Connected vehicles offer substantial potential for improving traffic safety and enhancing comfort services. However, maintaining consistent connections in dynamic vehicular environments remains a persistent challenge, especially due to the need for seamless handovers (HO) between cellular towers as vehicles travel at high speeds. The limited communication range often leads to frequent HOs and connection drops, which can degrade the reliability of services and resources. The virtual cell (VC) paradigm can help mitigate the challenges of the limited communication range in 5G networks for high-mobility, ultra-dense scenarios. To address these challenges, we propose a mobility-oriented approach using a multi-output regression model named MSVR to manage VCs. Our proposed approach ensures stable HO decision-making by dynamically managing VCs based on predictive mobility, considering network attributes and real-time data: speed, signal strength, and network quality. Realistic simulations and extensive result analyses have been conducted to demonstrate the effectiveness of the proposed MSVR approach over existing works in terms of throughput, frame loss ratio, number of HO, and size of VC.

Facing the growing demand for computational resources in Vehicular Edge Computing (VEC) services is a constant challenge for researchers. New scenarios, topologies, variables, and priorities emerge in an environment with many tasks. Making the best decision when selecting one task over another can prevent an important security task from being left out in favour of a simple entertainment task, for example. Thus, this work proposes a LOA-based priority-driven task allocation in the vehicular edge. The mechanism relies on the social behaviour of the Lion Optimization Algorithm (LOA) to select the best cluster and the best task, prioritizing the selection of tasks by category. However, the LOPRIVE has a behaviour of updating the position and checking the strength of the task to the target cluster, allowing the mechanism also to serve the other categories efficiently. The proposal was compared with other algorithms known in the literature and managed to serve more tasks, maximizing the allocated resources and maximizing the service and resources of tasks in priority categories.

Technological advancements in cellular networks have enabled to surpass many challenges in telecommunications, but some features remain restricted, such as throughput, packet loss, and latency. User equipment (UE), including mobile, smart devices, vehicles, IoT devices, and smart city infrastructure, requires seamless connectivity to share data and resources effectively. Multiple radio access technology (multi-RAT) scenarios offer a solution to the limitations of individual RATs by combining their strengths. Determining the optimal RAT for traffic steering (TS) in multi-RAT scenarios is challenging due to factors such as high mobility, ultra-dense networks, overall dynamic network conditions, and the unique needs of individual users. In this context, we propose a bi-level approach, called BIL-TS, which includes (i) centrally determining the optimality of RATs and (ii) locally making TS decisions. BIL-TS utilizes the Actor-Critic SARSA Reinforcement Learning (ACS-RL) in level (i) to evaluate the optimality of RATs by considering the entire network, and level (ii) leverages Linear Regression (LR) to make decisions of TS to the optimal RAT based on specific requirements of each UE. Simulation results show that our proposed BILTS approach significantly enhances efficiency in TS, resulting in higher throughput, reduced packet loss, and lower latency.

The 5 G networks and their Open Radio Access Networks (O-RAN) architecture face significant challenges in resource management due to their extended flexibility and technological diversity. O-RAN’s open, disaggregated architecture creates a heterogeneous environment that requires effective integration and analysis of data from various components. In this context, accurately predicting the computational utilization of virtual base stations (vBS) emerges as a critical challenge, essential for optimizing resource allocation and addressing the dynamic demands of 5 G and 6 G O-RAN networks. Traditional forecasting techniques often struggle with the complexity and variability of data in this scenario, necessitating advanced AI and ML approaches. We propose an Adaptive Neuro-Fuzzy Inference System (ANFIS) model for multi-target regression to predict CPU utilization in vBS. By combining neural networks and fuzzy logic, ANFIS enhances both prediction accuracy and interpretability, making it ideal for the complexities of O-RAN 5G/6G networks. Our model, tested on publicly available O-RAN datasets, outperforms traditional ML methods. These results position ANFIS as an effective tool for optimizing resource management and enabling transparent decision-making in 5G/6G infrastructures, providing valuable support for network operators seeking efficient and scalable solutions in the evolving ORAN landscape.

SpaceFOM is the reference standard adopted by space agencies for simulating space missions. Although specifically designed for handling space systems, it currently faces a significant limitation when simulating interplanetary missions: a fixed Federation Time Step. This constraint hinders accurate and flexible modeling of space missions, which limits the dynamic changes of simulation pace, especially during critical phases that require particular temporal granularities. This work proposes extending the SpaceFOM standard to address this issue by enabling dynamic adjustment of Federation Time Step granularity. The proposed solution allows fine-grained time steps for mission-critical phases and coarse-grained steps for extended phases while smoothly combining continuous temporal progression. Furthermore, the proposed solution is general-purpose and can be applied to other domains requiring dynamic temporal granularity.

5G technology offers high bandwidth, stability, and reliability among connected vehicles, which is necessary for increasing data sharing in intelligent transportation. While providing these benefits with its small cellular range and densification, it also presents a challenge in frequent handovers (HOs). This issue can result in unnecessary HO, HO failures, and ping-pong effects, negatively impacting service delivery and compromising safety data sharing. A learning-oriented proactive HO decision-making strategy can ensure connection stability by making HO decisions based on real-time scenarios. This paper presents a high mobility and ultra-dense network-aware proactive HO decision-making (PAHD) approach, efficiently ensuring stable connectivity by predicting future HO. PAHD consists of two parts (i) Gaussian Process Regression for mobility prediction and (ii) Bidirectional Long Short-Term Memory for the prediction of network traffic density. Realistic simulated analyses have shown that PAHD significantly improves efficiency in HO decision-making.

Ensuring connection stability is crucial for both vehicular safety and user experience. With the increasing amount of data sharing among connected vehicles, there is a need for more bandwidth, stability, and reliability. While 5G technology can offer these benefits with its small cellular range and densification, it also presents a challenge in frequent handovers (HOs). This issue can result in unnecessary HO, HO failures, and ping-pong effects, negatively impacting service delivery and compromising safety data sharing. To this end, we present High- mobility and Ultra-density Aware Handover decision-making (HMUD-H) approach using the SARSA Reinforcement Learning algorithm for connection management, which efficiently makes HO decisions to ensure stable connectivity. The HMUD-H algorithm is adaptable and can handle dynamic, highly mobile, and ultra-dense vehicular networks. Realistic simulated analyses have demonstrated that our algorithm significantly reduces the number of HOs, average cumulative HO time, HO failures, and ping-pong effects, thus improving overall connection stability.

The Internet of Vehicles is a true enabler of in-telligent transportation, but it also faces communication and connectivity challenges due to the heterogeneity and high mobility of vehicles. More reliably, defining the vehicular network edge substantially helps cope with the highly dynamic environment of vehicles. Previous approaches have employed intelligence, prediction, optimization, and incentive modelling strategies to reduce communication challenges. These approaches still experience flexibility, scalability, and latency challenges, reducing service quality. Our work integrates Graph Neural Networks (GNNs) and clustering methodologies to define and maintain the vehicular edge. We propose an iterative GNN-based vehicular edge clustering framework with three different iterative learning procedures to facilitate the extraction of intricate temporal, spatial and functional patterns within vehicular networks. The experiments demonstrate that the proposed approach provides a promising solution to incorporate insights from trending communication and mobility features and improve the responsiveness of sporadic connectivity in a dynamic environment.

5G and vehicular networks have enabled Intelligent Transportation Systems (ITS) with better safety and infotainment services where connected vehicles are critical components for data sharing. However, a stable connection is mandatory to transmit data successfully across the network. The 5G technology enhances bandwidth, stability, and reliability but suffers from low communication ranges, which results in frequent and unnecessary handovers and connection drops. In this paper, we introduce a user-centric approach, Factor-distinct SARSA Reinforcement Learning (FD-SRL), which combines a time series data-oriented model LSTM and adaptive method SARSA Reinforcement Learning for Virtual Cell (VC) and handover (HO) management. Our proposed approach maintains stable connections by reducing the number of HOs, given the fast-paced changes due to mobility, network load, and communication conditions. Realistic simulations demonstrated that FD-SRL reduced the number of HOs and the average cumulative HO time, showing potential improvements in connection stability for 5G-based ITS.

Connected vehicles enable massive data sharing and support intelligent transportation services. Consequently, a stable connection is compulsory to transmit across the network successfully, where 5G technology introduces more bandwidth, stability, and reliability. However, 5G communication is susceptible to frequent handovers and connection drops. A user-centric perspective helps cope with the smaller communication range in ultra-dense 5G networks. We thus introduce a Connectivity-oriented SARSA Reinforcement Learning (CO-SRL) algorithm for user-centric to efficiently handle virtual cell (VC) management and reduce the number of handovers (HO). The adaptability of the algorithm copes with high vehicular mobility and dynamic traffic and communication, deciding on in-rage cellular towers and VC size. Realistic simulated analyses showed CO-SRL reduced the number of handovers and the cumulative handover time.

With the increase of vehicles in cities and the technology used by these vehicles, there is also a need to use the technology made available by intelligent transport systems in an efficient and agile way. Edge computing services assist in the agile process of exchanging and sharing information and resources between vehicles. However, the limitations of edge services bring the need to optimize resource allocation processes. Thus, in this article we propose the MARIA, a mechanism for optimizing computational resources in Vehicular Ad Hoc Networks based on the particle swarm optimization bio-inspired algorithm. The ease of adaptation to various scenarios by a bio-inspired algorithm is presented in the work. In addition, the MARIA mechanism proved to be efficient when compared with techniques frequently used in the literature and was able to increase the amount of services accepted and reduced the amount of refused services.

Technological evolutions in intelligent transportation have enabled smart and connected vehicles to support novel safety and infotainment services. The provision of such services is guaranteed with effective sharing and allocation of resources for task offloading and processing. The use of vehicular fogs also helps this process by lowering the latency in communications and the resource share among the fog members. However, allocation in Fogs introduces challenges related to the intermittency of Fog vehicle nodes, clustering, topology changes, and resource allocation problems. The use of metaheuristic algorithms has been explored in several works to solve these optimization problems, such as resource allocation, clustering, task allocation, and network communications, especially regarding efficiency. We thus propose a bat bio-inspired decision-making algorithm for task allocation in vehicular fogs called AEGIS. AEGIS uses the cluster members and task parameters to do the decision-making process in the task allocation process that helps to choose the best vehicle of the fog to allocate a determined task. The AEGIS was compared to a GWO approach (meta-heuristic), Greedy, and Random (traditional) approaches. We considered allocated, denied, and lost tasks for the simulation criteria. AEGIS lost fewer tasks than the other algorithms and allocated more tasks than the traditional algorithms.

Vehicular fog computing is essential for quick data access to devices and applications, and caching data is a must that plays a vital role in enabling faster responses and making data more available. This work thus investigates using a multi-layer caching strategy to make data more readily available to the requesting nodes. The cache in the layers is distributive and updated on time based on the demand and criteria of the requests. We also distribute data using vehicular mobility by placing limited but significant data into the cache of the vehicle. These communication and data exchange types are standardized through policies in our methodology. This cache management design is extensively analyzed using established frameworks and vehicular networks through simulated environments.

The continuous emergence of new applications for Internet-connected road vehicles is imposing unprecedented re-source demand. Motivated by the incorporation of ever more resources into vehicles, this is a trend that, on the downside, is causing vehicular networks to become increasingly more challenging to manage. Departing from the proposition that computing capabilities can help overcome resource allocation problems in vehicular clouds (VCs), in this paper, we formulate ALTAIC, a coalition game to maximize resource utilization while dynamically load-balancing the usage among the VCs. First, we define a Shapley value-based strategy to determine the order in which the tasks are allocated. Then, with the marginal contribution of each task calculated, we employ a simple queue to allocate the tasks in VCs using these values. Finally, we conduct a comparative performance analysis of ALTAIC and relevant approaches. Simulation results show that the proposed solution allocates more tasks than the others and reduces 27.12% the load average of the VCs.

Vehicular Cloud Computing (VCC) exhibits many drawbacks with the demands of vehicular applications and intermittent network conditions. Vehicular Fog computing is a novel method for supporting and promoting the effective sharing of services and resources in urban areas. Diverse works on vehicular resource management have sought to handle the very dynamic vehicular environment using various methods, such as policy-based greedy and stochastic techniques. Nevertheless, high vehicular mobility poses many issues that compromise service consistency, efficiency, and quality. Adaptive vehicular Fogs incorporating Reinforcement Learning can deal with mobility and correctly distribute services and resources across all Fogs. Thus, we introduce an adaptive resource management model using cloudlet dwell time for resource estimation, mathematical formula for Fog selection, and reinforcement learning for iterative review and feedback mechanism for generating optimal resource allocation policy.

The allocation and management of vehicular resources are essential in enabling services in Vehicular Cloud networks. Combined Resource Units (CRUs) allow for relaxed resource management by utilizing vehicular resources clustered in virtualized units and easing the fulfillment of service requests. Previous works have used mobility-based models such as SMDP and MDP for resource allocation. However, these models have presented significant system overhead, which has impacted the network's performance. Therefore, this work proposes a game theory model for assigning CRUs to satisfy service requests. The utility function of CRUs is maximized by playing a non-cooperative game between service requests. Two different game models are implemented based on exhaustive search and pruning methods. These models use distinct utility functions, which differ in terms of distance and signal strength of the CRUs. Comparing the performance of the two models, the pruning model offers a 90% success rate towards satisfying service requests.

Traffic jams are an essential and continuous challenge in our cities, responsible for socioeconomic and environmental concerns and an ambitious traffic jams management agenda is urgent. The distributed solutions in the literature for Traffic Management Systems (TMS) are heavily based on beacon messages or proactive communication protocols to share vehicular traffic information among vehicles. Thus, these solutions are not scalable when the number of vehicles increases in the network — when there are traffic jams. To overcome these problems, we propose a new VANET-based traffic management system named CoNeCT: Predictive Congestion Control based on Collaborative Information Sharing for Vehicular Ad hoc Networks. CoNeCT's primary goal is to support vehicles' collaboration in analyzing, predicting, and managing congestion. The proposed system was designed to decrease the number of messages by using a novel road segment load assessment that improves traffic flow classification. Vehicles aware of traffic conditions share it with their neighbors, and they can also request traffic views whenever necessary. Additionally, vehicles can detect significant traffic variations and predict future traffic conditions to improve roads' overall traffic conditions, mitigating the congestion before it arises. Results obtained from an extensive performance analysis show CoNeCT's ability to reduce traffic congestion with a low impact on the wireless communication medium, outperforming the state-of-art systems.

As we move more deeply into information-oriented services and systems, we clearly observe the importance and impact of smart and connected vehicles for urban computing. New Cloud-enabled paradigms have boosted information and service sharing. However, such paradigms rely heavily on the underlying communication layer, inheriting the challenges originated from the high mobility of vehicles. Several works have been devised to cope with highly dynamic vehicular environments in support of effective resource management and allocation, which we discuss in the paper. Moreover, we propose a Fog paradigm solution to resource allocation using a hierarchical method in vehicular clouds. Our method is based on the Multiplicative Analytic Hierarchy Process (MAHP) proposed by Lootsma. MAHP is a branch of another method called Analytic Hierarchy Process proposed by Saaty. Therefore, we used MAHP in the decision-making of the resource allocation process using a Fog paradigm to select the best Fog to allocate certain services. We evaluated the proposed solution comparing to three other decision methods, GREEDY, RANDOM, and RELIABLE. The proposed Fog-oriented Hierarchical Resource Allocation Policy in Vehicular Clouds (FRACTAL) performed better than the other decision methods, fulfilling more services and consequently denying fewer services.

The growth of vehicles in cities brings significant socioeconomic problems and new challenges. With that growth, the amount of information generated by vehicles and their devices also increments and can improve the network using Vehicular Ad Hoc Networks (VANET). VANETs make the communication between vehicles and infrastructures possible to exchange information and share resources. To assist VANETs, another concept called Vehicular Cloud Computing (VCC) brings the Cloud paradigms to this scenario. In this paper, we propose a Bat Bio-inspired Mechanism for Resource Allocation in Vehicular Clouds, called NAUTILUS. The algorithm uses the metaheuristic to optimize the search process for defining pseudo-optimal decision-making of the allocation process in a Vehicular Cloud. We also consider a fog-based paradigm to assist the proposed mechanism in the allocation process. We allocate the following resources from the vehicles: storage, memory, runtime, and processing. The NAUTILUS was compared to two other algorithms that use traditional search techniques: a Greedy approach and an Analytic Hierarchy Process (AHP) approach. In the comparison process, we evaluate the number of blocked, attended, and denied services. Simulations results show that the NAUTILUS presented better efficiency than Greedy and AHP approaches in all three performance aspects: blocking fewer, attending more, and denying fewer services.

Urban computing has become a significant driver in supporting the delivery and sharing of services, being a strong ally to intelligent transportation. In an intelligent transportation scenario, smart vehicles present computing and communication capabilities that enable many autonomous vehicular safety and infotainment applications. Vehicular Fog computing appears as a new paradigm in enabling and facilitating efficient service and resource sharing in urban environments. Several vehicu-1ar resource management works have attempted to deal with the highly dynamic vehicular environment following diverse approaches, such as MDP, SMDP and policy-based greedy techniques. However, the high vehicular mobility causes several challenges compromising consistency and efficiency. RL-enabled adaptive vehicular Fogs can deal with the mobility for properly distributing load and resources over Fogs. Thus, we propose a mobility-based dwell time estimation method for accurately estimating vehicular resources in a Fog, leveraging the design of an adaptive and highly dynamic resource allocation model.

Recent advancements in smart and connected vehicles have led to increasing demand for vehicular communication and resources. It is thus essential to have a strategy that deals with resource management in vehicular Clouds where vehicles can form dynamic clouds for Edge computing. Such control is challenging due to high vehicle mobility and unstable connectivity. Previous works explored resource management problems in terms of vehicular mobility. However, stable connectivity is also crucial when talking about discovering and accessing resources from vehicles. Thus, we propose a resource management approach, COrRect, that makes use of connectivity and communication delays, along with the availability of heterogeneous resources from the vehicles to find the most suitable vehicle for resource/service provision. We have conducted a simulation experimental analysis to evaluate the performance of the proposed approach. Results have demonstrated that COrRect improved the selection of providers in terms of best connection conditions within highly-dynamic vehicular networks.

Vehicular Ad-hoc Networks (VANETs) have contributed significantly towards improving road traffic management and safety. VANETs, integrated with Vehicular Clouds, enable underutilized vehicular resources for efficient resource management, fulfilling service requests. This work introduces the concept of clustering resources from nearby vehicles to form Combined Resource Units (CRUs). These units contribute to fulfilling user service requirements while avoiding the difficult task of matching user requests with the available resources of an individual vehicle. CRU composition is helpful, especially for the heterogeneity of vehicles. The vehicle resources are clustered into CRUs based on three different sized pools, making the service matching process more time-efficient. Previous works have adopted stochastic models for resource clustering configurations. However, this paper adopts distinct search algorithms for CRU composition, which are computationally less complex. Results showed that light-weight search algorithms, such as SSA, achieved close to 80% of resource availability without over-assembling CRUs in higher density scenarios.

The growth in the number of traffic accidents has become a cause for concern in urban centers. As a result of the increase in population in large cities and the number of vehicles, the consequences of accidents and congestion can be even more significant, considering the impacts on the economy, environment and people's quality of life. Therefore, aiming to minimize these impacts, we present ALIVE, a distributed and Vanet-based solution that reduces congestion caused by different sources and, especially, from accident sources, contributing to efficient urban mobility in smart cities. The solution performs the detection of accidents and the dissemination of warning messages in multiple hops. Besides, the system can share the road traffic information with the nearby streets to improve traffic efficiency. We evaluated the proposed solution with PANDORA and NRR Traffic Management solutions. Simulation results indicate that the proposed solution reduces the average travel time, time lost, and the number of transmitted messages.

Road capacity infrastructure and temporary interruptions in trips constitute the main reasons behind the traffic jam phenomenon. City urbanization and growth further intensify these two reasons through the increase of work area and the demand for mobility. In such a scenario, several issues can emerge, such as higher mobility costs, more frequent traffic jams, more significant environmental damage, reduced quality of life, and more pollution. Therefore, this work presents a Proactive and Reactive Approach for Urban Traffic Management in Vehicular Networks, RIDER, to minimize traffic congestion. RIDER is a fully-distributed protocol that can assume proactive and reactive behaviors for sharing traffic condition information. Vehicles with traffic condition information can organize them-selves to improve traffic flow and reduce traffic congestion. In the proposed solution, vehicles monitor the road traffic condition and proactively share this information when needed, considering adaptive multi-hop communication. If vehicles do not have nearby road traffic information, they executed a reactive traffic information discovery. RIDER was evaluated and compared to previous works, regarding the number of transmitted messages, packet collisions, and traffic congestion metrics.

Vehicular Clouds heavily rely on the underlying Vehicular Networks to discover, announce, and exchange services and resources. The assembly and control of such clouds depend upon reliable and efficient data dissemination among vehicles and roadside units. Dissemination allows for critical information to be spread efficiently and widely through the VANETs. It is, therefore, imperative to create a dissemination algorithm that reduces the number of redundant messages. The biggest concern is to define an effective and efficient data dissemination algorithm that can be used in critical areas in traffic like intersections. Several issues must be considered in dense vehicular regions, such as broadcast storms and redundancy. Several approaches have already shown a reduction in the redundancy and overhead. Still, they may need to improve on the variance of a dynamically changing topology and exponential growth rate of messages sent. These approaches include a speed adaptive probabilistic dissemination, which is a lightweight approach. This work focuses on controlling the dissemination pace by applying timers and limiting forwarding messages on the speed adaptive broadcast algorithm to reduce overhead and redundancy while ensuring the broadcast is transmitted evenly. The timers are useful in high-density traffic and conclusively showed a broadcast overhead reduction.

Vehicular Cloud computing is new paradigm where vehicles collaboratively exchange data and resources to support services and problem-solving in urban environments. Characteristically, such Clouds undergo severe challenging conditions from the high mobility of vehicles, and by essence, they are rather dynamic and complex. Many works have explored the assembling and management of Vehicular Clouds with designs that heavily focus on mobility. However, a mobility-based strategy relies on the geographical position of vehicles and its feasibility has been questioned in some recent works. Therefore, we present a more relaxed Vehicular Cloud management scheme that relies on connectivity. This work models uncertainty and considers every possible chance a vehicle may be available through accessible communication means, such as V2X communications and the vehicle being in the range of RSUs for data transmissions. We utilize the MDP model to track the state of vehicles and when there are connected and available for transmission of the data.

The economic and environmental impacts caused by traffic congestion are increasing. Improvements in the cities road infrastructure for minimizing these impacts are pricey and do not happen immediately. Thus, in order to improve vehicular traffic flow in dense urban centers, we present REACT, a traffic management system to minimize vehicle congestion in Smart Cities. REACT is a traffic management system based on Vehicular communication, and it is divided into Request and Response phases. The Request phase allows vehicles to request traffic information from neighbor road segments. The Response supports vehicles to respond to the request with current road traffic information. The performance evaluation shows the ability of our solution to reduce traffic jams with a low communication overhead.

Intelligent Transport Systems rely extensively on the proper management of vehicular resources, as well as the underlying vehicular communication. Services and applications are made accessible through the communication of vehicles, which is expected to happen without any interruption; however, the high mobility of vehicles is detrimental towards their connectivity and communication stability. In this highly dynamic and heterogeneous vehicular scenario, we assume the existence of multiple communication interfaces and high movement speed of nodes. Thus, we propose the development of a flow mobility management policy. The policy is devised following a decentralized design, being flexible and capable of providing transparent flow management to users while maintaining continuous service access. We conducted experimental simulations for evaluating the proposed architecture, as well as making performance comparisons with three related flow management works. The results showed that the proposed policy reduces the delay of information exchange to approximately 50 ms, decreases handover time (0.5s), and maximizes the input of information around 95%.

In order to improve simulation performance and to integrate simulation resources among geographically distributed locations, the concept of distributed simulation is proposed. Several types of distributed simulation standards, such as DIS and HLA, are established to formalize simulations and achieve reusability and interoperability of simulation components. To implement these distributed simulation standards and to manage the underlying system of distributed simulation applications, we employ grid computing and cloud computing technologies. These tackle the details of operation, configuration, and maintenance of simulation platforms in which simulation applications are deployed. However, for modelers who may not be familiar with the management of distributed systems, it is challenging to make a simulation-run-ready environment among different types of computing resources and network environments. In this article, a new multi-layered cloud-based scheme is proposed for enabling modeling and simulation based on different distributed simulation standards. This scheme is designed to ease the management of underlying resources and to achieve rapid elasticity that can provide unlimited computing capability to end users; it considers energy consumption, security, multi-user availability, scalability, and deployment issues. A mechanism for handling diverse network environments is described; by adopting it, idle public resources can be easily configured as additional computing capabilities for the local resource pool. A fast deployment model is built to relieve the migration and installation process of this platform. An energy-saving strategy is utilized to reduce the consumption of computing resources. Security components are implemented to protect sensitive information and block malicious attacks in the cloud. In the experiments, the proposed scheme is compared with its corresponding grid computing platform; the cloud computing platform achieves similar performance, but incorporates many advantages that the Cloud can provide.

Big Data platforms are convoluted distributed systems which commonly comprise skill- and labour-intensive solution development to treat inherent Big Data application challenges. Several tools have been proposed to help developers and engineers to overcome the involved complexities in coordinating the execution of plenty processes/threads on multiple machines. However, no work so far has been able to combine both an accurate representation of Big Data jobs and realistic modeling of the behaviour of Big Data platforms at scale, including networking elements and data and job placement. In this paper, we propose BigDataNetSim, the first simulator which models accurately all the main components of the data movements in Big Data platforms (e.g., HDFS, YARN/MapReduce, network topologies, switching/routing protocols) in a large scale system. BigDataNetSim can serve as a valuable tool for engineering Big Data solutions, which includes set-up of systems, prototyping of jobs, and improvement of components/algorithms for Big Data platforms. We also demonstrate that BigDataNetSim can simulate a real Hadoop cluster with a high degree of accuracy in terms of data and job placements, being able to scale up to very large systems.

To support increasingly sophisticated sensors and resource-hungry applications with the current-used Lithium-based batteries and to augment mobile computing power further, the concept of the Cloudlet-based offloading is proposed which enables to migrate part of application computing tasks from battery-limited low-capacity mobile elements to local Cloudlets. However, due to the limited processing capability and the lack of fine-grain resource management schemes on the Cloudlet, the Cloudlet resources can be quickly overloaded especially in the large-scale multi-user offloading scenarios. As a result, a considerable number of offloading requests are forwarded to the remote Cloud, which may significantly increase the communication overhead for the energy-sensitive mobile offloading tasks. In this paper, we develop and formulate a novel task-centric energy-aware Cloudlet-based Mobile Cloud model to address this issue. We concern the offloading performance, scalability, security, and availability problems, aiming at increasing the Cloudlet processing throughput, reducing the energy cost on the remote Cloud, and improving offloading execution efficiency and energy-efficiency on the mobile devices. A Cloudlet task-based offloading mechanism is proposed to achieve fine-grain energy-aware offloading resource preparation and scheduling on the Cloudlet. A Cloud task-centric scheduling algorithm is presented for the green collaborative offloading processing between Cloudlet and remote Cloud. The experiment results demonstrate that the energy-aware offloading model can efficiently enhance offloading performance for mobile devices, and the offloading scheduling schemes for the Cloudlet and remote Cloud outperform the traditional protocol class.

Modeling and simulation have shown essential for forecasting load and resource availability in large-scale complex scenarios. The growth of urban environments, as well as the use of ICT in enabling applications and services, has encouraged several works on the modeling of transportation. High mobility of vehicles in such a context consists of a significant challenge in modeling traffic. Several microscopic and macroscopic models have been designed aiming to represent the movement of vehicles accurately in road segments, involving different levels of complexity, precision, and realism. Out of these models, Free-flow models have shown useful due to being light and reasonably accurate for estimating load in short-time predictions. A recent free-flow traffic flow modeled using queues assumed constant vehicle speed along the road segment; this assumption may lead to a lack of realism and accuracy. Therefore, we propose a free-flow model based on this previous work where the road segment is split into several intervals, representing the oscillations of the speed of vehicles. The proposed model has shown correctness comparable to the previous free-flow model, considering that it has included speed varying behavior of vehicles.

Mobile devices nowadays can connect to the network very conveniently using cellular data network or WiFi. However, latency is still a challenge caused by the stability and the availability of the network, mainly in the context of mobile environments. In this paper, we propose an architecture based on a Cloudlet model using CAching and pre-FEtching scheme (CAFE scheme) to improve data access efficiency. The prefetching scheme on the Cloud enables the retrieval of specific data in advance based on specific information of users. On the Cloudlet, a caching technique selectively stores data passing through the Cloudlet. The classification of data into specific and general takes both individual access behavior and common trends into consideration. Compared to an original model, the experiment results show that our architecture can really decrease latency and improve data access efficiency when users request data from a Cloudlet and the Cloud.

Mobile devices are now capable of handling many daily computing tasks that used to be accomplished by desktops or servers. However, these improvements also introduce resource-hungry mobile applications that require richer resource-hungry computing features and more complex functions. Mobile Cloud Computing (MCC) addresses these limitations considering the nature of mobility; this innovative strategy provides external computing and storage capability so that tasks can be offloaded. The concept Cloudlet model is proposed to perform as the local resource pool that receives outsourced tasks. However, Cloudlets are restricted by the computing power and storage capacity, limiting the scale of offloading devices. In this paper, a Cloudlet-based task offloading model is proposed. By utilizing caching technologies and N-to-N resource scheduling, Cloudlets can support a larger number of mobile devices compared to previous models. Based on the experimental results, the proposed scheduling model cloud achieve better overall efficiency on energy consumption and task execution.

HLA-based simulation systems are prone to load imbalances due to lack management of shared resources in distributed environments. Such imbalances lead these simulations to exhibit performance loss in terms of execution time. As a result, many dynamic load balancing systems have been introduced to manage distributed load. These systems use specific methods, depending on load or application characteristics, to perform the required balancing. Load prediction is a technique that has been used extensively to enhance load redistribution heuristics towards preventing load imbalances. In this paper, several efficient Time Series model variants are presented and used to enhance prediction precision for large-scale distributed simulation-based systems. These variants are proposed to extend and correct the issues originating from the implementation of Holt's model for time series in the predictive module of a dynamic load balancing system for HLA-based distributed simulations. A set of migration decision-making techniques is also proposed to enable a prediction-based load balancing system to be independent of any prediction model, promoting a more modular construction.

Cloud computing environments mainly focus on the delivery of resources, platforms, and applications as services to users over the Internet. Cloud promises users access to as many resources as they need, making use of an elastic provisioning of resources. The cloud technology has gained popularity in recent years as the new paradigm in the IT industry. The number of users of Cloud services has been increasing steadily, so the need for efficient task scheduling is crucial for maintaining performance. In this particular case, a scheduler is responsible for assigning tasks to virtual machines efficiently, it is expected to adapt to changes along with defined demand. In this paper, we suggest an elastic scheduler that is able to alter its focus based on the current requirements demanded by the cloud service provider and the user of those services. The Elasticity Based Scheduling Heuristic (EBSH) suggested is measured against the bio-inspired optimization algorithms such as Ant Colony Optimization (ACO) and Honey Bee Optimization (HBO). Also, a networking algorithm is used in this study, namely Random Biased Sampling (RBS). The presented EBSH shows superior performance because of its ability to adapt to changes.

Mobile Cloud Computing is a concept infrastructure wherein Cloud Computing resources are utilized to offload tasks from mobile elements. The combination of Cloud Computing and Mobile Computing increases the complexity of modeling and performance evaluation in terms of task scheduling policies, energy consumption model, the mobility of mobile devices over a layered architecture of both local Clouds and remote Cloud data centers. In this paper, a distributed HLA-based Cloud toolkit is proposed, enabling the modeling and simulation of Mobile Cloud Computing environment. The proposed toolkit simulates the behaviors of the Mobile system and the Cloud infrastructure, within which different resource scheduling policies can be evaluated in a repeatable manner. In addition, an HLA-based simulation scheduling scheme is proposed, trying to improve simulation execution efficiency by automatically parallelizing and distributing simulation components. A Cloud-based simulation resource management paradigm is also implemented, handling the configuration and maintenance issues regarding underlying system resources and simulation data. Based on the experiments, the proposed toolkit can achieve better simulation execution efficiency, with consideration of both Cloud and Mobile behaviors, compared to current Cloud Computing environment simulators.

A Vehicular Cloud generally focuses on several aspects, which include providing a set of computational services at low cost to vehicle drivers; minimizing traffic congestion, accidents, travel time, and environmental pollution; and ensuring the use of low energy and real-time services of software, platforms, and infrastructure with QoS to drivers. The largest challenge in vehicular mobile Clouds consists of creating a mechanism for management and search resources that does not depend on roadside infrastructure, which consequently requires the spontaneous and dynamic creation of a Cloud through the resources shared by vehicles. Therefore, the system must enable collaboration and co-operation between vehicles so that they may establish connections to provide resources. To address this challenge, we propose a peer-to-peer protocol to assist in the discovery and management of resources in a vehicular mobile Cloud without depending on the support of a roadside infrastructure so that vehicles are expected to organize themselves and establish collaborations to manage and share their resources. Simulation results show that the proposed approach introduces a short search time of approximately 0.5 (ms) to seek resources in one hop and approximately 0.9 (ms) to seek more of a hop in resources. Furthermore, the proposed protocol enables a high availability of resources, about 87%.

The choice of a suitable path for packet transmission represents a fundamental issue to any routing protocol. The principle of choosing the shortest path is also no longer a good option for route selection since many other aspects may influence communication performance. In that sense, the link quality of the path is more important than the length of the path in a wireless network because of the unstable conditions of the channel. Expected Transmission Count (ETX) is a widely-used routing metric, which servers into the selection of the path with fewer transmissions for a successful packet delivery. However, other factors should be also considered for route selection, such as the re-transmission delay. In this paper, a comprehensive analysis on the effect of the link layer delay to the transmission throughput is presented and discussed. Given the analysis outcome, this work proposes a routing metric based on the link layer delay for IEEE 802.11. This metric allows to determine the delay of each hop along the path to evaluate the quality of the path as whole. Experimental simulation results reveal that the proposed routing metric achieves better results when compared to two other known approaches.

Due to the nature of communication, mobility and portability in Mobile Computing, the handling of limited computing, storage and network capabilities become increasingly important especially when more features and richer functionality are required today. Cloud Computing, as an elastic computing utility provisioning framework, is shown to be a promising approach, addressing the concerns in Mobile Computing. Many achievements have been made by researchers regarding how to offload computational tasks from mobile systems to the Cloud. However, the proposed offloading methodologies are mainly from the perspectives of mobile application level, focusing on static estimation, dynamic partitioning, cloning, transmission overhead evaluation and migration. Issues related to multi-core Cloud systems are not fully considered, such as overall energy consumption of Cloud systems, information security, usability and availability. In this paper, a platform-based system model is designed from the view of the Cloud platform, trying to enable these Cloud benefits in addition to offloading, and to provide better execution efficiency and overall energy reduction by utilizing the proposed platform level scheduling. Based on the experiments, the proposed platform scheduling can achieve greater energy reduction with little computing overhead on the management node, compared to application-level scheduling methods.

Vehicular network applications may be benefited by the use of simultaneous network interfaces to maximize through-put and reducing latency. In order to take advantage of all radio interfaces of the vehicle and to provide a good quality of service for vehicular applications, we have developed an architecture that performs the management of the flow mobility based on some classes of application for vehicle network. Our goal is to minimize the time of handover between the rings of flows in order to meet the minimum requirements of vehicular applications, as well as to maximize the throughput. Simulations have been conducted to analyze the performance of the proposed architecture by comparing it to other previously devised architectures. As a result, the proposed architecture presented a low delivery time of messages, packets with lower loss and lower delay.

The accelerated growth of applications and services in intelligent transportation systems (ITS) are driven by interests from the public and private sectors. The intent to utilize the onboard resources, along with the advanced methods of managing the available computing capabilities in the conventional cloud, has led to the high popularity of Vehicular Clouds. Likewise in Vehicular Networks, vehicles provide the building blocks for forming these particular clouds, which can enable a large number of applications and services that can benefit the whole transportation system, as well as the drivers, passengers, and pedestrians. However, due to its high mobility, Vehicular Clouds show several inherent challenges, which increase complexity and restrict the design of solutions. Determining the number of vehicles and their time of availability in a given region through a model works as a critical stepping stone for enabling vehicular clouds, as well as any other system involving vehicles moving over the traffic network. Therefore, by implementing proper traffic models, we present a comprehensive stochastic analysis about the distribution of the number of vehicles inside a road segment in this paper. According to real parameters, we show that certain classes of applications are feasible even for highly mobile scenarios.

Summary Simulations have been used extensively for evaluating scenarios, which are very difficult, costly or impractical to implement in real systems. Testing in a synthetic, realistic environment provides a means to determine the viability of solutions. Simulations have proved to be very useful in the verification of algorithms and protocols, offering tools for testing them in different situations. The simulation of vehicular area networks pose additional challenges as realistic mobility models are crucial and must be incorporated in scenario elements while applications and communication protocols are tested. Several simulators and simulation frameworks have been designed that aim to synthetically reproduce communication and mobility of vehicles as realistically as possible. The majority of such simulators merge pre‐existing networking and mobility simulators, which add issues regarding compatibility and realism. Such simulators present limited run‐time 3D visualization tools, essential for providing immersive environments. Therefore, in this paper, we propose real‐time simulation and 3D visualization for vehicular networks of realistic scenarios. This proposed simulation system generates output in real time, making use of 3D‐modelled real‐world maps and effectively generating visualization as elements are updated in the simulation. Experiments have been conducted with simulation and visualization components to evaluate delays and performance of the proposed simulator. Copyright © 2016 John Wiley & Sons, Ltd.

Summary An examination of the multidimensional range query in existing peer‐to‐peer (P2P) overlay networks indicates that multidimensional range queries are sensitive to underlying topologies; this is because partitioning and mapping of multidimensional data space are two interconnected parts of a process that must be carried out cooperatively. The first section focuses on how to preserve data localities, whereas the second section concerns how to accommodate and maintain data localities at the P2P overlay layer. There are many studies that have been conducted on the first section since 1966, and those works that are well accepted are mostly based on recursive decomposition, which forms a tree structure in nature. However, less effort has been made to provide comparable support from the P2P overlay layer. In our previous work, we proposed the Hierarchically Distributed Tree (HD Tree) in order to better support multidimensional range queries in the P2P overlay network. This paper further explores error‐resilient routing and load balancing strategies that can be employed in the HD Tree. We also provide a complete set of experimental results for all routing operations: Join and Leave of nodes, range queries at different levels of selectivity, and the dynamic load balancing scheme. Comparisons are made by conducting simulations under both the ideal and the error‐prone routing environment and within various ary HD Trees. The experimental results show that load balancing in the HD Tree can be adjusted dynamically and globally, and it is actually a trade‐off between distributing the basic load and the involvement of nodes in range querying. The experimental results also indicate that a maximum of 10 percent of routing nodes’ failures do not have significant effects on the performance of range queries. However, a lower ary HD Tree appears to have better routing performance, whereas a higher ary HD Tree achieves a higher fault‐tolerant capacity. Nevertheless, the performance of range queries in a higher ary HD Tree can be further optimized if all possible routing options can be fully explored in the error‐prone routing environment. Copyright © 2013 John Wiley & Sons, Ltd.

Load management is an essential and important factor for distributed simulations running on shared resources due to load imbalances that can caused considerable performance loss. This feature is essential for High Level Architecture (HLA)-based simulations since the HLA framework does not present the ability to manage resources or help detect load imbalances that could directly cause decrease of performance. A migration-aware dynamic balancing system has been designed for HLA simulations to offer an efficient load-balancing scheme that works in large-scale environments. This system presents some limitations on estimating costs and benefits, so we propose an enhancement to this existing load balancing system, which improves the accuracy of generating federate migrations. The proposed scheme aims to precisely estimate the migration delay and gain by analyzing the load on shared resources, preventing the issuing of migrations costly towards simulation execution time. Upon a performance analysis, the proposed decision-making analysis scheme has shown an improvement on decreasing the number of migrations and consequently decreasing execution time.

The HLA framework is widely used to formalize simulations and achieve reusability and interoperability of simulation components. In order to manage the underlying system of HLA-based simulations, Grid Computing and Cloud Computing are employed to tackle the details of operation, configuration, and maintenance of simulation platforms that simulation applications run on. However, to make a simulation-run-ready environment among different types of computing resources and network environments is challenging, especially for modelers who may not be familiar with the management of distributed systems. In this article, we propose a new cloud-based scheme for HLA based simulations, aiming to ease the management of underlying resources, particularly for those located on geographically distributed locations, and to achieve rapid elasticity that can provide adequate computing capability to end users. An approach for handling diverse network environments is given, by adopting it, idle public resources can be easily configured as additional computing resources for the local cloud infrastructure. In the experiments, compared with its corresponding Grid Computing platform, this Cloud Computing platform achieves a similar performance but with many advantages that Cloud can provide, such as energy consumption, security, and multi-user availability.

TCP suffers degradation in wireless networks, which is caused by the improper, static definitions on the lower layers. In order to improve the TCP performance in wireless networks, the strategy of the lower layer should be reconsidered. In this paper, the TCP transmission is flattened in order to combine the TCP segment transmission and TCP Acknowledgement (ACK) transmission into one transmission. The performance of TCP is also analyzed, in order to find out the corresponding parameters affecting the TCP throughput. Based on the analysis, contention window size shows as one of the parameters that greatly affects the TCP performance. A discrete-time Markov decision process is adopted in order to solve the TCP throughput maximization. Based on the analysis, a TCP-distributed algorithm is proposed. Several simulations are conducted to verify the improvements of TCP-distributed by comparing both TCP Reno and TCP Vegas. Simulation results show that TCP-distributed can perform better than the two TCP variants, and it also limits the delay in an acceptable range.

Several research studies have introduced an efficient and intelligent path recommendation protocols for vehicular networks. Communications among traveling vehicles and located roadside units (RSUs) have been utilized to investigate the traffic distribution over the road network. This helps construct the optimal path towards each targeted destination located on the road network. However, none of the previous proposed protocols in this field have specifically considered potential faults among the nodes of the vehicular networks or potential link failures. In this paper, we present a fault tolerant distributed-based path recommendation (TD-PR) protocol. Our protocol detects and tolerates faults occur among nodes and/or communication links. We present TD-PR protocol in this paper and report on its performance evaluation. Our simulation experiments show that TD-PR improves the success rate significantly over our previously proposed path recommendation protocol (ICOD). The success ratio is improved in roadside failure and link failure scenarios. In general TD-PR has better performance in terms of decreasing the traveling time and traveling distance compared to ICOD in these scenarios.

Cloud computing has been receiving a growing attention due to its features on the provisioning of computing resources for distributed processing. A lot of interest lays on the enabled benefits of flexible and elastic management of cheap and reliable computing sources though virtualization. The cheap characteristic brought from virtualization helps data centers to host more than one operating system on any machine. However, the time-sharing nature of virtualization and the existence of more software layers lead to larger delays in execution and network communications. The bundling of network messages directed to the same destination can be used as means to reduce the number of network transfers. In this paper, the effect of different message bundling aspects closely related to cloud applications is investigated. Analysis shows that message bundling effectively enhances communication efficiency; however, it is directly influenced by parameters such as number of messages, length of the bundling cycle, and number processing units.! Thus, the highest performance gain is achieved by flexibly adjusting according bundling parameters towards data communication in large-scale cloud environments.

HLA-based distributed simulations tend to suffer from load imbalances and degradation in performance as a result of running on distributed environment. High-Level Architecture (HLA) is a general purpose framework that eases the implementation of distributed simulations on top of dedicated resources without worrying about the computing infrastructure. Due to the high cost of hardware and other factors, some companies have ditched the concept of dedicated resources and shifted towards shared ones which revealed some HLA weaknesses, out of which, dynamic reaction to load imbalances and managing federates on the shared resources. Therefore, different efforts have proposed numerous dynamic load balancing systems to offer a balancing feature to running distributed simulations. In order to perform the load balancing task, these proposed systems gather and make use of a number of simulation and load metrics. Load prediction is a metric that is computed to provide load projections and prevent any prospective load imbalances by migrating federates from an overloaded shared resource to an underloaded shared resource. This work touches the federate migration decision-making process, which is the last step of the balancing task. The proposed federate migration decision-making methods are to overcome the dependency on predefined thresholds in previous work and offer dynamic decisions to migrate federates.

Evaluation and validation of algorithms and protocols in vehicular area networks is challenging and requires the support of simulators in most cases due to the restrictions on cost and scalability. Consequently, there is a need to identify or build a simulator that best fits into the characteristics of VANets. Such simulators need to reproduce the communication of networks together with the mobility of vehicles in a given simulated area. Many simulators and simulation frameworks have been developed, most of them combining pre-existing mobility and networking simulators in one solution. However, these simulators present limited features on 3D visualization. In this paper, we propose a real-time, realistic 3D visualization for VANet simulations, which makes use of 3D-modeled real-world maps; the proposed system effectively generates the intended visualization. Experiments have been conducted to evaluate the performance on the synchronization between simulation and visualization components through an analysis of overhead and delays.

Due to the dependency of HLA-Based distributed simulations on the resources of distributed environments, simulations can face load imbalances and can suffer from low performance in terms of execution time. High-Level Architecture (HLA) is a framework that simplifies the implementation of distributed simulations, and, it has been built with dedicated resources in mind. As technology is nowadays shifting towards shared resources, the following two weaknesses have become apparent in HLA: managing federates and reacting towards load imbalances on shared resources. Moreover, a number of dynamic load management systems have been designed in order to provide a solution to enable a balanced simulation environment on shared resources. These systems use some specific techniques depending on certain simulation or load aspects, to perform the balancing task. Load prediction is one such technique that improves load redistribution heuristics by preventing load imbalances. In this work, we present a number of enhancements for a prediction technique and compare their efficiency. The proposed enhancements solve observed problems with Holt's implementations on dynamic load balancing systems for HLA-Based distributed simulations and provide better forecasting. As a result, these enhancements provide better forecasting for the load of the shared resources.

As the scale of distributed simulations, such as HLA-base simulations, grow due to the complexity in modelling real systems, balancing systems become essential for enabling the execution of these simulations or preventing simulation performance loss. This loss is greatly produced by non-dedication or heterogeneity of resources in large-scale environments or by simulations that might generate dynamic load oscillations. In light of avoiding load imbalances in such conditions, a distributed balancing system has been developed. However, in order to enable load balancing systems to effectively and properly react to load imbalances, they need to be properly deployed and configured in environments where are occasionally unknown due to the lack of previous information about them or to the dynamic changes that might occur as time passes. As a result, an autonomous configuration scheme is proposed to increment the distributed balancing system and allow it to flexibly adapt to the environment where it is deployed. Experiments have been performed in defined scenarios in order to assess the benefit in incorporating such dynamic configuration solution on the balancing system.

SUMMARY Because of the availability of shared resources, substantial efforts have been applied to the development of large‐scale distributed simulations, and performance has become an essential aspect that can be impaired by heterogeneity and availability of resources, dynamic, unpredictable load imbalances, and communication delays. In order to manage and keep such distributed simulations consistent, the high level architecture (HLA) standard has been designed; however, it does not provide any solution that directly solves simulation performance issues. Many balancing approaches have been proposed in order to offer a suboptimal balancing solution, but they are limited to certain simulation aspects, are specific to determined applications, or are unaware of the HLA‐based simulation characteristics. In light of considering both computational and communication aspects for HLA‐based simulations, a centralized hierarchical balancing scheme was proposed. This scheme presents several drawbacks that make it susceptible to bottlenecks, overheads, global synchronization, and single point of failure. Therefore, a scheme based on a distributed algorithm to re‐arrange the computational and communication load is proposed. Experiments have been performed to evaluate the effectiveness of the distributed scheme when compared with the scheme based on a centralized redistribution algorithm. The results showed that the distributed balancing technique could provide similar performance gain or even improve it for some specific cases. Copyright © 2011 John Wiley & Sons, Ltd.

Load balancing is a vital mechanism for improving the performance of distributed simulations or even for enabling their execution. A balancing technique has been designed in order to provide a balancing scheme for HLA-based simulations on non-dedicated resources. However, this technique lacks efficiency by producing large amounts of unnecessary federate migrations, so a self-adaptive mechanism has been introduced in the technique in order to correct its balancing responsiveness. As a drawback, the self-adaptation technique assumes that only the frequency of federate migrations represents the balancing efficiency. This leads the scheme to present static parameters regardless of the conditions of the environment, which in turn can restrict the balancing response to imbalances. Thus, awareness to migration delays is inserted into the self-adaptive balancing scheme in order that more precise and more realistic analysis of balancing efficiency can be enabled. Experiments have been conducted to show the performance gain of the proposed scheme when compared to the distributed and self-adaptive load balancing systems.

Due to the dependency on the resources, HLA-based simulations can experience load imbalances and consequently loose execution performance. Such imbalances are originated from external background load, inappropriate deployment of simulation entities, heterogeneity of resources, and dynamic load variations. The High Level Architecture (HLA) was developed aiming to facilitate the creation and control of distributed simulations through a design framework and management services, but such an architecture does not offer solutions for solving load imbalance issues. In order to provide mechanisms for preventing performance loss caused by load imbalances in distributed simulations, numerous balancing approaches have been developed. The majority of these mechanisms present application-specific solutions or limited awareness of environment characteristics. To cope with this problem, a distributed dynamic balancing scheme has been designed, but its redistribution algorithm, as other developed balancing schemes, is limited to just correct load distribution issues and does not react properly in presence of abrupt load changes due to be based on recent load status. Therefore, a predictive balancing scheme is proposed to provide a method to decrease the number of precipitated migration moves and to detect and prevent load imbalances based on load variation tendencies. In order to observe and evaluate the proposed scheme, experiments have been performed to compare performance gain and efficiency with the distributed balancing scheme.

Large-scale HLA-based simulations are susceptible to performance issues caused by load imbalances. High Level Architecture (HLA) was designed to organize distributed simulations, but it does not provide any mechanism to prevent imbalances. Also, several load balancing schemes have been proposed to properly re-arrange simulation load, but they do not present all the features needed for large-scale environments. A hierarchical, distributed balancing scheme has been designed to support the execution of such simulations, however, due to the balancing inter-relations, heterogeneity of resources, and cyclic load changes, the scheme reacts unnecessarily. Thus, a self-adaptive balancing scheme is proposed in order to dynamically re-configure the redistribution scheme and avoid the ones that are needless. Experimental results showed that the adaption technique improved the balancing efficiency since less migrations were required to achieve similar or better performance.

In large-scale distributed simulations, communication aspects are highly significant due to their direct influence on performance. The High Level Architecture (HLA) provides services for managing such simulations and reducing their communication overhead. However, HLA does not present any solution for the communication latencies caused by the network distances among simulation elements. Several dynamic balancing schemes have been proposed attempting to provide a general best solution for the performance issues caused by computation and communication imbalances. Amongst these schemes, some just perform a limited redistribution of communication load. Based on a proximity analysis of federate interactions, a distributed dynamic scheme for balancing the communication load of HLA-based simulations is devised. The design of this distributed scheme aims at improving fault tolerance, decreasing communication and computation overload, and avoiding bottlenecks in the system. The distributed balancing system, organized in the hierarchical structure, monitors simulations, redistributes load, and migrates federates. Experiments have been realized to compare the proposed distributed scheme with a centralized scheme and to prove its effectiveness for large-scale HLA-based simulations.

HLA-based simulations, as any distributed computing application, can undergo critical performance issues due to load imbalances on large-scale, heterogeneous, non-dedicated distributed systems. Such imbalances are produced by HLA simulation entities that can dynamically change their computation and communication load during their execution time, so an initial static load deployment is incapable of providing simulations complete and even distributed resources usage. Moreover, because the computing resources are non-dedicated, unknown external applications can generate load for any computing resource, increasing the imbalances' unpredictability. Thus, in order to re-allocate resources for an HLA simulation during its execution time, an hierarchical dynamic load balancing system is introduced. The system manages a simulation's workload by monitoring the distributed load through the MDS Grids' service; by identifying load imbalances according to a load sharing policy; by re-allocating resources according to defined policies; and by migrating federates through the GRAM Grids' service, a migration proxy, and peer-to-peer state transfer. By keeping the load evenly partitioned on the distributed system, such a devised system successfully improved the simulations' performance. The experimental results and comparative analyses between balanced and non-balanced simulations proved the efficiency of the proposed dynamic load balancing system.

Federate migration is a fundamental mechanism for large-scale distributed simulations. It provides the means for simulation load-balancing and thus improves the simulation's overall performance. Given its importance for simulations, several federate migration approaches have been proposed in the literature. Some approaches freeze the entire simulation, others use third-party mechanisms to transport data, and others make use of unnecessary communication and computing. Thus, in order to minimize the time spent on federate migration, we introduce a simulation agent that manages the migration steps, as well as the migrating federate's communication with the other simulation entities. The use of the agent simplifies the message management transparently and avoids redundant usage of network and computing resources. We demonstrate through simulation experiments that our approach decreases federate migration latency, improving the performance of HLA simulations that run over large-scale environments.