Claudia Di Napoli

Using chemical reactions to model service composition

Internet is evolving from a network of computers and information into a network of services allowing applications to be built by selecting services and composing them in a loosely coupled manner. These Service Based Applications (SBA) are composed of a number of possibly independent services that are provided by many actors under different conditions (like price, time to deliver, and so on). Service provision conditions may change in time depending on provider policies or other environmental changes, so it is necessary to organize compositions of services on demand in response to dynamic requirements and circumstances. In this paper we propose to use a chemical computational model to address this problem by decoupling the process of finding services composing an SBA requested by a user, from their actual enactment. An SBA request is described in terms of an abstract workflow where only service functionalities of the single components and their execution order (i.e. the application control flow) are specified, along with parameters representing the conditions under which the user expects the application to be delivered. The proposed approach allows to model the process of instantiating the required functionalities with actual service implementations as an evolving and always running middleware mechanism that can take into account the current state of the context when the composition is required. Furthermore, the evolutionary nature of the chemical system provides a form of adaptation since once compositions of services are computed with the available services, new compositions can be computed as soon as new services become available or the conditions of existing ones change.

A Chemical Based Middleware for Workflow Instantiation and Execution

Service-Oriented Architecture (SOA) is widely adopted today for building loosely coupled distributed applications. With this approach, new applications can be built by creating a business process that defines a concrete workflow composed of different partner services available via the network. In this scenario, the bindings between the business process and partner services are predefined by statically referencing the corresponding endpoints. This paper proposes a middleware architecture for dynamical workflow instantiation and execution. Using this middleware, partner services are selected and bound in the run-time and the aggregated QoS values are ensured to satisfy the requester’s end-to-end QoS requirement. The selection is based on both non-functional requirement (such as price and response time) of the global composition, and Quality of Service (QoS) performance of each candidate service. The implementation is based on chemical computing, a parallel and autonomic computing paradigm that allows to model workflow instantiation and execution as an evolving and adaptable process.

A network of cellular automata for a landslide simulation

Cellular Automata (CA) offer a promising computational model to simulate complex phenomena that are characterized by different coupled parameters that account for the interactions of their different components. In this paper we present the Cellular Automata Network (CAN) computational model and its application to the simulation of debris/flow phenomena. We carried out some experimentations developing a CAN application that implements the SCIDDICA model (Simulation through Computational Innovative methods for the Detection of Debris flow path using Interactive Cellular Automatamodel), tested on the landslide occurred in Sarno (Italy) in 1998. CAN model allows to represent each component of a physical system in terms of cellular automata, and the interactions among these components in terms of a network of cellular automata. The adoption of the CAN model allows to exploit two different types of parallelism: the data parallelism that comes from the use of Cellular Automata classical model, and the task parallelism that could occur introducing the network of Cellular Automata.

A continuation-based framework for economy-driven grid service provision

The management of computational resources is a crucial aspect in grid computing because of the decentralized, heterogeneous and autonomous nature of these resources that usually belong to different administrative domains and are provided in dynamic and changing environments. For this reason more sophisticated computing methodologies are necessary to provide these resources in a flexible manner. In particular, the possibility of controlling the execution of services in grid is a crucial aspect in order to change service execution policies at run-time. In the present work an infrastructure to model service providers is proposed to allow for flexible provision of grid services, i.e. to allow providers to dynamically control the execution of services according to the changing conditions of the environment where they operate in. The infrastructure is based on continuations, a programming paradigm that allows to control the state of program execution at application-level without involving the operating system stack. This approach makes the proposed infrastructure a flexible and easily programmable middleware to experiment different scheduling policies in service-oriented scenarios.

Exploring Multi-level Parallelism in Cellular Automata Networks

Usually physical systems are characterized by different coupled parameters accounting for the interaction of their different components. The Cellular Automata Network (CAN) model [1] allows to represent each component of a physical system in terms of Cellular Automata (CA) [9], and the interaction among these components in terms of CA networks. In this paper we report our experimentations in exploiting two different kinds of parallelism offered by the CAN model using policies for network restructuring and thread assignment. At this purpose we used a prototype graphic tool (CANviz) designed to let the user experimenting heuristics to effciently exploit two-level parallelism in CAN applications.

A portable parallel environment for complex systems simulation through cellular automata networks

Abstract In this paper we describe the Parallel Environment for Cellular Automata Network Simulation (PECANS) we developed to deal with the simulation of complex systems. It provides a user-friendly environment in which all the aspects of the simulation activity can be easily carried out. In particular we describe how the PECANS environment allowed us to easily port it on different architectures, both sequential and parallel ones. Some experimental results are also reported.

MIDAS: a cloud platform for SOA testing as a service

The increasing adoption of the service-oriented architecture approach makes software quality more challenging since traditional approaches for software testing are inadequate for SOA-based applications. The MIDAS platform is an integrated platform for SOA testing automation, designed and architected according to the SOA computing paradigm, and deployed on a public cloud infrastructure to tackle the variability in the computational resources necessary for testing SOA applications. It is available to end users as a testing as a service on a self-provisioning, pay-per-use, elastic basis. The cloud-based software architecture envisioned for the MIDAS platform and the strategies adopted for both its development, and its deployment on the target cloud infrastructure are here discussed, together with the solution designed and implemented in order to monitor the usage of MIDAS services and resources. Also, the strategy adopted to provide the MIDAS platform with the management of elastic resources is outlined.

A SOA Testing Platform on the Cloud: The MIDAS Experience

The widespread use of SOA-based applications is posing new challenges to software testing since conventional methodologies are inadequate to cover the specific testing requirements of these applications. In this context, the European FP7 Project MIDAS aims at building an integrated platform for SOA testing automation designed and architected according to the SOA computing paradigm. In order to address the variability in the computational resources necessary for testing SOA applications of different complexity coming from different users, the MIDAS platform is designed as an integrated Testing as a Service (TaaS) framework deployed on a public Cloud infrastructure, available on a self-provisioning, pay-per-use, elastic basis. This paper introduces and discusses the Cloud-based software architecture envisioned in the MIDAS project by detailing the strategies adopted for its deployment on the target Cloud infrastructure (Amazon AWS).

A Methodology to Annotate Cultural Heritage Digital Video

The goal of our work is to provide a well-structured methodological approach for annotating digital video in the domain of cultural heritage, taking into account that the methodologies and technologies currently available do not allow to make the annotation process completely automatic. In the proposed approach an interaction with the user that decides both the video segments to be annotated, and the set of labels to be associated to a segment is always required. This approach uses a predefined set of categories (world-view), hierarchically structured, which classifies the video subjects describing the contents shown in a video. This approach leads both to a way of integrating heterogeneous sources of information, and to structure information in such a way to retrieve video segments through a simple query language.

Adaptive instantiation of service workflows using a chemical approach

Service oriented technologies allow Service Based Applications (SBAs) to be easily built by composing independent services available in a network and provided by many actors under conditions that may change in time. Therefore services need to be dynamically selected and composed when an SBA is required along with parameters representing the service delivery conditions. In this paper we propose to use a chemical computational approach to model the process of selecting the required service functionalities with the required conditions as an evolving and always running middleware mechanism. The chemical evolving behaviour of the middleware allows to take into account environmental changes coming from both the providers and users side.