance of an applicationto multiple customers typically in a single-instance multitenantarchitecture model. Therefore the applications must bescalable, multi-tenant aware and configurable. In this paper weshow how the services in a service-oriented SaaS applicationcan be deployed using different multi-tenancy patterns. Wedescribe how the chosen patterns influence the customizability,multi-tenant awareness and scalability of the application. Usingthe patterns we describe how individual services in a multitenantaware application can be not multi-tenant aware whilemaintaining the overall multi-tenant awareness of the application.We show based on a real-world example how the patternscan be used in practice and show how existing applicationsalready use these patterns.

Horizontal and vertical combination of multi-tenancy patterns in service-oriented applications

Software as a service (SaaS) providers exploit economies of scale by offering the same instance of an application to multiple customers typically in a single-instance multi-tenant architecture model. Therefore the applications must be scalable, multi-tenant aware and configurable. In this article, we show how the services in a service-oriented SaaS application can be deployed using different multi-tenancy patterns. We describe how services in different multi-tenancy patterns can be composed on the application level. In addition to that, we also describe how these multi-tenancy patterns can be applied to middleware and hardware components. We then show with some real world examples how the different multi-tenancy patterns can be combined.

Aggregation of Service Level Agreements in the Context of Business Processes

One can observe that more and more companies are focusing on their core compenetency and are outsourcing parts or even complete business processes to service providers. Service Level Agreements (SLAs) are used as a contractual basis to define certain non-functional properties (e.g. response time) a service has to provide. To determine the SLA of a business process a priori, the SLAs of the invoked services need to be aggregated into a single SLA for the business process. This paper defines a method how a service provider can aggregate the SLAs of the individual services within a business process into a single SLA. This provides a service provider with the capability to annotate the service that the business process implements with an appropriate SLA. We propose a framework that helps in carrying out the SLA aggregation within business processes. The framework consists of two parts: (i) a formal model for SLAs and (ii) a concept to aggregate SLAs, where algorithms can be plugged in.

Cafe: A Generic Configurable Customizable Composite Cloud Application Framework

In this paper we present Cafe (Composite Application Framework) an approach to describe configurable composite service-oriented applications and to automatically provision them across different providers. Cafe enables independent software vendors to describe their composite service-oriented applications and the components that are used to assemble them. Components can be internal to the application or external and can be deployed in any of the delivery models present in the cloud. The components are annotated with requirements for the infrastructure they later need to be run on. Providers on the other hand advertise their infrastructure services by describing them as infrastructure capabilities. The separation of software vendors and providers enables end users and providers to follow a best-of-breed strategy by combining arbitrary applications with arbitrary providers. We show how such applications can be automatically provisioned and present an architecture and a prototype that implements the concepts.

Enabling Adaptation of Pervasive Flows: Built-in Contextual Adaptation

Adaptable pervasive flows are dynamic workflows situated in the real world that modify their execution in order to adapt to changes in the execution environment. This requires on the one hand that a flow must be context-aware and on the other hand that it must be flexible enough to allow an easy and continuous adaptation. In this paper we propose a set of constructs and principles for embedding the adaptation logic within the specification of a flow. Moreover, we show how a standard language for web process modeling (BPEL) can be extended to support the proposed built-in adaptation constructs.

Customer-defined service level agreements for composite applications

One key factor for successful outsourcing of applications to a provider is to guarantee a certain level of service. Hence, the service provider and service consumer agree on service level agreements (SLA) that define the behaviour and quality of the non-functional properties. Unlike many existing approaches, which only allow the customer to choose between a small set of predefined service levels, our approach helps to cope with the situation that customers want to model their own service level according to their individual business needs. As a first step, we introduce an application model which enables the outsourcing of applications. In addition, we present corresponding models for describing Quality of Service (QoS) requirements and SLAs. To assure the requested QoS, the provider has to determine the QoS requirements of each application component. In this work, we determine the QoS requirement of a component by disaggregating the QoS requirements stated in the SLA. After the disaggregation the provider deploys each component of the application in an optimal way based on its QoS requirements, determined during the disaggregation.

Dynamic Composition of Pervasive Process Fragments

A critical aspect for pervasive computing is the possibility to discover and use process knowledge at run time depending on the specific context. This can be achieved by using an underlying service-based application and exploiting its features in terms of dynamic service discovery, selection, and composition. Pervasive process fragments represent a service-based tool that allows to model incomplete and contextual knowledge. We provide a solution to automatically compose such fragments into complete processes, according to a specific context and specific goals. We compute the solution by encoding process knowledge, domain knowledge and goals into an AI planning problem. We evaluate our approach on different scenarios stress testing the main characteristics of pervasive process fragments.

Process Fragment Composition Operations

The construction kit principle is a well-known software engineering paradigm to foster reusability. In case the construction kit principle is applied at runtime it is even a way to implement flexibility. In today’s workflow technology the construction kit principle is applied, e.g. in in hierarchical modeling approaches using sub processes. In this paper we propose a construction kit application based on process fragment. In contrast to sub processes, process fragments represent on-complete process knowledge, which needs to be integrated with further process knowledge to become a complete process model. Integrating one process fragment with another process fragment requires complex composition operations, since process fragments do not represent the implementation of a single abstract activity like in sub processes, but are knitted together on the same level of granularity. The advantage of process fragments lies in their means to represent noncomplete process knowledge. In this paper we propose a formal process fragment modeling language, which is based on current workflow standard languages, like BPMN. Based on this modeling language we design a basic operation set, which allows to compose process fragments.

A Taxonomy for Cloud Data Hosting Solutions

Cloud computing allows reducing capital expenditure by using resources on demand. We investigate how to build a database layer in the Cloud and present pure and hybrid Cloud data hosting solutions. The solutions are organized in a taxonomy. The properties used for organization are: application layer, deployment model, location, service model, data store type, and compatibility. Using the taxonomy, existing Cloud data hosting solutions are categorized.

Performance Optimizations for Interacting Business Processes

Choreographies describe the interaction behavior of processes at design time: a choreography defines when messages have to be exchanged between the involved processes during their runtime. In the context of Web services and the de-facto workflow language BPEL, SOAP is used to encode the messages. When complex messages are exchanged between the processes, this can become costly and time consuming with respect to the overall execution time of a choreography. In this work, we suggest three different performance optimization techniques for workflow engines to reduce the number of message exchanges between the interacting processes and hence, to decrease the execution times and costs of the choreographies: intra-engine transport, service request caching, and inline execution. We describe how these techniques are implemented in a workflow engine. Performance measurements are carried out to determine the performance improvements that are achieved with each optimization technique. We further show that the optimizations also affect the energy consumption of the workflow engine.