|Title||Advancement of farming by facilitating collaboration : reference architectures and models for farm software ecosystems|
|Author(s)||Kruize, Jan Willem|
|Source||University. Promotor(en): Adrie Beulens, co-promotor(en): Huub Scholten; Sjaak Wolfert. - Wageningen : Wageningen University - ISBN 9789462579668 - 242|
LEI Innovation, Risk and Information Management
|Publication type||Dissertation, internally prepared|
|Keyword(s)||farming - information technology - computer software - farms - models - farm management - information systems - landbouw bedrijven - informatietechnologie - landbouwbedrijven - modellen - agrarische bedrijfsvoering - informatiesystemen|
|Categories||Information and Communication Technology (General) / Farm Management / Agricultural Engineering (General)|
Since time began, mankind has been threatened by the combination of growing populations and diminishing resources. Present-day, this threat is very pertinent as mankind is challenged by a growing world population that is expected to exceed 10 billion in 2050, while resources diminish. Simultaneously, increase of food production should be accomplished in a sustainable manner as consumers require food to be produced environmentally-friendly. Moreover, consumers require safe food produced in transparent agri-food supply chain networks. Farm enterprises can contribute by advancing their management to increase food production in a sustainable, safe and transparent manner. A well-known advanced farm management style, which is knowledge and information intensive, is precision agriculture. Precision agriculture increases the profitability of crop production, while simultaneously reducing the negative environmental impact by tight monitoring and control, in which applications rates of agricultural inputs are adjusted to local needs. Such advanced farm management requires integrated farm information systems as it is knowledge and information intensive. However, advancement is hindered because of interoperability issues between software systems of multiple vendors. An integrated farm information system, containing components of multiple vendors, is required as single organisations cannot develop all technical solutions and ICT Components (e.g. tractors, implements, FMIS, decision support tools) that farmers require. A global overarching system, developed by a single vendor, that can support all business functions of farmers is therefore neither a feasible nor, from a competitive point of view, a desirable solution in agriculture. To realize farm enterprise integration we combine the approaches ICT Mass Customisation with Best-of-Breed. ICT mass customisation combines advantages of standard and customised software by enabling on-demand configuration of information systems from standard components with standardised interfaces. These ICT components can be supplied by different software vendors, which allow Best-of-Breed solutions. By realization of these approaches farm enterprise integration can improve. A farm enterprise can be an arable farm, livestock farm or horticultural farm. In this thesis we focus on arable farm enterprises.
To enable farm enterprise integration we have developed six artefacts that are presented in this thesis which are:
The Reference Architecture of Agricultural Enterprises (RAAgE) 1.0 that can describe farm enterprise architectures in a uniform and efficient manner;
A problem description, which is a case specific instantiation of RAAgE 1.0 generalized to a generic problem description;
An ontology that supports communication between collaborating actors and components;
Reference Architecture for Farm Software Ecosystems that defines generic relationships between actors and components;
RAAgE 2.0 that is a technical reference model to support configuration of business processes and ICT components, which is based on RAAgE 1.0;
Prototype software that serves as a proof of concept substantiating that all previous components will provide a solution for integration problems at farm enterprises.
RAAgE 1.0 supports designing enterprise architectures in a uniform and efficient manner. The reference model is described in a standard modelling language, named ArchiMate, and shows the interrelations between the business, application and technology layers of farm enterprises. The reference model includes an ontology to provide a concise and precise, formal specification of the object system. This is required to have a shared understanding and effective communication between researchers, farmers, software developers and other stakeholders involved. This ontology is used and extended in other parts of our research. The architectural descriptions can depict the relations between farm business processes and the ICT Components used. The model is validated by two experts that have experience in developing reference architectures and models.
A detailed problem description is created using RAAgE 1.0 to gain insight in the cause and nature of integration problems at farm enterprises. To find these problems a method was developed and applied in a case study research including three arable farm enterprises producing potatoes. These farm enterprises focused on improving their management and invested in new technologies for innovation. Within multiple steps of the method the architectural descriptions developed with RAAgE 1.0 facilitated communication and provided insight into problems of farm enterprises to achieve more advanced farm management. The case specific problems, described by instantiating RAAgE 1.0, have been analysed and formulated as more generic problems for farm enterprise integration. These generic problem descriptions have been validated with national and international experts. Based on this research we found that the cause and nature of current integration problems in farming are that ICT components used within the same farm enterprise:
have partly overlapping and partly unique application services, functions and interfaces (that are non-standard);
are missing required application services, functions and interfaces,
have disjoint data repositories;
have inadequate and incomplete data exchange as semantics are not unambiguously defined;
are hard to configure while this configuration is not supported by an actors and tools.
A design, addressing these problems is expected to solve current integration bottlenecks. First, this design must enable smooth data handling and seamless data exchange between ICT Components to solve inadequate and incomplete data exchange and enable integration of data repositories of multiple vendors. Second, it must include a configuration approach to link ICT Components to each other in a meaningful and coherent way. This should be supported by actors that are willing to configure ICT Component of multiple vendors into an integrated solution. Third, the design must enable the formation of a software enterprise to address the previous points and to organize collaboration between actors involved. This software enterprise should focus both on improving interoperability to contribute in solving problems with partly overlapping and partly unique application services, functions and interfaces as well as on organizing the development of missing application services, functions and interfaces.
To address these integration challenges a Reference Architecture for Farm Software Ecosystems and RAAgE 2.0 were developed, focusing on both technical and organizational aspects.
From literature we found that collaboration can take place within Software Ecosystems. Software Ecosystems are defined as the interaction of a set of actors on top of a common technological platform that results in a coherent set of ICT components or Services. They can provide an effective way to construct large software systems on top of a software platform by combining components, developed by actors that are part of different organisations. To support instantiation of Software Ecosystems for farming, a Reference Architecture was developed. This Reference Architecture describes how software developers, farmers and other stakeholders can collaborate to enable development, configuration and instantiation of integrated software solutions. More specifically, it can be used to map, assess, design and implement Farm Software Ecosystems to help to decrease current integration problems. The reference architecture comprises five main components:
Actors, which are basically app developers, business architects/software developers and end-users, i.e. farmers that finally use the configured ICT components and services;
Platform that enables configuration of Atomic Application Components into integrated information systems for farmers;
Open software enterprise that manages the relation between the actors and the platform;
Business services that support software configuration, development and hosting;
ICT Components that are configured application components from multiple vendors allowing seamless data exchange based on standards
After the design the reference architecture was first verified based on the requirements. Second, semi-structured interviews were held with experts to validate the model. Moreover, the assessment and mapping functionally was validated by using the reference architecture in a case study, in which two existing farm software ecosystems were assessed and mapped.
The Reference Architecture for Farm Software Ecosystems mainly addresses the organizational part of this research question. The technical part on the configuration of different ICT components into integrated solutions was not yet sufficiently covered in the Reference Architecture for Farm Software Ecosystems. Therefore we designed RAAgE 2.0 to improve the integrating capabilities of ICT Components, focussing on configuration and ICT Mass Customisation. In this research RAAgE 1.0 was extended into RAAgE 2.0 supporting technical aspects related to configuration of ICT Components by providing a hierarchical configuration methodology. This methodology divides configuration in two steps (i) business process configuration and (ii) software configuration. To enable business process configuration the model comprises three reference models, i.e. on products, processes and resources. The dependencies between these models are defined in rules that define possible combinations of products, processes and resources and that constrain the configuration of farm-specific models i.e. instances. The reference model also includes a configuration tree and templates. Templates describe a set of pre-configured product, process and resource models for typical cases. Variety in farm business processes can be modelled with business process variants. Such a variant realizes a similar kind of business services (e.g. basic fertilization, precision fertilization). Each variant has partly overlapping business processes and resources and unique ones. RAAgE 2.0 provides insight into these specific and generic parts. The other part of the methodology, software configuration, is divided in two additional sub-steps. The first sub-step is to create configuration templates that describe the required (generic) application services (capability types) to support specific business process variants. These configuration templates describe the interactions between the capability types. This sub step is typically performed by a business architect in close collaboration with software developers. The second sub-step is the selection and configuration of the specific capability of a capability type. Capabilities can be offered by atomic application components of multiple vendors that need to be selected. This second sub-step is performed by a business architect, in close collaboration with a farmer. With this extension RAAgE 2.0 supports (i) development of ICT components that fit within an ICT Mass Customisation and Best-of-Breed approach, (ii) selection of ICT components based on business processes that they should support and (iii) getting insight into configuration of different ICT components into an integrated farm information system.
To substantiate that our artefacts contribute to realizing ICT Mass Customisation in combination with Best-of-Breed in arable agriculture a proof of concept was developed. A proof of concept is defined as a phase in development, in which experimental hardware or software is constructed and tested to explore and demonstrate the feasibility of a new concept. Realizing ICT Mass Customisation requires: (i) software modularity, (ii) an information integration platform, (iii) component availability, (iv) configuration support and (v) reference information models. To fulfil these requirements a design was developed and instantiated for a specific use case on late blight protection in potato growing for a specific farmer in The Netherlands. For that purpose we:
configured the business processes that are involved in late blight protection using RAAgE 2.0 to identify which advanced ICT components are needed to support this process for this farmer;
developed the required advanced ICT components that were identified in the previous step using the FIspace platform. These components were provided by different app developers from 5 different European countries;
configured a composite application component within the FIspace platform using the configuration framework of RAAgE 2.0. This included involvement of 5 different European organizations;
instantiated and executed the application component within the FIspace platform for this specific farmer.
This resulted in prototype software that showed how we can configure business processes and multi-vendor atomic application components into a composite component to support late blight protection in potatoes for a specific farmer. It was made plausible that this approach is also applicable to other cases to create software able to support other business processes in agriculture.
Within this research we developed artefacts and substantiated that they facilitate collaboration between the actors involved and can help to develop ICT Components that improve farm enterprise integration. Still, to make ICT Mass Customisation and Best-of-Breed a more common practice, future research is required. In this research we recommend to focus on:
Development of business models to gain insight into the motives of software developers to become part of Farm Software Ecosystems. Insight into these motives can enhance the adoption of Software Ecosystems for agriculture, which makes the concept of ICT Mass Customisation more feasible.
Improving configuration of atomic application components and supporting tools as this is currently still cumbersome. We recommend focussing on one specific case to dig into all details of the case. Such a detailed description will be re-usable for many other farm business processes such as fertilization, other types of crop protection, seeding and harvesting.
Although, there are still hurdles to take we recommend continuing this research line as it can result in improved farm enterprise integration and adoption of advanced farm management styles by famers. This can enable farm enterprises to increase food production, while producing in a sustainable, safe and transparent manner.