Accurate detection and prediction of required water for irrigating a farm plays an important role in both the quality and quantity of its products. Machine learning is considered as an appropriate solution to detect and predict the required water of a farm if enough and proper data can be provided. To this end, the farms of a region need to share their data. However, many farms because of privacy issues may not want to share their data with others. In our previous work, we proposed Federated Learning (FL) as a solution. While FL can provide an accurate prediction for the required water of farms, it ensures data privacy. In this paper, we implement FL for predicting the required water of a farm. We compare the performance of four different versions of FL (i.e., FedAvg, FL with feedback, FL with influence and FL with feedback and influence) to find which one is more appropriate for the application of predicting the required water of a farm. The results show that FL with influence can provide the best performance.

The Passive Optical Network (PON) has been evolving continuously in terms of architecture and capacity to keep up with the demand for high-speed Internet access in the access network segment. Recently, integration of Software-Defined Networking (SDN), which provides programmable and (logically) centralized network control, with PON has attracted intensive research interest to further enhance PON performance and reduce operational and capital expenditure. Although PON is regarded as an energy-efficient access network solution, it would be a major contributor for increasing energy consumption in the access network segment because of its higher penetration rate than other access network technologies. Over the past several years, the major standardization bodies like IEEE and ITU-T, and research communities have engaged in introducing energy-efficient PON solutions. This article familiarizes readers with PON evolution in terms of capacity, architecture, and its integration with virtualization and SDN based control. We present a comprehensive survey of the energy conservation research efforts in PON starting from conventional PON to SDN based PON leveraging virtual and physical network functions. This article also presents contemporary energy-efficient standardization activities in IEEE and ITU-T. To the best of our knowledge, to date, this article is the first most comprehensive survey on energy saving research and standardization on PON. We summarize the lessons learned from the recent advancements, identify important challenges ahead and outline several future research directions that can contribute to further advancement of energy-efficient PON.

The Artificial Intelligence of Things (AIoT), a combination of the Internet of Things (IoT) and Artificial Intelligence (AI), plays an increasingly important role in smart agriculture (SA). AIoT has been adopted in many applications including agriculture, such as crop yield estimation, soil and water conservation, pest and disease detection and supply chain management. While there are plenty of studies on AIoT applications in healthcare, smart cities, manufacturing, and transportation, SA still has a small share of the reported research. This paper presents a comprehensive review of the existing literature in AIoT and Federated Learning (FL) for SA. It identifies current and potential challenges and provides research direction for the future investment in both academia and industry.

Driven by rising energy costs and growing environmental concerns, sustainability-aware network operations in 5G-Advanced research and standardization have gained increased attention. In this paper, we introduce a set of potential solutions with focus on energy efficiency and carbon emission reduction. These solutions are proposed to be studied and included in the 5G 3GPP standard. In particular, we focus on solutions aimed at the cross-domain energy optimization (encompassing Core Network (CN), Radio Access Network (RAN) and Edge), inter-operator optimizations and the service specific implications of these optimizations (e.g., energy-aware charging and analytics).

The irrigation scheduling system can play an important role in an environment with multiple farms and shared and limited available water. An accurate prediction of required water for every farm can help the irrigation scheduling system to manage better and optimize the usage of water. However, many farmers are reluctant to share their farms data with others. As a solution, this paper proposes to use the Federated Learning (FL) technique for predicting irrigation requirements where the farmers do not need to share their farm data. To our knowledge, this paper is the first one that proposes using FL for solving the problem mentioned above. We propose a use case to show the feasibility of the idea and how FL can be adapted and utilized for predicting irrigation requirements in an environment with multiple farms and a shared water source.

Arid and semi-arid regions are characterized by the limited available water. We consider a scenario in which there is one shared water resource among multiple farms as well as a water-sharing policy. It is required to irrigate all or some of these farms currently, but there is not enough water to properly irrigate them. This paper proposes a Multi-farm Irrigation Scheduling System, called MISS. The objective is to save the farms that need urgent irrigation while still satisfying all farm owners according to the water-sharing policy. MISS uses a two-level decision-making method where it considers a priority for the farms that need immediate irrigation as well as a priority for blocks of every farm. MISS is designed based on a realistic scenario where, because of the limited shared water, we sometimes are not able to save all the farms.

This paper defines a smart home use case to automatically adjust home temperature and/or hot water. The main objective is to reduce the energy consumption of cooling, heating and hot water systems in smart homes. To this end, the residents set a temperature (i.e., X degree Celsius) for home and/or hot water. When the residents leave homes (e.g., for work), they turn off the cooling or heating devices. A few minutes before arriving at their residences, the cooling or heating devices start working automatically to adjust the home or water temperature according to the residents’ preference (i.e., X degree Celsius). This can help reduce the energy consumption of these devices. To estimate the arrival time of the residents (i.e., drivers), this paper uses a machine learning-based street traffic prediction system. Unlike many related works that use machine learning for tracking and predicting residents’ behaviors inside their homes, this paper focuses on predicting resident behavior outside their home (i.e., arrival time as a context) to reduce the energy consumption of smart homes. One main objective of this paper is to find the most appropriate machine learning and neural network-based (MLNN) algorithm that can be integrated into the street traffic prediction system. To evaluate the performance of several MLNN algorithms, we utilize an Uber’s dataset for the city of San Francisco and complete the missing values by applying an imputation algorithm. The prediction system can also be used as a route recommender to offer the quickest route for drivers.

A cluster-based crop recommendation system categorizes the crop candidates into several groups or classes (e.g., based on soil and environment parameters similarities). After receiving a request from a farmer, it recommends the most appropriate group of crops to the farmer. Proposing a group of crops (i.e., more than one crop) can allow farmers to consider their personal interests as well. In addition, the cluster-based crop recommendation system can reduce the complexity and utilized resources (e.g., compute resource). As the main contribution of this paper, we evaluate the performance of different Machine Learning (ML) algorithms to find the most appropriate one for using in the cluster-based crop recommendation system.

Smart Agriculture plays an important role in providing foods and manufacturing products. It covers various food-related applications, from weed and pest identification to crop recommendation. Selecting an appropriate crop to grow for a farm directly affects the quality and quantity of products. This paper proposes a user-friendly crop recommendation system called UACR. The users only need to insert their farm location and the system can automatically obtain other necessary information (e.g., rain level, soil type) based on the requested location and recommends the best crop(s) for growing in the location. To support and implement the proposed system, an architecture based on Artificial Intelligence of Things (AIoT) technology, as one of the main enablers for smart agriculture, is also proposed.

Finding an available parking place has been considered a challenge for drivers in large-size smart cities. In a smart parking application, Artificial Intelligence of Things (AIoT) can help drivers to save searching time and automotive fuel by predicting short-term parking place availability. However, performance of various Machine Learning and Neural Network-based (MLNN) algorithms for predicting parking segment availability can be different. To find the most suitable MLNN algorithm for the above mentioned application, this paper evaluates performance of a set of well-known MLNN algorithms as well as different combinations of them (i.e., known as Ensemble Learning or Voting Classifier) based on a real parking datasets. The datasets contain around five millions records of the measured parking availability in San Francisco. For evaluation, in addition to the cross validation scores, we consider resource requirements, simplicity and execution time (i.e., including both training and testing times) of algorithms. Results show that while some ensemble learning algorithms provide the best performance in aspect of validation score, they consume a noticeable amount of computing and time resources. On the other hand, a simple Decision Tree (DT) algorithm provides a much faster execution time than ensemble learning algorithms, while its performance is still acceptable (e.g., DT’s accuracy is less than 1% lower than the best ensemble algorithm). We finally propose and simulate a recommendation system using the DT algorithm. We have found that around 77% of drivers can not find a free spot in their selected destinations (i.e., street or segment) and estimated that the recommendation system, by introducing alternative closest vacant locations to destinations, can save, in total, 3500 min drivers searching time for 1000 parking spot requests. It can also help to reduce the traffic and save a noticeable amount of automotive fuel.

As an increasing amount of data processing is done at the network edge, high energy costs and carbon emission of Edge Clouds (ECs) are becoming significant challenges. The placement of application components (e.g., in the form of containerized microservices) on ECs has an important effect on the energy consumption of ECs, impacting both energy costs and carbon emissions. Due to the geographic distribution of ECs, there is a variety of resources, energy prices and carbon emission rates to consider, which makes optimizing the placement of applications for cost and carbon efficiency even more challenging than in centralized clouds. This paper presents a Dynamic Energy cost and Carbon emission-efficient Application placement method (DECA) for ECs. DECA addresses both the initial placement of applications on ECs and the re-optimization of the placement using migrations. DECA considers geographically varying energy prices and carbon emission rates as well as optimizing the usage of both network and computing resources at the same time. By combining a prediction-based A* algorithm with a Fuzzy Sets technique, DECA makes intelligent decisions to optimize energy cost and carbon emissions. Simulation results show the ability of DECA in providing a tradeoff and optimizing energy cost and carbon emission at the same time.

In recent years, the number of objects connected to the internet have significantly increased. Increasing the number of connected devices to the internet is transforming today’s Internet of Things (IoT) into massive IoT of the future. It is predicted that, in a few years, a high communication and computation capacity will be required to meet the demands of massive IoT devices and applications requiring data sharing and processing. 5G and beyond mobile networks are expected to fulfill a part of these requirements by providing a data rate of up to terabits per second. It will be a key enabler to support massive IoT and emerging mission critical applications with strict delay constraints. On the other hand, the next generation of software-defined networking (SDN) with emerging cloudrelated technologies (e.g., fog and edge computing) can play an important role in supporting and implementing the above-mentioned applications. This paper sets out the potential opportunities and important challenges that must be addressed in considering options for using SDN in hybrid cloud-fog systems to support 5G and beyond-enabled applications.

Online Social Network Services (OSNSs) are changing the fabric of our society, impacting almost every aspect of it. Over the past few decades, an aggressive market rivalry has led to the emergence of multiple competing, “closed” OSNSs. As a result, users are trapped in the walled gardens of their OSNS, encountering restrictions about what they can do with their personal data, the people they can interact with, and the information they get access to. As an alternative to the platform lock-in, “open” OSNSs promote the adoption of open, standardized APIs. However, users still massively adopt closed OSNSs to benefit from the services’ advanced functionalities and/or follow their “friends,” although the users’ virtual social sphere is ultimately limited by the OSNSs they join. Our work aims at overcoming such a limitation by enabling users to meet and interact beyond the boundary of their OSNSs, including reaching out to “friends” of distinct closed OSNSs. We specifically introduce Universal Social Network Bus (USNB) , which revisits the “service bus” paradigm that enables interoperability across computing systems to address the requirements of “ social interoperability .” USNB features synthetic profiles and personae for interaction across the boundaries of closed and open and profile- and non-profile-based OSNSs through a reference social interaction service . We ran a 1-day workshop with a panel of users who experimented with the USNB prototype to assess the potential benefits of social interoperability for social network users. Results show the positive evaluation of users for USNB, especially as an enabler of applications for civic participation. This further opens up new perspectives for future work, among which includes enforcing security and privacy guarantees.

In order to improve locality aspects, new Cloud-related architectures such as Edge Computing have been proposed. Despite the growing popularity of these new architectures, their energy consumption has not been well investigated yet. To move forward on such a critical question, we first introduce a taxonomy of different Cloud-related architectures. From this taxonomy, we then present an energy model to evaluate their consumption. Unlike previous proposals, our model comprises the full energy consumption of the computing facilities, including cooling systems, and the energy consumption of network devices linking end users to Cloud resources. Finally, we instantiate our model on different Cloud-related architectures, ranging from fully centralized to completely distributed ones, and compare their energy consumption. The results show that a completely distributed architecture, because of not using intra-data center network and large-size cooling systems, consumes between 14 and 25 percent less energy than fully centralized and partly distributed architectures, respectively. To the best of our knowledge, our work is the first one to propose a model that enables researchers to analyze and compare energy consumption of different Cloud-related architectures.

Network Functions Virtualization (NFV) is a promising solution to provide cost-efficient, scalable and rapid deployment of network services. It allows the implementation of fine-grained services as a chain of Virtual Network Functions (VNFs). In order to place the VNF chains in the network, several cost-efficient methods have been already proposed. However, a few works have considered order of VNFs to reduce the cost. In this paper, we propose a Simple and Effective Technique (SET), which can be easily combined with VNF placement methods to dramatically improve their cost efficiency by considering different possible orders for the VNFs in the chain. As a proof-of-concept, we combine the proposed technique with one of the recent cost-efficient VNF placement and chaining algorithms called CCVP. The results show that the combination can yield significantly better cost than CCVP operating solo.

Value-added services (VASs) are an integral part of todays Content Delivery Networks (CDNs). They can be implemented as a chain of Virtual Network Functions (VNFs). These chains need to be placed in an efficient way in CDNs in order to optimize quality of service (QoS) for end-users(EUs) while minimizing cost for providers. We formulate the problem as an Integer Linear Program (ILP) and propose a Pattern-based Cost-efficient Proactive VNF placement and chaining (PCPV) algorithm. The objective is to find the optimal number, location and chaining of VNFs in such a manner that the cost is minimized while QoS is met. Apart from cost minimization, the support for large-scale CDNs with a large number of physical machines (servers) and EUs is an important feature of the proposed algorithm. Through simulations, the algorithm behavior for small-scale to large scale CDN networks is analyzed.

Network services have been significantly increased in today's enterprise networks. The time and cost of deploying these services are recently considered as critical challenges for enterprise networks. Network Functions Virtualization (NFV) is a promising solution to offer cost-efficient, scalable and more rapid deployment of such services. It allows the implementation of fine-grained services as a chain of Virtual Network Functions (VNFs). These chains need to be placed in the network. The chain placement is critical since it effects on both quality of service (QoS) and the provider cost. This paper formulates the problem of VNF placement and chaining as an Integer Linear Program (ILP) and proposes a Cost-efficient Centrality-based VNF Placement and chaining algorithm (CCVP). The objective is to find the optimal number of VNFs along with their locations in such a manner that the provider cost is minimized. Apart from cost minimization, the support for large-scale environments with a large number of servers and end-users is an important feature of the proposed algorithm. Finaly, the algorithm behavior is analyzed through simulations.

Distributed clouds have recently attracted many cloud providers and researchers as a topic of intensive interest. High energy costs and carbon emissions are two significant problems in distributed clouds. Due to the geographic distribution of data centers (DCs), there are a variety of resources, energy prices and carbon emission rates to consider in a distributed cloud, which makes the placement of virtual machines (VMs) for cost and carbon efficiency even more critical than in centralized clouds. Most previous work in this field investigated either optimizing cost without considering the amount of produced carbon or vice versa. This paper presents a cost and carbon emission-efficient VM placement method (CACEV) in distributed clouds. CACEV considers geographically varying energy prices and carbon emission rates as well as optimizing both network and server resources at the same time. By combining prediction-based A* algorithm with Fuzzy Sets technique, CACEV makes an intelligent decision to optimize cost and carbon emission for providers. Simulation results show the applicability and performance of CACEV.

User Location Discovery (ULD) is a key issue in smart home ecosystems, as it plays a critical role in many applications. If a smart home management system cannot detect the actual location of the users, the desired applications may not be able to work successfully. This article proposes a new taxonomy with a broad coverage of ULD methods in terms of user satisfaction and technical features. In addition, we provide a state-of-the-art survey of ULD methods and apply our taxonomy to map these methods. Mapping contributes to gap analysis for existing ULDs and also validates the applicability and accuracy of the taxonomy. Using this systematic approach, the features and characteristics of the current ULD methods are identified (i.e., equipment and algorithms). Next, the weaknesses and advantages of these methods are analyzed utilizing ten important evaluation metrics. Although we mainly focus on smart homes, the results of this article can be generalized to other spaces such as smart offices and eHealth environments.