Data that farmers generate and collect are comparable to the business secrecies of other economic actors because the information from data means an advantage in knowledge and competition.However, as data are not physical and are easily duplicable, it was until today not possible to define a right on data ownership.Within the intellectual property law, it is difficult to dispense justice on a copyright law basis, concerning data that are regularly not the result of thinking, creating processes.Mere, unordered collections of “raw data” are not protected in this context, and also the GDPR does not apply in such cases.But farmers’ data in the normal cloud-based communication ways are exposed to several third parties.Data security, data ownership, and data safety, also for impersonalized data, need to be addressed for example by entering the corresponding parties into contracts so that impersonal data are as save as personal data are.Another possibility to personalize data is offered by the block chain technology itself and furthermore the use of NFT.Data stored in this way in a block chain still can be copied but the authenticity of data is safely defined.Furthermore, decentrality amongst clouds is recommendable to increase data security.Vogel summarizes, that data sovereignty is hardly guaranteed and whoever is in possession of the data can use it as wished.Undesirable market dynamics that could lead to extortion of individuals and the lack of criteria defining data sovereignty led the German Conference of Justice Ministers to reject the creation of defined data sovereignty.Contracts with the according service providers could bring data sovereignty to farmers but may also lead to unfavourable “lock-in” effects.To strengthen data sovereignty several activities took and still take place.There was an industry recommendation in Germany heading to assure full control of data sovereignty and rights of use to the farmer.On a European level, there is “The EU code of conduct on agricultural data sharing” heading the same aims and centralizing clear contracts between collaborative parties after the principles of the code.

The number of signatories representing farmers, industries, and cooperatives is gratifying and promising.However, still,vertical rack system several security issues on different layers of digital farming systems are present.Therefore, concepts are needed which provide suggestions of systems that give control to the data originators or trusted parties of them and provide open, simple, and interoperable solutions, facilitating the introduction and continuation in digitization for every farm size.In the authors’ point of view, absolute data sovereignty belongs to the farmers concerning any data which are generated in the fields of their responsibility.Farmers are the owners of these data, this needs to be legally protected and substantiated by unique identification methods of the corresponding data.In the near future, this needs to be taken care of, from the beginning of a cooperation, service, or machine purchase.The EU code of conduct is a good basis for the beginning of a legal process.Farmers are encouraged to use the “Code of conduct” for the critical examination of appropriate solutions.Defining the term IoT as a technology still seems misleading.Paraforos unites numerous definitions to the common denominator of a technological paradigm.One might consider IoT as an integrating network of technologies interacting and exchanging data in an ideally interoperable way.Kim categorizes the applications of IoT fourfold in management systems, monitoring systems, control systems, and unmanned machinery, which include respectively a perception layer where physical properties are recognized, the network layer realizing M2M communication, and the application layer where the data is being used or processed to information.Chaudhary conducted a case study on AGCO’s Fuse Technology’s ‘Connected Farm Services’ as a commercial IoT example covering farm management, standard field works, monitoring, and dealer telemetries.They mention as a major vulnerability issue the centrally connected network and the therefore comprehensive need of cyber security measures.IoT offers practical and also monetary benefits at farm level if it is tailored to the needs of the user as realized in the specific example of sugar cane production shown by van de Vooren.Increased invention and application of IoT solutions lead to a strongly increased number of devices and data traffic which reveals the transmission and computing limits of cloud solutions.

Therefore, applying edge and/or fog computing, data processing is being decentralized from the cloud, on or close to the data acquiring device, to the edge of the network, leveraging this problem.This results in a lowered latency by avoiding edge to cloud or edge to enterprise server round trips.By processing computing workload on edge devices or edge servers/gateways network congestion is minimized.Furthermore, data privacy and security are increased by installing access control options at the corresponding edge device/ gateway.Also, storage and intelligence capacities from the cloud can be mirrored on the edge servers.Fog computing is usually localized one level beyond edge computing in the network.Like this, a decentralized and redundant infrastructure evolves and leads to more independence from centralized cloud solutions.As it was mentioned by Jha and Patidar in a market report: “The global autonomous farm equipment market is projected to expand at over 10% CAGR through 2031, and top a market valuation of US$ 150 billion by 2031”.Numerous Start-Up companies all over the world develop robots and autonomous systems for agricultural purposes.Also here the amount of data that is and will be generated and has to be transferred, increases, and has to be processed in high quality and quantity to ensure continuous functionality to the autonomous systems also using AI.The actual working speed and efficiency of autonomous systems are still low, which makes working hours in narrow time windows of certain works even more crucial/critical.Furthermore, a full customer service is needed, which also is a matter of costs.For technical issues or hardware problems, a service technician needs to be present near-time, and also for remote services in case of software issues farmers need real-time online support.Thus, these solutions need a solid and seamless digital infrastructure to exploit the full potential of each device.Due to the competition of OEMs, whose interest in data sharing is limited, the development of infrastructure is always lagging behind the development of single devices.But to be forearmed for the use of autonomous systems, the infrastructure conceptualization and development should be forced.A nice example shows Saito by using XML standards for directing robots to target plants.Today AI in agriculture is used in decision support systems, expert systems, and agricultural predictive analytics.Digital twin methods are dealt with for modelling future scenarios and preventing disadvantageous circumstances.Furthermore, data over periods of decades could reveal regionally optimized crop rotations, cultivar selections, and cultivation strategies by the application of AI.

However, to reach this, the form of data storage, transmission, and processing must orient on international standards to ease the interoperable interactions of systems that can reflects entire agricultural processes.Slurry application is an often delegated application for small and medium-sized farms today.Customers can order the service of a self propelled slurry applicator using precision farming technologies like auto-steer, online NIRS nutrient analysis, and site specific slurry application including section control, which farmers themselves would not invest for, for their own farms alone.If VRA is conducted, the data transmission goes via a USB stick.Which can be inconvenient, due to the risk of data and hardware loss, and can be time-consuming if changes in the application map are necessary.The data upload to the cloud can be tedious in rural areas because of narrow bandwidth and low mobile network coverage but also occur to be minimized by weak performance of the cloud services.FMISs lack interfaces for seamless data transmission and task execution.Task documentation also demands increased knowledge and skills of drivers to organize and overview data of multiple farms.The billing of the single tasks hereafter is done by hand and transferred to the computer manually for finalizing the invoicing.To summarize, high-tech machines and digital farming components are available and implemented but are barely used to their full potential up to automated documentation and billing, due to the lack of infrastructure and/or not interoperable or isolated components.To meet the responsibility for a critical infrastructure and the weather-dependent and therefore, time-critical conditions in agriculture, specific requirements concerning the digital infrastructure are to be fulfilled.If services or devices, which generate or need data, do not work at the application date, in most cases farmers will continue without it.This chapter aims to specify the requirements of regional and on-farm ICT infrastructures.Farmers’ independence of the susceptibility of centralized systems and the straightforward inclusion of small and middle scaled farms are the main focus.

For the application of field measures like seeding, plant protection, or fertilization, various information can be used and are required to achieve maximum efficiency.The existing, actual, and forecasted data of soil, plants, and weather conditions are decisive.Therefore, these agronomic data need to be easily accessible to farmers if they don’t acquire them themselves.The structure and the format of these data must further be readable and processible in farmers’ FMISs.The combination and systematic processing of these data should be straightforward via accessible algorithms and knowledge bases which are accessed and used by the management software of the farmers.Like this,mobile grow rack farmers have the information to make profound decisions which they need directly in crop management and field applications.Additionally, to existing data like yield maps or satellite images, these data are to be combined and supplemented in the prescription map by merging also dynamic and non-deterministic parameters by farmers as Heiß et al. showed in their work.Most FMISs and other digital farming components are and will be based on cloud computing solutions.Reasons are the advantages for the companies like better customer support, instant new updates, and decentral data management.However, to correspond to the need for resilience, centralized cloud-based services can become redundant and fail-safe by decentralization.In Fig.1 it is highlighted where decentralization can be realized.Decentralization is required within the cloud layer, by placing functionalities redundantly within further clouds ideally with geographically remote located servers.In the lower layer of fog computing, an additional, regional server location, driven by an MR or a local government institution , can be implemented to expand decentralization on a regional level.Server maintenance and corresponding storage, back up and computing capacities must be provided by the responsible institution.At the farm level, a farm server tailored to the farms’ needs is required.It must be able to define the access rights to farm data for third parties and has to ensure the required data protection and data privacy.The red arrows in Fig.1 show the current means of data acquisition, communication, and processing in the corresponding clouds: From sensor/ device to OEM cloud to FMIS and back or further to other destinations.

The path over the cloud is the preferred way in the suggested FDFS.Alternatively, and/or additionally, in case of disconnection or outage, data can be sent to the farm server and be processed there or on the district servers to maintain functionalities.For the interpretation of this data, a certain level of intelligence in offline software on the farm server or secondary servers is needed because farmers hardly deal with raw data.The prerequisite for the common communication path in Fig.1 and as well for the connection between farm and regional server locations is the expansion of broadband internet over landline connection and mobile network coverage especially to rural areas where farms are located.For realizing the resilient communication path , a local on-farm network is required to sustain data communication between sensors/devices and farm server/farm PC.To enable communication during internet outages between farmers, farmers and contractors/MR, etc., a network communication technology is required which ensures data communication over an area expansion that covers the region of the farmers, their partners, and contractors.Here only small and necessary amounts of data need to be sent for example to communicate basic data of an application task.Furthermore, especially for small and middle scaled farmers who have to communicate with different brands through MR and contractors, interoperability of digital technologies is a major requirement.This concerns all components of an FDFS beginning at data structures on sensor level and ending at interfaces to sales and purchase.Finally, the FDFS needs to ensure data safety and security.The more digitization proceeds the more important this requirement becomes.The systems must be protectable against unwanted and malicious access.Traceability of data must be given so that it is known to the farmers who use their data and for which purpose.Furthermore, information about the production process is offered to the customer which increases trust in farmers’ practices.