Article

PORA: Predictive Offloading and Resource Allocation in Dynamic Fog Computing Systems

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

In multi-tiered fog computing systems, to accelerate the processing of computation-intensive tasks for real-time IoT applications, resource-limited IoT devices can offload part of their workloads to nearby fog nodes, whereafter such workloads may be offloaded to upper-tier fog nodes with greater computation capacities. Such hierarchical offloading, though promising to shorten processing latencies, may also induce excessive power consumptions and latencies for wireless transmissions. With the temporal variation of various system dynamics, such a trade-off makes it rather challenging to conduct effective and online offloading decision making. Meanwhile, the fundamental benefits of predictive offloading to fog computing systems still remain unexplored. In this paper, we focus on the problem of dynamic offloading and resource allocation with traffic prediction in multi-tiered fog computing systems. By formulating the problem as a stochastic network optimization problem, we aim to minimize the time-average power consumptions with stability guarantee for all queues in the system. We exploit unique problem structures and propose PORA, an efficient and distributed predictive offloading and resource allocation scheme for multi-tiered fog computing systems. Our theoretical analysis and simulation results show that PORA incurs near-optimal power consumptions with queue stability guarantee. Furthermore, PORA requires only mild-value of predictive information to achieve a notable latency reduction, even with prediction errors.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... Thus, in distributed systems with nodes that may span over large geographical areas (such as in fog computing), measuring the network proximity and organizing the nodes accordingly is necessary to ensure the efficiency of the hierarchical structure. However, most current approaches for creating fog computing systems rely on hierarchical structures but do not take actual proximity measurements, and do not discuss how the nodes of such a structure are discovered and organized in a way that fulfills latency-and bandwidth-related goals (e.g., [13,77,118]). This might not pose a concern for statically configured small-scale systems, but may become a problem when the task of creating the system structure needs to be automated, e.g., for large-scale systems. ...
... For instance, Kiani et al. [118] propose a hierarchical structure with geographically distributed compute nodes, and investigate related performance benefits based on simulations that consider either negligible or significant delay between compute nodes. Gao et al. [77] present a hierarchical fog computing system with dynamic resource allocation based on traffic prediction, and show latency reductions using extensive simulations. Adhikari et al. [13] propose a hierarchical fog computing system for executing applications based on priorities. ...
... Flat measurements Scalability Presented structure structure based on analysis system Adhikari et al. [13] Simulated Gao et al. [77] Simulated Kiani et al. [118] Simulated Skarlat et al. [207] Coordinates Implemented Saurez et al. [199] Coordinates Implemented Mortazavi et al. [160] Emulated Nguyen et al. [168] Not defined Casadei and Viroli [47] Simulated Rabay'a et al. [184] Simulated Yu et al. [241] Not defined Song et al. [210] Implemented Tato et al. [216] Hops, latency Simulated Jiang and Tsang [96] Simulated ...
Thesis
Full-text available
Fog computing is a novel computing paradigm which enables the execution of applications on compute nodes which reside both in the cloud and at the edge of the network. Various performance benefits, such as low communication latency and high network bandwidth, have turned this paradigm into a well-accepted extension of cloud computing. So far, many fog computing systems have been proposed, consisting of distributed compute nodes which are often organized hierarchically in layers. Such systems commonly rely on the assumption that the nodes of adjacent layers reside close to each other, thereby achieving low latency computations. However, this assumption may not hold in fog computing systems that span over large geographical areas, due to the wide distribution of the nodes. In addition, most proposed fog computing systems route the data on a path which starts at the data source, and goes through various edge and cloud nodes. Each node on this path may accept the data if there are available resources to process this data locally. Otherwise, the data is forwarded to the next node on path. Notably, when the data is forwarded (rather than accepted), the communication latency increases by the delay to reach the next node. This thesis aims at tackling these problems by proposing distributed algorithms whereby the compute nodes measure the network proximity to each other, and self-organize accordingly. These algorithms are implemented on geographically distributed compute nodes, considering image processing and smart city use cases, and are thoroughly evaluated showing significant latency- and bandwidth-related performance benefits. Furthermore, we analyze the communication latency of sending data to distributed edge and cloud compute nodes, and we propose two novel routing approaches: i) A context-aware routing mechanism which maintains a history of previous transmissions, and uses this history to find nearby nodes with available resources. ii) edgeRouting, which leverages the high bandwidth between nodes of cloud providers in order to select network paths with low communication latency. Both of these mechanisms are evaluated under real-world settings, and are shown to be able to lower the communication latency of fog computing systems significantly, compared to alternative methods.
... Therefore, we assume that the incoming URLLC traffic to base station bs i∈{1..I} during time slot t denoted with Ai,u(t) follows a Pareto distribution with an arrival rate λ u i and parameters ai(t) and xi(t) [5,8]. In addition, an amount Ai,m(t) ≤ A max of mMTC traffic arrives to base station bsi; such that A max is a positive constant and E[Ai,m(t)] = λ m i [18]. Each base station bsi is equipped with a learning module customized to IMSTN that can predict the future mMTC traffic by leveraging the spatio-temporal correlation between the neighboring base stations and between the subsequent satellites. ...
... In each small base station bsi, there are different types of queues as depicted in Fig. 4 [18,20]: c) Terrestrial queues Q Ter i,κ , where κ ∈ {m, u}, which buffers the traffic of type κ that will be offloaded to the terrestrial backhaul. When κ = m, the traffic type is mMTC and when κ = u, the traffic type is URLLC. ...
... .., (Wi − 1)} and the arrival queue A −1 i,m (t) evolve according to the following dynamics [18,20]: ...
Preprint
Full-text available
The integration between the satellite network and the terrestrial network will play a key role in the upcoming sixth-generation (6G) of mobile cellular networks thanks to the wide coverage and bandwidth offered by satellite networks. To leverage this integration, we propose a proactive traffic offloading scheme in an integrated multi-satellite terrestrial network (IMSTN) that considers the future networks' heterogeneity and predicts their variability. Our proposed offloading scheme hinges on traffic prediction to answer the stringent requirements of data-rate, latency and reliability imposed by heterogeneous and coexisting services and traffic namely enhanced mobile broadband (eMBB), massive machine-type communications (mMTC) and ultra-reliable low latency communication (URLLC). However, the fulfilment of these requirements during offloading in dynamic IMSTN comes at the expense of significant energy consumption and introduces inherently supplementary latency. Therefore, our offloading scheme aims to balance the fundamental trade-offs first between energy consumption and the achievable data-rate and second between energy consumption and latency while meeting the respective needs of the present traffic. Our findings prove the importance of the cooperation between the multi-satellite network and the terrestrial network conditioned by traffic prediction to enhance the performance of IMTSN in terms of latency and energy consumption.
... The CloudSim platform allows comparison of processor productivity, resource utilization, task completion, and other factors. In [19], As a result of the study, the PREDICTIVE Offloading and Resource Allocation (PORA) method was proposed, which is an efficient and distributed resource allocation scheme for multilayer fog computing systems, and the study examined the effects of dynamic discharge and resource allocation on traffic forecasting in these systems. The simulation results showed that the proposed method was an efficient online method that minimized energy consumption and ensured queue stability while applying the predicted discharge. ...
... where α implies the maximum amount of CPU power consumption and is calculated based on formula (19). ...
... Resource utilization can be minimized and allocation process optimized Propose Real-time TRAM algorithm based on fog nodes to maximize the allocation of efficient cloud energy resources [19] Content courtesy of Springer Nature, terms of use apply. Rights reserved. ...
Article
Full-text available
Fog computing can be an effective way to improve quality of services and solve network problems, with the demand for real-time, latency-sensitive applications increasing as well as limitations such as network bandwidth and Internet of Things users' resources. Due to the fact that different tasks in a network can create overhead that can reduce the quality of service, dynamic voltage, and frequency scaling along with a ranking function and a high number of physical and virtual machines were used in this research. A profit function phase is used to analyze network tasks in order to improve QoS by sending them to physical machines and sending them via the network to physical machines. The simulation results demonstrate that this method is the most effective in allocating radio and computational resources to IoT devices in fog computing. A comparison is presented in the results section between the proposed method and the SPA, Markov-Fog, and TRAM methods. Criteria for evaluating performance include the response time for heterogeneous environments, energy consumption against tasks and users, memory processing, energy consumption for physical and virtual machines, and network profitability.
... Some of these include autonomic computing [93], accelerated gradient with Lagrangian dual theory [94], Q-learning [95,96], deferred acceptance [97], online optimization [98], Interior-point [98,99], Hungarian [100,101], game theory [102][103][104], matching game [105], sub-gradient [106,107], and outer approximation [108]. Many researchers have also come up with their own algorithms to address various optimization problems in fog computing services [109][110][111][112][113][114]. ...
... Gao et al. [113] Own Integer Programming Sun et al. [124] Own Linear Programming Elgendy et al. [242] Branch and Bound Integer Programming Fan et al. [169] Lagrangian New algorithm proposed by the authors Any exiting algorithm could be used as stated by the authors Cat Swarm optimization tency-and cost-related metrics were considered in [234] and [172] which include time to execute a task in the fog, the transmission time between an end device and a fog node, and the cost of processing the task in a fog node. A joint optimization of bandwidth, and network and resource utilization was implemented in [88]. ...
... As shown in Fig. 11, latency-related metrics also play a significant role in this category with 17 papers utilizing this objective metrics, followed by energy-based metrics with 12 papers. Among these papers, some included only one metric in their objective functions such as [98], [113], [245], [246], [169], [187], [211], and [221]. For instance, Lee et al. [98] addressed the problems of fog node placement and task allocation aiming at minimizing the computation time and communication latency. ...
... By far, the most recent offspring of the physical separation of functional units is fog. It's a computational layer that delivers processing, networking, and storage services near the perception layer, which houses the sensors and actuator [3,4]. Latency-critical applications in mobile networks have recently received a lot of attention and have seen progress. ...
... Z.LI et al. [1,3,5] Due to advantages in both communication and computing, the fog radio access network (F-RAN) is a potential architecture for future networks that can effectively boost network performance. In a device-to-device (D2D) enabled F-RAN, the data caching mechanism is examined from a social aspect. ...
... Research [3] The issue of online content placement with undetermined content popularity was investigated. It suggested LACP, a multiplay Thompson sampling-based learning-aided content placement strategy. ...
Article
Full-text available
The rapid advancement in the Internet of things applications generates a considerable amount of data and requires additional computing power. These are serious challenges that directly impact the performance, latency, and network breakdown of cloud computing. Fog Computing can be depended on as an excellent solution to overcome some related problems in cloud computing. Fog computing supports cloud computing to become nearer to the Internet of Things. The fog's main task is to access the data generated by the IoT devices near the edge. The data storage and data processing are performed locally at the fog nodes instead of achieving that at cloud servers. Fog computing presents high-quality services and fast response time. Therefore, Fog computing can be the best solution for the Internet of things to present a practical and secure service for various clients. Fog computing enables sufficient management for the services and resources by keeping the devices closer to the network edge. In this paper, we review various computing paradigms, features of fog computing, an in-depth reference architecture of fog with its various levels, a Review Article Abdulqadir et al.; AJRCOS, 10(4): 1-18, 2021; Article no.AJRCOS.71449 2 detailed analysis of fog with different applications, various fog system algorithms, and also systematically examines the challenges in Fog Computing which act as a middle layer between IoT sensors or devices and data centers of the cloud.
... Computational offloading enables workload/computational tasks to be shared between IoT devices, fog nodes, and cloud servers [11][12][13][14]. When computational offloading occurs between fog nodes, this is called "fog cooperation" [15], in which overloaded fog nodes send part of their workload to other underloaded fog nodes to meet their QoS requirements [16,17]. Resource management can involve multiple factors, saving energy consumption in the fog environment is one of these factors, and is considered in this work. ...
... Regarding online deployment of computational offloading, few studies have addressed this, such as [15][16][17][24][25][26]. Yousefpour et al. [16] suggested a delay-minimization approach to reduce overall service delay. ...
... Their numerical results indicated the effectiveness of their proposed model in terms of minimizing overall latency in comparison with different algorithms. In a fog-cloud computing system, Gao et al. [17] investigated the issue of dynamic computational offloading and resource allocation. In order to reduce energy consumption and delay while having a stable queueing status, the authors formulated the problem as a stochastic network optimization problem. ...
Article
Full-text available
Fog computing is a potential solution to overcome the shortcomings of cloud-based processing of IoT tasks. These drawbacks can include high latency, location awareness, and security—attributed to the distance between IoT devices and cloud-hosted servers. Although fog computing has evolved as a solution to address these challenges, it is known for having limited resources that need to be effectively utilized, or its advantages could be lost. Computational offloading and resource management are critical to be able to benefit from fog computing systems. We introduce a dynamic, online, offloading scheme that involves the execution of delay-sensitive tasks. This paper proposes an architecture of a fog node able to adjust its offloading threshold dynamically (i.e., the criteria by which a fog node decides whether tasks should be offloaded rather than executed locally) using two algorithms: dynamic task scheduling (DTS) and dynamic energy control (DEC). These algorithms seek to minimize overall delay, maximize throughput, and minimize energy consumption at the fog layer. Compared to other benchmarks, our approach could reduce latency by up to 95%, improve throughput by 71%, and reduce energy consumption by up to 67% in fog nodes.
... From the last decade, resource contribution in distributing computing has discovered great attention among researchers which incorporate cloud computing, [19][20][21][22] edge computing, 23,24 and fog computing. [25][26][27] The problem of resource allocation is referred to the efficient management of idle storage, computing, and processing resources. The problem of resource allocation is approached using several perspectives which involve optimization computing, game theory, intelligent decision making using machine learning. ...
... The problem of resource allocation is approached using several perspectives which involve optimization computing, game theory, intelligent decision making using machine learning. Gao et al 26 formulated an optimization problem by minimizing the overall time and power usage in a multi-tier fog environment. The designed problem is solved using Lyapunov optimization. ...
Article
Full-text available
In the IoT‐cloud environment, the growing amount of spawn data may limit performance in terms of communication latency, network traffic, processing power, and energy usage. The introduction of fog computing extends the cloud services nearer to the edge of the network. Since these lightweight fog servers are not able to fulfill the demand of every user node and process each offloaded task due to the limited computation resources. Accordingly, an efficient resource management scheme is required to proficiently handle fog resources. The profit‐driven nature of both the fog service providers and user nodes increases the possibility of malicious activity while resource trading for their advantages or to privilege a bunch of devices. In this article, we designed a trusted and fair incentive mechanism that encourages buyers and sellers to trade by leveraging the benefits of blockchain and smart contracts. Especially, a combinatorial double auction employed market model is proposed which satisfies different economical properties such as individual rationality, budget balance, and truthfulness. Blockchain‐driven decentralized fog environments prevent the tampering of trade‐related information by the malicious nodes. Simulation results indicate that the proposed combinatorial double auction significantly improves the network utilization by improved winner determination and pricing model.
... The preferences of fog and IoT devices are defined considering fog nodes attempt to minimize their operation cost (i.e., energy consumption) and minimize traffic cost, while IoT devices are interesting in minimizing the execution time. Offloading and resource allocation problems in a multi-tier fog network are address in [18]. The solution attempts to minimize long-term time-average expectation of power consumption in fog tiers. ...
... We considered a matrix of visited to check all possible combinations of coalitions during the merge and split operations. The merge process (lines [11][12][13][14][15][16][17][18][19][20][21][22] is started by selecting a random local coalition F L i that did not check its merging Calculate preference list of offloaded-tasks and sort lists in decreasing order 5: while there is any unmatched offloaded-task do 6: for t=1:OT do 7: if t is unmatched and it's preference list is not empty then 8: t propose to it's top-ranked choice s in the preference list 9: remove s from preference list of t 10: ...
Article
Full-text available
Fog computing provides a distributed computing paradigm that executes interactive and distributed applications, such as the Internet of Things (IoT) applications. Large-scale scientific applications, often in the form of workflow ensembles, have a distributed and interactive nature that demands a dispersed execution environment like fog computing. However, handling a large-scale application in heterogeneous environment of fog computing requires harmonizing heterologous resources over the continuum from the IoT to the cloud. This paper investigates offloading and task allocation problems for orchestrating the resources in a fog computing environment where the IoT application is considered in the form of workflow ensembles. We called Offload-Location a mechanism which has been designed to find offloading coalition structure alongside a matching algorithm for allocating the offloaded tasks to fog/cloud resources. The introduced solution attempts to minimize the execution time and minimize the price paid to servers for executing the tasks provided that Quality of Service (QoS) requirements of the ensemble’s deadline and budget are retaining. These objectives lead to maximizing the number of completed workflows of the ensemble as an ultimate goal. The appropriate performance of this mechanism is studied under different workflow applications and circumstances.
... This approach scaled the problem size to exploit the computation power of the controller, and the cost-based switching of the allocation streamlined the system's accuracy. A predictive offloading and resource allocation (PORA) system has been created to address the service offloading issue derived from the traffic prediction of fog servers, in an attempt to minimize the energy consumption of the system [29]. A task selection-task allocation (TS-TA) method was established to effectively allocate resources using a genetic algorithm. ...
... The fog-cloud environment is evolving rapidly, and machine learning algorithms learn from the evolving environment and adapt accordingly [15]. Predicting the workload using machine learning provides an essential advantage in planning and scheduling the resources [16]. Content courtesy of Springer Nature, terms of use apply. ...
Article
Full-text available
In recent, there has been a huge increase in the number of context-aware and latency-sensitive IoT applications. Executing these applications on traditional cloud servers is infeasible due to strict latency requirements. Emerging edge technologies such as fog/edge computing, cloudlets, edge clouds etc. have been proposed recently to fulfill latency requirements of these applications. In these edge technologies, computing infrastructure is available near to the end-user devices. Scheduling the IoT applications on heterogeneous and distributed fog/edge nodes is an important and complex research problem and has been extensively studied in these domains by applying various traditional approaches. In recent times, there has been tremendous growth in machine learning research and its applications in many domains. This work makes a detailed study of the machine learning techniques and their applicability for the fog/edge computing by reviewing various machine learning applications and opportunities in fog/edge computing. As most of the existing works in machine learning confines to resource allocation, therefore, a fog application classifier and scheduling model is proposed that would help researchers to understand how a resource allocation and scheduling problem can be solved. We present detailed algorithms that schedule the IoT applications on multiple fog layers based on applications’ QoS. The performance of the proposed work has been validated through simulation study. Various challenges in the realization of machine learning in fog/edge computing along with the future possibilities are also presented.
... Therefore, when the workload on a particular fog node increases, it can offload part of its workload to the next layer's nodes, and this is where optimization is needed. The authors in [106] mentioned a novel approach for predictive offloading and stochastic network optimization in resource allocation by using a queuing model for optimization. ...
Article
Full-text available
With the rapid growth in the data and processing over the cloud, the accessibility of those data has become easier. On the other side, it poses many technical and security challenges to the users of those provisions. Fog computing makes these technical issues manageable up to some extent. Fog computing is one of the promising solutions for handling the big data produced by the IoT, which is often security-critical and time sensitive. Massive IoT data analytics by fog computing structure is emerging and requires extensive research for more proficient knowledge and smart decisions. Though advancement in Big Data Analytics is taking place, it does not consider Fog Data Analytics. But, there are many challenges, including heterogeneity, security, accessibility, resource sharing, network communication overhead, real-time data processing of complex data, etc. This paper explores various research challenges and their solution using the next-generation Fog Data Analytics and IoT networks. We also performed an experimental analysis based on fog computing and cloud architecture. The result shows that fog computing outperforms cloud in terms of network utilization and latency. Finally, the paper is concluded with future trends.
... In addition, it has a scalability issue as it needs to know the states of all fog nodes to make decisions. The authors of [18] proposed a multi-tiered fog computing system that dynamically offloads tasks and allocate resources based on traffic predictions. They did not differentiate the task length from its processing time, which generally are two different parameters for each task. ...
Conference Paper
The Internet of Things (IoT) has grown at a rapid pace in recent years. It requires a large amount of data and massive computational resources, thus the concept of Fog Computing (FC) has emerged. FC attempts to overcome network latency by bringing computational resources closer to IoT devices. One important part of FC is an offloading mechanism t o make proper decisions for better utilizing of FC node(s), especially for real-time (low latency and high throughput) applications. Generally, offloading policies a re categorized a s centralized and distributed. However, by growing numbers of IoT devices which leads to expansion of FC layer beyond the initial configurations, centralized scheduling solutions for time-sensitive tasks suffers from two major challenges: first, i ncreasing c omplexity, and second, non-fault tolerating. In order to address these issues, scalable decentralized/distributed approaches have been developed to schedule tasks through an autonomous collaboration between a small number of nodes (neighbors). Without a thorough picture of the network or nodes' state, it is difficult to design algorithms that make optimum decisions. This paper presents a scalable algorithm for offloading time-sensitive t asks through a semi-network aware distributed scheduling mechanism. Based on the evaluation results obtained for acceptance rate, response time, and network resource usage, the proposed method outperforms the state-of-the-art on average.
... What to offload determines the offloaded task and its form, whether to perform fine-grain offloading such as [10,11] to transmit tasks partially or course grain offloading [12,13] to transmit the entire task. Second question, when to offload answer that tasks should be offloaded to minimize the latency [14], or to solve the problem of energy consumption [15,16] or some other QoS factors [17,18] should be considered while transmitting the tasks. Where to offload selects the best suitable computation device to execute the offloaded tasks which can be using fog computing, Mobile Edge Computing (MEC) [16,19], or cloudlets [20]. ...
Article
Full-text available
With the surge of intelligent devices, applications of Internet of Things (IoT) are growing at a rapid pace. As a result, a massive amount of raw data is generated, which must be processed and stored. IoT devices standalone are not enough to handle large amount of data. Hence, to improve the performance, users started to push some jobs to far-situated cloud data centers, which would lead to more complications such as high bandwidth usage, service latency, and energy consumption. Fog computing emerges as a key enabling technology that brings cloud services closer to the end-user. However, owing to the unpredictability of tasks and Quality of Service (QoS) requirements of users, efficient task scheduling and resource allocation mechanisms are needed to balance the demand. To handle the problem efficiently, we have designed the task offloading problem as Markov Decision Process (MDP) by considering various user QoS factors including end-to-end latency, energy consumption, task deadline, and priority. Three different model-free off-policy Deep Reinforcement Learning (DRL) based solutions are outlined to maximize the reward in terms of resource utilization. Finally, extensive experimentation is conducted to validate and compare the efficiency and effectiveness of proposed mechanisms. Results show that with the proposed method, on average 96.23% of tasks can satisfy the deadline with an 8.25% increase.
... A new resource scheduling algorithm based on optimized fuzzy clustering is introduced to separate the resources, by which the scale of the resource gets reduced (Nguyen, et al., 2019;Mohammad, et al., 2018;Qayyum, et al., 2015;Mukherjee, 2019). Nowadays, renowned meta-heuristics algorithms such as Particle Swarm Optimization (PSO), Dragonfly Algorithm (DA), Monarch Butterfly Optimization (MBO), etc (Beno, et al., 2014;Zhang and Li, 2018;Gu, et al., 2018;Gao, et al., 2020) are exploited for resource allocation strategy in Fog computing. The existing algorithms have some drawbacks like the PSO algorithm has a low convergence rate Furthermore, the DA method is easy to implement, and it needs few parameters for tuning, and also it suits for different applications. ...
Article
Full-text available
Fog computing is a decentralized computer system where data, processing, storage, as well as applications are located anywhere between the cloud and data source. Fog computing takes the cloud closer to users, decreasing the latency and allows the deployment of new delay-sensitive appliances. An important feature of a fog-cloud network is the process of decision-making on assigning the resources to execute the tasks of application. This paper aims to propose a resource allocation strategy for fog computing that determines the effective process under the consideration of the objectives that includes the constraints like credibility score, concurrency, price affordability and task time computation. Moreover, the credibility score is determined based on the execution efficiency, Service response rate, access reliability and Reboot rate. Thereby, the optimal allocation of resources is handled by a new Hybrid Monarch-Dragon Algorithm (HM-DA) that hybrids the concept of Dragonfly Algorithm (DA) and Monarch Butterfly Optimization (MBO) algorithm.
... A typical fog computing system consists of a set of fog nodes at different locations. These nodes are deployed on the periphery of the network and have flexible resource configurations such as storage, computing, and network bandwidth [6]. If the cloud is far from the edge node, the link connected to the cloud may not be reliable, which also causes a lot of delay. ...
Article
Full-text available
Delay-sensitive tasks account for an increasing proportion of all tasks on the Internet of Things (IoT). How to solve such problems has become a hot research topic. Delay-sensitive tasks scenarios include intelligent vehicles, unmanned aerial vehicles, industrial IoT, intelligent transportation, etc. More and more scenarios have delay requirements for tasks and simply reducing the delay of tasks is not enough. However, speeding up the processing speed of a task means increasing energy consumption, so we try to find a way to complete tasks on time with the lowest energy consumption. Hence, we propose a heuristic particle swarm optimization (PSO) algorithm based on a Lyapunov framework (LPSO). Since task duration and queue stability are guaranteed, a balance is achieved between the computational energy consumption of the IoT nodes, the transmission energy consumption and the fog node computing energy consumption, so that tasks can be completed with minimum energy consumption. Compared with the original PSO algorithm and the greedy algorithm, the performance of our LPSO algorithm is significantly improved.
... For the resource allocation problem, although the multilayer structure of edge computing can reduce the processing delay, it can also increase the power consumption and delay of transmission through the server in the network. The paper [16] mainly studies the dynamic task arrival and resource allocation problem with traffic prediction in edge computing system and describes it as a stochastic network optimization problem. The simulation results show that the power consumption is improved when the processing delay is stable. ...
Article
Full-text available
The Internet of Things (IoT) is rapidly growing and provides the foundation for the development of smart cities, smart home, and health care. With more and more devices connecting to the Internet, huge amounts of data are produced, creating a great challenge for data processing. Traditional cloud computing has the problems of long delays. Edge computing is an extension of cloud computing, processing data at the edge of the network can reduce the long processing delay of cloud computing. Due to the limited computing resources of edge servers, resource management of edge servers has become a critical research problem. However, the structural characteristics of the subtask chain between each pair of sensors and actuators are not considered to address the task scheduling problem in most existing research. To reduce processing latency and energy consumption of the edge-cloud system, we propose a multilayer edge computing system. The application deployed in the system is based on directed digraph. To fully use the edge servers, we proposed an application module placement strategy using Simulated Annealing module Placement (SAP) algorithm. The modules in an application are bounded to each sensor. The SAP algorithm is designed to find a module placement scheme for each sensor and to generate a module chain including the mapping of the module and servers for each sensor. Thus, the edge servers can transmit the tuples in the network with the module chain. To evaluate the efficacy of our algorithm, we simulate the strategy in iFogSim. Results show the scheme is able to achieve significant reductions in latency and energy consumption.
... Zeng et al. discussed wireless energy harvesting in edge computing for mobile devices [16]. In some works, cloud and edge computing were combined to improve the system performance [19]. ...
Article
Full-text available
Flexible resource scheduling and network forecast are crucial functions to enhance mobile vehicular network performances. However, BaseStations (BSs) and their computing unit which undertake the functions cannot meet the delay requirement because of limited computation capability. Offloading the time-sensitive functions to User Equipment (UE) is believed to be an effective method to tackle this challenge. The disadvantage of the method is offloading occupies communication resources, which deteriorate the system capability. To better coordinate offloading and communication, a multi-connectivity enhanced joint scheduling scheme for distributed computation offloading and communication resources allocation in vehicular networks is proposed in this article. Computation tasks are divided into many slices and distributed to UEs to aggregate the computation capability. A communication-incentive mechanism is provided for involving UEs to compensate the loss of UEs, while multi-connectivity is adopted to enhance the system throughput. We also defined offloading failure ratio as a conclusive condition for offloading size by analyzing the movement of UEs. By a two-step optimization, the co-scheduling of offloading size and throughput is solved. The system-level simulation results show that the offloading size and throughput of the proposed scheme are larger than comparisons when the time constraint is tight.
... In [26], [27], the authors proposed an active load balancing scheme based on the traffic prediction, the balance between the computing loading and the long-term system energy consumption minimization is considered mainly. In [28], the author proposed a non-cooperative game task offloading strategy. ...
Article
Full-text available
Vehicular edge computing (VEC) has emerged as a promising paradigm to ensure the real-time task processing caused by the emerging 5G or high level intelligent assisted driving applications. The computing tasks can be processed via the edge services deployed at the roadside units (RSUs) or moving vehicles. However, the high dynamic topology of the vehicular communication system and the time-varying available computing resources in RSUs make a challenge of the efficient task offloading of vehicles. In this paper, we consider an efficient task offloading scheme for VEC networks based on trajectory prediction, we focus on the serving handover between the adjacent RSUs. The moving vehicles can cooperate with RSUs or the surrounding vehicles for task processing. To reduce the latency of task transmission between vehicles, we present a cooperative vehicle selection method based on trajectory prediction. Then, we propose an efficient task offloading scheme based on deep reinforcement learning (DRL), while the dynamically available computing and communication resources are considered jointly. The simulation results show that the proposed task offloading scheme has great advantages in improving the utility of vehicles.
... Task offloading problem is resolved by designing different efficient and optimization schemes. In [23], an offloading strategy based on Lyapunov optimization is presented. A multitier fog architecture is presented where all the offloaded nodes are arranged in increasing order of their computation power. ...
Article
Full-text available
With the elevation of terminal devices, network traffic has also grown with rapid speed. To relieve cloud computing constraints on timely delivery, energy consumption, and congestion, fog computing is introduced to provide proximate and spot-on services to network devices. In fog environment, resource-deprived user nodes can offload their multi-task to network periphery situated fog nodes. However, due to the decentralized and untrusted behavior of fog nodes; trade, and pricing related sensitive information can be tampered by an unauthorized entity for their benefits. In this paper, we design a joint resource allocation and pricing scheme using blockchain employed smart contract. Blockchain is used to overcome untrusted behavior by preventing the malicious node from tampering with trade information. However, a smart contract employed fair and secure payment among user nodes and fog nodes. Particularly, we propose descending combinatorial auction-based dynamic pricing schemes, in which any user node can request bundle of resources and fog nodes compete with each other to serve their request. Experimental results show that the proposed pricing scheme helps fog nodes to earn significant profit by providing their resource efficiently.
... IoT resources are arranged in a hierarchical order: cloud, fog nodes and edge devices (maximum to minimum resources). Different task offloading options that can be considered include [115]:-• Device-to-device offloading. ...
Article
Full-text available
Fog computing is a promising technology that can provide storage and computational services to future 6G networks. To support the massive Internet of Things (IoT) applications in 6G, fog computing will play a vital role. IoT devices and fog nodes have energy limitations and hence, energy-efficient techniques are needed for storage and computation services. We present an overview of massive IoT and 6G enabling technologies. We discuss different energy-related challenges that arise while using fog computing in 6G enabled massive IoT. We categorize different energy-efficient fog computing solutions for IoT and describe the recent work done in these categories. Finally, we discuss future opportunities and open challenges in designing energy-efficient techniques for fog computing in the future 6G massive IoT network.
... Additionally, network providers can also leverage the predictions for determining caching [30] demand for MECs. Advanced task offloading models can be integrated with the proposed model to optimize the computational resources further [31]. Moreover, the optimization of cooperation among the MEC servers [32] could be considered when the network providers deploy the servers. ...
Preprint
A smart city improves operational efficiency and comfort of living by harnessing techniques such as the Internet of Things (IoT) to collect and process data for decision making. To better support smart cities, data collected by IoT should be stored and processed appropriately. However, IoT devices are often task-specialized and resource-constrained, and thus, they heavily rely on online resources in terms of computing and storage to accomplish various tasks. Moreover, these cloud-based solutions often centralize the resources and are far away from the end IoTs and cannot respond to users in time due to network congestion when massive numbers of tasks offload through the core network. Therefore, by decentralizing resources spatially close to IoT devices, mobile edge computing (MEC) can reduce latency and improve service quality for a smart city, where service requests can be fulfilled in proximity. As the service demands exhibit spatial-temporal features, deploying MEC servers at optimal locations and allocating MEC resources play an essential role in efficiently meeting service requirements in a smart city. In this regard, it is essential to learn the distribution of resource demands in time and space. In this work, we first propose a spatio-temporal Bayesian hierarchical learning approach to learn and predict the distribution of MEC resource demand over space and time to facilitate MEC deployment and resource management. Second, the proposed model is trained and tested on real-world data, and the results demonstrate that the proposed method can achieve very high accuracy. Third, we demonstrate an application of the proposed method by simulating task offloading. Finally, the simulated results show that resources allocated based upon our models' predictions are exploited more efficiently than the resources are equally divided into all servers in unobserved areas.
... Additionally, network providers can also leverage the predictions for determining caching [30] demand for MECs. Advanced task offloading models can be integrated with the proposed model to optimize the computational resources further [31]. Moreover, the optimization of cooperation among the MEC servers [32] could be considered when the network providers deploy the servers. ...
Article
Full-text available
A smart city improves operational efficiency and comfort of living by harnessing techniques such as the Internet of Things (IoT) to collect and process data for decision making. To better support smart cities, data collected by IoT should be stored and processed appropriately. However, IoT devices are often task-specialized and resource-constrained, and thus, they heavily rely on online resources in terms of computing and storage to accomplish various tasks. Moreover, these cloud-based solutions often centralize the resources and are far away from the end IoTs and cannot respond to users in time due to network congestion when massive numbers of tasks offload through the core network. Therefore, by decentralizing resources spatially close to IoT devices, mobile edge computing (MEC) can reduce latency and improve service quality for a smart city, where service requests can be fulfilled in proximity. As the service demands exhibit spatial-temporal features, deploying MEC servers at optimal locations and allocating MEC resources play an essential role in efficiently meeting service requirements in a smart city. In this regard, it is essential to learn the distribution of resource demands in time and space. In this work, we first propose a spatio-temporal Bayesian hierarchical learning approach to learn and predict the distribution of MEC resource demand over space and time to facilitate MEC deployment and resource management. Second, the proposed model is trained and tested on real-world data, and the results demonstrate that the proposed method can achieve very high accuracy. Third, we demonstrate an application of the proposed method by simulating task offloading. Finally, the simulated results show that resources allocated based upon our models' predictions are exploited more efficiently than the resources are equally divided into all servers in unobserved areas.
Article
With the arrival of the Internet of Things (IoT) era, multi-tier computing has attracted significant attention. The multi-tier computing can organize many computing devices and provide sufficient computing resources to support various IoT applications. However, due to the complex architecture and the dynamic system status of the multi-tier computing network, task offloading for multi-tier computing is still challenging. This paper proposes a novel task offloading method named OSTTD, to deal with the Offloading of Splittable Tasks with Topological Dependence in multi-tier computing networks. OSTTD formulates the task offloading as a sequential decision-making problem and learns the task offloading policy by Deep Reinforcement Learning (DRL). Compared with existing task offloading approaches, OSTTD is the first method in which the topological dependence among sub-tasks of the splittable task is fully considered. OSTTD makes offloading decisions intelligently based on the dynamic system status and can be applied to various multi-tier network topology structures. To verify the effectiveness of OSTTD, we extend and build a latency-aware multi-tier computing simulation platform. Extensive simulations show that OSTTD can significantly reduce the task processing time, thus, improving the overall task processing efficiency in multi-tier computing networks.
Article
The integration of cloud and IoT edge devices is of significance in reducing latency of IoT stream data processing by moving services closer to the edge-end. In this connection, a key issue is to determine when and where services should be deployed. Common service deployment strategies used to be static based on the rules defined at the design time. However, dynamically changing IoT environments bring about unexpected situations such as out-of-range stream fluctuation, where the static service deployment solutions are not efficient. In this paper, we propose a dynamic service deployment mechanism based on the prediction of upcoming stream data. To effectively predict upcoming workloads, we combine the online machine learning methods with an online optimization algorithm for service deployment. A simulation-based evaluation demonstrates that, compared with those state-of-the art approaches, the approach proposed in this paper has a lower latency of stream processing.
Article
Since the beginning of the century, the internet of things (IoT) and cloud computing have dominated industry and academic research by raising the standard of living. In order to meet the demanding expectations of end users, both technologies recently encountered significant challenges. During the pandemic, the entire world experiences unexpected developments that force millions of people to stay at home and rely on remote services to perform both their personal and professional tasks. The use of fog computing is expanding significantly in response to the growing demand for network services. Although the growth of fog nodes around edge devices may also introduce malicious activities that can impair regular service operations. Moreover, it becomes more difficult to select a fog node in a distributed fog environment that will offer facilities that are reliable and secure. As a result, we present a resource allocation mechanism in this study where a resource-constrained user node chooses a trusted fog device that is one hop distant to handle various application service requests. We also provide an overload detection technique for the fog network, which is based on the service deadline and fog node service arrival rate. A fog-to-fog offloading method is proposed in the event of overload to select the best suitable node for preserving service quality. However, the suggested approach significantly outperforms and is more effective than the current baseline offloading methodologies. Results show that with the proposed method, average latency, and energy consumption are reduced by 65.15% and 67.94%, respectively, whereas a total of 91.07% of tasks are executed successfully.
Article
Full-text available
In recent years, the emergence of the internet of things (IoT) has accelerated the quality of day-to-day tasks. With the rapid development of IoT devices, the cloud computing paradigm has become an attractive solution by facilitating on-demand services. However, the remote location of third-party cloud reduces the overall user experience while increasing the latency involved. Fog computing has been introduced as a promising solution by improving the overall service quality. Fog computing comprises distributed and heterogeneous fog nodes with limited resource capabilities. Therefore, managing fog resources while satisfying users' service quality is a challenging task. This study attempts to conduct a systematic review by examining high-quality research papers published between 2018 and April 2022. This paper aims to address current trends, challenges, and theoretical gaps with a wide range of open issues to guide the researchers and practitioners interested in carrying out the further research in the current domain.
Article
In edge computing, the app vendors hire resources from edge servers and allocate them to app users to overcome the challenge of the limited computing capacities of their IoT devices. An app vendor intends to provide app services to the maximum number of users with the least number of edge servers in order to make efficient use of edge resources while reducing overall system costs. However, when an edge server has to serve more app users than its capacity, the Quality of Service (QoS) deteriorates. Thus, establishing a trade-off between cost and QoS is a critical challenge in the process of allocating edge computing resources to users. It is referred to as the App User Allocation (AUA) problem. To solve the AUA problem, we propose a distributed game-theoretic approach that finds a Pure Nash Equilibrium (PNE) as the optimal stable solution. We first model the AUA problem as a constrained optimization problem and then introduce a User Allocation Game (UAGame) to solve it. This UAGame employs a distributed Edge Server Allocation (ESA) algorithm to reach PNE. The time complexity of the ESA algorithm is reduced by the edge server clustering. It has also been shown that the UAGame is a potential game, and therefore the ESA algorithm is guaranteed to converge at PNE. The performance of the ESA algorithm has also been studied theoretically and validated numerically.
Article
Full-text available
Resource allocation in fog computing is a rigorous and challenging task and the allocation of appropriate resources to tasks generated by IoT users depends upon the QoS requirements of applications used by IoT users. Due to heterogeneity, mobility, uncertainty and limited availability of resources, the challenge of efficient resource allocation in fog computing cannot be addressed with traditional resource allocation strategies. Researchers are still facing problem in selecting an efficient resource allocation algorithm for wide variety of applications. This research study represents a systematic literature analysis of resource allocation in the fog computing. The current status of resource allocation in fog computing is distributed in several categories such as auction-based techniques, heuristics techniques and metaheuristic techniques etc. Methodological analysis of resource allocation techniques based on meta-heuristic approaches has been presented in this research paper. This research work will assist the researchers to find the important parameters of resource allocation algorithms and will also help in selecting appropriate resource allocation algorithm for tasks generated by IoT users.
Article
The exponential growth in Internet of Things (IoT) devices and the limitations of cloud computing in terms of latency and quality of service for time-sensitive applications have led to the unfolding of the efficient middleware technology called fog. Fog computing circumvents the limitations of the cloud by creating a seamless continuum between the things/IoT/end-user devices and the cloud and reducing the effective distance. However, fog computing faces challenges for offloading tasks for their remote computation at some level. Hence, the optimality of task offloading is the primary research area in fog computing. Several contemporary papers exist on this important subject. The research gap in reviewing all these task offloading algorithms has motivated us for their presentation in the form of a detailed survey in this paper. There exist some survey papers which deal with the task offloading. However, none of them has covered the basics of optimization techniques and their solution approaches. The primary objective of this paper is to provide the readers with a complete overview of the journey from a task offloading idea to its mathematical problem formulation and finally to its solution with all details of optimization techniques. We begin by introducing fog computing, and task offloading process followed by several task offloading factors governing decision-making process and their surveys. A section is fully dedicated to the survey of offloading objectives with examples. We also present several optimization approaches used in task offloading. Finally, the last section dedicates to the challenges and future direction in fog computing. The outcomes of the survey will benefit readers in learning the optimization used in task offloading, and it will also provide them a systematic design of offloading scheme with specific objectives.
Article
The fifth generation (5G) communication network has been developed rapidly in the past few years, which provides substantial bandwidth capacities and higher quality of service (QoS). 5G technology commercialization includes mobile edge computing and communication technologies. The increasing deployment of mobile edge computing empowered 5G infrastructures facilitates further intelligent data processing in smart city scenarios. The future innovative city applications require the seamless integration of mobile edge computing and communication. This article gives a comprehensive vision of the next generation communication networks integrated with distributed computing in both academia and industry. We propose the network for AI (NET4AI) and edge-mesh computing architectures, enabling cloud-native capabilities to be expanded to the communication network to realize integrated communication and computing services. Notably, a flexible KubeEdge Wireless (KEW) platform is inspired and discussed for dynamic services, such as service migration, shedding light for academia and industry. copy 2002-2012 IEEE. © 2022 Institute of Electrical and Electronics Engineers Inc.. All rights reserved.
Article
The integration between the satellite network and the terrestrial network will play a key role in the upcoming sixth-generation (6G) of mobile cellular networks thanks to the wide coverage and bandwidth offered by satellite networks. To leverage this integration, we propose a proactive traffic offloading scheme in an integrated multi-satellite terrestrial network (IMSTN) that considers the future networks’ heterogeneity and predicts their variability. Our proposed offloading scheme hinges on traffic prediction to answer the stringent requirements of data-rate, latency and reliability imposed by heterogeneous and coexisting services and traffic namely enhanced mobile broadband (eMBB), massive machine-type communications (mMTC) and ultra-reliable low latency communication (URLLC). However, the fulfilment of these requirements during offloading in dynamic IMSTN comes at the expense of significant energy consumption and introduces inherently supplementary latency. Therefore, our offloading scheme aims to balance the fundamental trade-offs first between energy consumption and the achievable data-rate and second between energy consumption and latency while meeting the respective needs of the present traffic. Our findings prove the importance of the cooperation between the multi-satellite network and the terrestrial network conditioned by traffic prediction to enhance the performance of IMTSN in terms of latency and energy consumption.
Article
Full-text available
Network slicing and mobile edge computing are key technologies in 5G/6G networks. Stable allocation of slicing resources is a challenge while communication and computation offloading coexist. To meet the service requirements of communication and computation offloading, a radio access network model which consists of enhanced mobile broadband (eMBB) slicing users, ultra‐reliable low latency communication (URLLC) slicing users, base stations integrated with MEC servers, and fog nodes with computing power are proposed. By employing the Lyapunov optimisation technique, a slicing resource allocation algorithm is proposed to maintain the queue stability and minimise the energy consumption of the system. For eMBB and URLLC slices, long time interval and short time interval resource allocation algorithms are proposed respectively. Simulation results show that the proposed algorithm can guarantee the long‐term stability of the network system while ensuring the service requirements.
Article
Owing to the limited computing capability of edge cloudlets, use of cloud platforms to provide computation‐intensive services as an assistant should be studied. Offloading, among the key techniques in the cloud‐edge environment, is plagued by the serious problem of privacy disclosure. In addition, the edge does not have access control of terminals, and the controller working the edge that makes offloading decisions, must not be allowed access to any information regarding the contents of the terminal. Thus, this study introduced a system architecture wherein the data was separated from the control plane to minimize systemic risk and optimize the offloading process towards makespan and cost reduction. Traditional methods of privacy preservation are dependent on encryption of total or content awareness. Edge privacy preservation renders maintaining edge delay advantages a challenge because traditional architectures result in edges being engaged with data encryption and decryption. Moreover, in content applications such as video compression and intelligent collaborative applications, users are unaware of content considered private data, and the relative privacy sensitivity and security priority of tasks. The proposed solution avoids side‐channel attacks by eliminating the individual characteristics via risk‐free constraints under a certain amount of entropy at the security level. Further, a smart optimum approach based on MOEA/D was employed to strike the balance between privacy preservation and the cost/latency of activity. This reduces the burden on edge servers, enhances security based on a better offloading strategy, and protects against threats and attacks using a reasonable system architecture. Through experiments on different types of workflow parameters the validity and effectiveness of the offloading algorithm with privacy preservation was demonstrated.
Chapter
Runtime IoT data fluctuation brings challenges for optimizing the resource allocation for a data stream processing (DSP) flow in a cloud-edge environment. It can result in extra high latency for a flow. Optimized strategy of dynamic resource allocation is still hard to design to timely dealing with the IoT data fluctuation. In this paper, the above challenge is abstracted and redefined as the service deployment problem. An improved GA optimization algorithm, integrating with the IoT data fluctuation prediction ability, is proposed to handle IoT data fluctuations during the running of a DSP flow. Effectiveness of the proposed approach is evaluated based on the real datasets from a real application.
Chapter
The ever-growing number of vehicles brings forth challenges in traffic management. This causes various traffic management issues in urban cities around the world. Some of the issues are: delay in emergency/alarming situations, non-deterministic waiting time of local transport, increased fuel consumption, etc. To help the people travelling by local transport in the cities by knowing the position of the bus, at a specific time, would ease them from indefinite wait or pass over of bus. In this chapter, our focus is to provide a trouble free, smart and innovative IoT-based traffic assistant that can solve real time transport related problems. A Hierarchical Peer Connected Fog Architecture (HPCFA) is proposed to lower latency time and computational overhead. In HPCFA, the fog nodes are organized in a hierarchy where the peer fog nodes present at the same level are also interconnected with each other. The data from IoT devices equipped on the roads will capture the position of the vehicle which is then transmitted to the nearest fog node. This fog node will further transmit the information through HPCFA to the user. Using HPCFA, the total energy consumption is also reduced to some extent. The proposed architecture is very flexible, as it works both with fog nodes or without fog nodes and directly with the cloud. Further, an android application is also developed for the proposed architecture. The simulations and results are also displayed.
Article
Full-text available
This paper presents an all-inclusive review of the optimization methods and their applications in fog computing. We put forward a taxonomy of the optimization techniques and then discuss their applications to solving various optimization problems in this field. Moreover, we develop a glossary of different optimization metrics in the literature and classify various evaluation environments used to test the solutions of different algorithms. The distribution of the relevant publications and the threats to the validity of the study are also discussed. Finally, we present the challenges in the existing literature and the future trends for research in fog computing optimization.
Article
Resource allocation and task scheduling is a complex task in fog computing environment because of the inherent heterogeneity among the fog devices. The proposed work attempts to solve the problem by using the popular multi criteria decision making methods such as AHP and TOPSIS. The goal of this paper is to propose a model for performance oriented task - resource mapping in a fog computing environment. MIPS, RAM & storage, uplink latency, downlink latency, uplink bandwidth, downlink bandwidth, trust, cost per MIPS, cost per memory, cost per storage and cost per bandwidth are the various performance characteristics considered in this work for task – resource mapping. Two different multi-criteria decision making methods are employed in order to assess the performance characteristics of the fog devices. In the first method, Analytic Hierarchy Process (AHP) is used for both priority weight calculation and ranking of fog devices. In the second method, AHP is used for priority weight calculation, based on the weights yielded by AHP, Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) algorithm is executed in order to rank the fog devices. Then the fog devices can be allocated to the tasks based on its rank. Furthermore, a motivational example is also demonstrated to validate the proposed method. Simulation results show that the proposed technique exhibits superior performance over other scheduling algorithms in the fog environment by incorporating performance, security, and cost metrics into scheduling decisions.
Article
The fog radio access network (Fog-RAN) has been considered a promising wireless access architecture to help shorten the communication delay and relieve the large data delivery burden over the backhaul links. However, limited by conventional inflexible communication design, Fog-RAN cannot be used in some complex communication scenarios. In this study, we focus on investigating a more intelligent Fog-RAN to assist the communication in a high-speed railway environment. Due to the train’s continuously moving, the communication should be designed intelligently to adapt to channel variation. Specifically, we dynamically optimize the power allocation in the remote radio heads (RRHs) to minimize the total network power cost considering multiple quality-of-service (QoS) requirements and channel variation. The impact of caching on the power allocation is considered. The dynamic power optimization is analyzed to obtain a closed-form solution in certain cases. The inherent tradeoff among the total network cost, delay and delivery content size is further discussed. To evaluate the performance of the proposed dynamic power allocation, we present an invariant power allocation counterpart as a performance comparison benchmark. The result of our simulation reveals that dynamic power allocation can significantly outperform the invariant power allocation scheme, especially with a random caching strategy or limited caching resources at the RRHs.
Article
In fog-assisted IoT systems, it is a common practice to offload tasks from IoT devices to their nearby fog nodes to reduce task processing latencies and energy consumptions. However, the design of online energy-efficient scheme is still an open problem because of various uncertainties in system dynamics such as processing capacities and transmission rates. Moreover, the decision-making process is constrained by resource limits on fog nodes and IoT devices, making the design even more complicated. In this paper, we formulate such a task offloading problem with unknown system dynamics as a combinatorial multi-armed bandit (CMAB) problem with time-averaged energy consumption constraints. Through an effective integration of online learning and online control, we propose a Learning-Aided Green Offloading (LAGO) scheme. In LAGO, we employ bandit learning methods to handle the exploitation-exploration tradeoff and utilize virtual queue techniques to deal with the time-averaged constraints. Our theoretical analysis shows that LAGO reduces the average task latency with a tunable sublinear regret bound over a finite time horizon and satisfies the time-averaged energy constraints. We conduct extensive simulations to verify such theoretical results.
Article
For NFV systems, the key design space includes the function chaining for network requests and the resource scheduling for servers. The problem is challenging since NFV systems usually require multiple (often conflicting) design objectives and the computational efficiency of real-time decision making with limited information. Furthermore, the benefits of predictive scheduling to NFV systems still remain unexplored. In this article, we propose POSCARS, an efficient predictive and online service chaining and resource scheduling scheme that achieves tunable trade-offs among various system metrics with stability guarantee. Through a careful choice of granularity in system modeling, we acquire a better understanding of the trade-offs in our design space. By a non-trivial transformation, we decouple the complex optimization problem into a series of online sub-problems to achieve the optimality with only limited information. By employing randomized load balancing techniques, we propose three variants of POSCARS to reduce the overheads of decision making. Theoretical analysis and simulations show that POSCARS and its variants require only mild-value of future information to achieve near-optimal system cost with an ultra-low request response time.
Article
In fog-assisted Internet-of-Things systems, it is a common practice to cache popular content at the network edge to achieve high quality of service. Due to uncertainties, in practice, such as unknown file popularities, the cache placement scheme design is still an open problem with unresolved challenges: 1) how to maintain time-averaged storage costs under budgets; 2) how to incorporate online learning to aid cache placement to minimize performance loss [also known as (a.k.a.) regret]; and 3) how to exploit offline historical information to further reduce regret. In this article, we formulate the cache placement problem with unknown file popularities as a constrained combinatorial multiarmed bandit problem. To solve the problem, we employ virtual queue techniques to manage time-averaged storage cost constraints, and adopt history-aware bandit learning methods to integrate offline historical information into the online learning procedure to handle the exploration–exploitation tradeoff. With an effective combination of online control and history-aware online learning, we devise a cache placement scheme with history-aware bandit learning called CPHBL . Our theoretical analysis and simulations show that CPHBL achieves a sublinear time-averaged regret bound. Moreover, the simulation results verify CPHBL’s advantage over the deep reinforcement learning-based approach.
Conference Paper
Advancement in IoT research encouraged the generation of ubiquitous computing devices and changed their way of utilization. The high spawning of data through these computing devices and its computation demands the evolvement of a new paradigm. Instead of using a centralized cloud-based system, a new distributed computing environment has been emerged for feature computational and storage capabilities at the periphery of the network called fog computing. The main idea behind fog computing is to utilize the processing and storage resources of fog nodes by offloading the computation-intensive IoT tasks. In the fog-enabled Cloud/IoT architecture, resource-constraint IoT devices offload their resource-hungry tasks to nearby deployed fog nodes to reduce the processing time and fast execution. Fog enabled processing guarantees fast response time and better energy management. Therefore, we are proposing a optimal offloading decision strategy and resource allocation scheme in fog atmosphere by minimizing the total energy consumption of IoT users concerning the latency of communication and computation. To optimally offload the tasks to fog node an optimization problem is designed and an improved version of branch-and-bound tree is proposed for its evaluation. Some of the metrics we consider for performance evaluation are average task latency, average energy consumption per task and average remaining energy per IoT node. Simulation results show that using proposed offloading decision strategy, network performance improve significantly.
Article
Fog computing extends cloud computing capabilities to the edge of the network and reduces high latency and network congestion. This paradigm enables portions of a transaction to be executed at a fog server and other portions at the cloud. Fog servers are generally not as robust as cloud servers; at peak loads, the data that cannot be processed by fog servers is processed by cloud servers. The data that need to be processed by the cloud is sent over a WAN. Therefore, only a fraction of the total data needs to travel through the WAN, as compared with a pure cloud computing paradigm. Additionally, the fog/cloud computing paradigm reduces the cloud processing load when compared with the pure cloud computing model. This paper presents a multiclass closed-form analytic queuing network model that is used by an autonomic controller to dynamically change the fraction of processing between edge and cloud servers in order to maximize a utility function of response time and cost. Experiments show that the controller can maintain a high utility in the presence of a wide variations of request arrival rates for various workloads.
Chapter
Full-text available
The disastrous phenomena of landslide is known as downward and outward movement of the materials forming the slope under the influence of gravitational forces as well as human interference. The state of Himachal Pradesh, India often gets disrupted due to monsoonal rainfall triggered mass movements in the form of debris flows, slides, falls, etc. from the past three decades. To assess the loss of life and property because of landslide activities in the region through hazard zonation mapping is an important task which is helpful for future developmental and constructional activities along the transport road corridors. Road section from Lahru up to Chamba district headquarters via Banikhet in Chamba, Himachal Pradesh was chosen for landslide hazard zonation mapping in the present study based on the numerous landslide causative factors in the form of thematic maps delineated through ArcGIS. The raster and vector type of datasets were collected from the available sources for the software analysis. In addition to this, desk and field investigations were performed for selecting the study area and to collect landslide related information from various departments and the local public in the region. The certainty factor approach was adopted for delineating landslide hazard zonation map of the study area to categorize the map into five hazard zones. Depending upon the severity, landslide hazard map (LHZ) was further divided into five different categories like very low (17%), low (27%), moderate (26%), high (20%), and very high (10%). The accuracy of the hazard zonation map was evaluated using the area under the curve method. This map will be helpful for various stakeholders like geologists, geo-technicians, engineers, Public Works Department for future development along the selected road corridor to mitigate landslide geohazard activities.
Chapter
Fog Computing has emerged as an area that provides an efficient platform for computing to support sustainable development. This latest computing paradigm could be an expansion of cloud computing. The main focus of fog computing is to decrease the load on the cloud due to an extreme number of IoT devices increased in the last few years. Sustainable resource allocation is the most promising challenge in fog computing that can save bandwidth, reduce latency, and energy consumption. In this article, a comprehensive review is done on resource allocation techniques used by many researchers in fog computing to find shortcomings for further improvement. The future direction and research gaps are also discussed at the end, so the author can easily find the way to do future research.
Chapter
Microplastic pollution is one of the emerging challenges aquatic and terrestrial environment is facing. Plastic particles having diameter less than 5 mm are considered to be microplastics. These are becoming the pollutants of increasing concern. Microplastics are complex contaminants and remain in environment for a longer duration. Due to production of enormous amounts of plastics annually, they are released into the environment as fragments, fibres, foams, microfibres and so on. These fragments of plastics when enter the pathway of water become a threat for aquatic species as well as human beings. In this review, recent studies and research on microplastic sources, their environmental impacts and various removal strategies are summed up. Effect of microplastic pollution at various trophic levels of aquatic and terrestrial ecosystems has been discussed. Microplastic composition and distribution with the context of various water bodies in the world and technology employed in the analysis with their shortcomings have been investigated in this review. In the end, some challenges and future line prospectus are suggested to increase development of more research in this particular area of concern.
Article
Full-text available
The Internet of Things (IoT) connects a huge number of resource-constraint IoT devices to the Internet, which generate massive amount of data that can be offloaded to the cloud for computation. As some of the applications may require very low latency, the emerging mobile edge computing (MEC) architecture offers cloud services by deploying MEC servers at the mobile base stations (BSs). The IoT devices can transmit the offloaded data to the BS for computation at the MEC server. Narrowband Internet of Things (NB-IoT) is a new cellular technology for the transmission of IoT data to the BS. In this paper, we propose a joint computation offloading and multi-user scheduling algorithm in NB-IoT edge computing system that minimizes the long-term average weighted sum of delay and power consumption under stochastic traffic arrival. We formulate the dynamic optimization problem into an infinite-horizon average-reward continuous-time Markov decision process (CTMDP) model. In order to deal with the curse-of-dimensionality problem, we use the approximate dynamic programming techniques, i.e., the linear value-function approximation and TD learning with post-decision state and semi-gradient descent method, to derive a simple algorithm for the solution of the CTMDP model. The proposed algorithm is semi-distributed, where the offloading algorithm is performed locally at the IoT devices, while the scheduling algorithm is auction-based where the IoT devices submit bids to the BS to make the scheduling decision centrally. Simulation results show that the proposed algorithm provides significant performance improvement over the two baseline algorithms and the MUMTO algorithm which is designed based on the deterministic task model.
Article
Full-text available
Cooperation between the fog and the cloud in mobile cloud computing environments could offer improved offloading services to smart mobile user equipment (UE) with computation intensive tasks. In this paper, we tackle the computation offloading problem in a mixed fog/cloud system by jointly optimizing the offloading decisions and the allocation of computation resource, transmit power and radio bandwidth, while guaranteeing user fairness and maximum tolerable delay. This optimization problem is formulated to minimize the maximal weighted cost of delay and energy consumption (EC) among all UEs, which is a mixed-integer non-linear programming problem. Due to the NP-hardness of the problem, we propose a low-complexity suboptimal algorithm to solve it, where the offloading decisions are obtained via semidefinite relaxation and randomization and the resource allocation is obtained using fractional programming theory and Lagrangian dual decomposition. Simulation results are presented to verify the convergence performance of our proposed algorithms and their achieved fairness among UEs, and the performance gains in terms of delay, EC and the number of beneficial UEs over existing algorithms.
Conference Paper
Full-text available
While mobile edge computing (MEC) alleviates the computation and power limitations of mobile devices, additional latency is incurred when offloading tasks to remote MEC servers. In this work, the power-delay tradeoff in the context of task offloading is studied in a multi-user MEC scenario. In contrast with current system designs relying on average metrics (e.g., the average queue length and average latency), a novel network design is proposed in which latency and reliability constraints are taken into account. This is done by imposing a probabilistic constraint on users' task queue lengths and invoking results from extreme value theory to characterize the occurrence of low-probability events in terms of queue length (or queuing delay) violation. The problem is formulated as a computation and transmit power minimization subject to latency and reliability constraints, and solved using tools from Lyapunov stochastic optimization. Simulation results demonstrate the effectiveness of the proposed approach, while examining the power-delay tradeoff and required computational resources for various computation intensities.
Article
Full-text available
Driven by the visions of Internet of Things and 5G communications, recent years have seen a paradigm shift in mobile computing, from the centralized Mobile Cloud Computing towards Mobile Edge Computing (MEC). The main feature of MEC is to push mobile computing, network control and storage to the network edges (e.g., base stations and access points) so as to enable computation-intensive and latency-critical applications at the resource-limited mobile devices. MEC promises dramatic reduction in latency and mobile energy consumption, tackling the key challenges for materializing 5G vision. The promised gains of MEC have motivated extensive efforts in both academia and industry on developing the technology. A main thrust of MEC research is to seamlessly merge the two disciplines of wireless communications and mobile computing, resulting in a wide-range of new designs ranging from techniques for computation offloading to network architectures. This paper provides a comprehensive survey of the state-of-the-art MEC research with a focus on joint radio-and-computational resource management. We also discuss a set of issues, challenges and future research directions for MEC research, including MEC system deployment, cache-enabled MEC, mobility management for MEC, green MEC, as well as privacy-aware MEC. Advancements in these directions will facilitate the transformation of MEC from theory to practice. Finally, we introduce recent standardization efforts on MEC as well as some typical MEC application scenarios.
Conference Paper
Full-text available
Despite the broad utilization of cloud computing, some applications and services still cannot benefit from this popular computing paradigm due to inherent problems of cloud computing such as unacceptable latency, lack of mobility support and location-awareness. As a result, fog computing, has emerged as a promising infrastructure to provide elastic resources at the edge of network. In this paper, we have discussed current definitions of fog computing and similar concepts, and proposed a more comprehensive definition. We also analyzed the goals and challenges in fog computing platform, and presented platform design with several exemplar applications. We finally implemented and evaluated a prototype fog computing platform.
Article
Full-text available
In this work we present a large scale comparison study for the major machine learning models for time series forecasting. Specifically, we apply the models on the monthly M3 time series competition data (around a thousand time series). There have been very few, if any, large scale comparison studies for machine learning models for the regression or the time series forecasting problems, so we hope this study would fill this gap. The models considered are multilayer perceptron, Bayesian neural networks, radial basis functions, generalized regression neural networks (also called kernel regression), K-nearest neighbor regression, CART regression trees, support vector regression, and Gaussian processes. The study reveals significant differences between the different methods. The best two methods turned out to be the multilayer perceptron and the Gaussian process regression. In addition to model comparisons, we have tested different preprocessing methods and have shown that they have different impacts on the performance.
Article
Internet of Things (IoT) as a prospective platform to develop mobile applications, is facing with significant challenges posed by the tension between resource-constrained mobile smart devices and low-latency demanding applications. Recently, mobile edge computing (MEC) is emerging as a cornerstone technology to address such challenges in IoT. In this paper, by leveraging social ties in human social networks, we investigate the optimal dynamic computation offloading mode selection to jointly minimize the total tasks’ execution latency and the mobile smart devices’ energy consumption in MEC-aided low-latency IoT. Different from the previous studies, which mostly focus on how to exploit social tie structure among mobile smart device users to construct the permutation of all the feasible modes, we consider dynamic computation offloading mode selection with social awareness-aided network resource assignment, involving both the computing resources and transmit power from heterogeneous mobile smart devices. On one hand, we formulate the dynamic computation offloading mode selection into the infinite-horizon time-average renewal-reward problems subject to time average latency constraints on a collection of penalty processes. On the other hand, an efficient solution is also developed, which elaborates on a Lyapunov optimization based approach, i.e., drift-plus-penalty (DPP) algorithm. Numerical simulations are provided to validate the theoretical analysis and assess the performance of the proposed dynamic social-aware computation offloading mode selection method considering different configurations of the IoT network parameters.
Article
Driven by the vision of edge computing and the success of rich cognitive services based on artificial intelligence, a new computing paradigm, edge cognitive computing (ECC), is a promising approach that applies cognitive computing at the edge of the network. ECC has the potential to provide the cognition of users and network environmental information, and further to provide elastic cognitive computing services to achieve a higher energy efficiency and a higher Quality of Experience (QoE) compared to edge computing. This article first introduces our architecture of the ECC and then describes its design issues in detail. Moreover, we propose an ECC-based dynamic service migration mechanism to provide insight into how cognitive computing is combined with edge computing. In order to evaluate the proposed mechanism, a practical platform for dynamic service migration is built up, where the services are migrated based on the behavioral cognition of a mobile user. The experimental results show that the proposed ECC architecture has ultra-low latency and a high user experience, while providing better service to the user, saving computing resources, and achieving a high energy efficiency.
Article
In this paper, we consider the problem of task offloading in a software-defined access network, where IoT devices are connected to fog computing nodes by multi-hop IoT access-points (APs). The proposed scheme considers the following aspects in a fog-computing-based IoT architecture: 1) optimal decision on local or remote task computation; 2) optimal fog node selection; and 3) optimal path selection for offloading. Accordingly, we formulate the multi-hop task offloading problem as an integer linear program (ILP). Since the feasible set is non-convex, we propose a greedy-heuristic-based approach to efficiently solve the problem. The greedy solution takes into account delay, energy consumption, multi-hop paths, and dynamic network conditions, such as link utilization and SDN rule-capacity. Experimental results show that the proposed scheme is capable of reducing the average delay and energy consumption by 12% and 21%, respectively, compared with the state of the art.
Article
Utilizing the intelligence at the network edge, edge computing paradigm emerges to provide time-sensitive computing services for Internet of Things. In this paper, we investigate sustainable computation offloading in an edge-computing system that consists of energy harvesting-enabled mobile devices (MDs) and a dispatcher. The dispatcher collects computation tasks generated by IoT devices with limited computation power, and offloads them to resourceful MDs in exchange for rewards. We propose an online Rewards-optimal Auction (RoA) to optimize the long-term sum-of-rewards for processing offloaded tasks, meanwhile adapting to the highly dynamic energy harvesting (EH) process and computation task arrivals. RoA is designed based on Lyapunov optimization and Vickrey-Clarke-Groves auction, the operation of which does not require a prior knowledge of the energy harvesting, task arrivals, or wireless channel statistics. Our analytical results confirm the optimality of tasks assignment. Furthermore, simulation results validate the analytical analysis, and verify the efficacy of the proposed RoA.
Article
With proliferation of computation-intensive Internet of Things (IoT) applications, the limited capacity of end devices can deteriorate service performance. To address this issue, computation tasks can be offloaded to the Mobile Edge Computing (MEC) for processing. However, it consumes considerable energy to transmit and process these tasks. In this paper, we study the energy efficient task offloading in MEC. Specifically, we formulate it as a stochastic optimization problem, with the objective of minimizing the energy consumption of task offloading while guaranteeing the average queue length. Solving this offloading optimization problem faces many technical challenges due to the uncertainty and dynamics of wireless channel state and task arrival process, and the large scale of solution space. To tackle these challenges, we apply stochastic optimization techniques to transform the original stochastic problem into a deterministic optimization problem, and propose an energy efficient dynamic offloading algorithm called EEDOA. EEDOA can be implemented in an online way to make the task offloading decisions with polynomial time complexity. Theoretical analysis is given to demonstrate that EEDOA can approximate the minimal transmission energy consumption while still bounding the queue length. Experiments results are presented which shows the EEDOA's effectiveness.
Article
The deployment of cloud and edge computing forms a three-tier mobile computing network, where each task can be processed locally, by the edge nodes, or by the remote cloud server. In this paper, we consider a cooperative three-tier computing network by leveraging the vertical cooperation among devices, edge nodes and cloud servers, as well as the horizontal cooperation between edge nodes. In this network, we jointly optimize the offloading decision and the computation resource allocation to minimize the average task duration subject to the limited battery capacity of devices. However, the formulated problem is a large-scale mixed integer non-linear optimization problem with the growing number of base stations and devices, which is NP-hard in general. To develop an efficient offloading scheme with low complexity, we conduct a series of reformulation based on reformulation linearization technology (RLT) and further propose a parallel optimization framework by utilizing alternating direction method of multipliers (ADMM) method and difference of convex functions (D.C.) programming. The proposed scheme decomposes the large-scale problem into some smaller subproblems, which are done across the multiple computation units in a parallel fashion to speed up the computation process. Simulation results demonstrate that the proposed scheme can obtain a near-optimal performance with low complexity, and can reduce up to 24% of the task duration compared with other schemes. Simulation also shows how much the vertical and horizontal computation cooperations affect the system performance under different network parameters.
Article
Fog computing is considered as a promising technology to meet the ever-increasing computation requests from a wide variety of mobile applications. By offloading the computation-intensive requests to the fog node or the central cloud, the performance of the applications, such as energy consumption and delay, are able to be significantly enhanced. Meanwhile, utilizing the recent advances of social network and energy harvesting techniques, the system performance could be further improved. In this paper, we take the social relationships of the energy harvesting MDs into the design of computational offloading scheme in fog computing. With the objective to minimize the social group execution cost, we advocate game theoretic approach and propose a dynamic computation offloading scheme designing the offloading process in fog computing system with energy harvesting MDs. Different queue models are applied to model the energy cost and delay performance. It can be seen that the proposed problem can be formulated as a Generalized Nash Equilibrium Problem (GNEP) and we can use exponential penalty function method to transform the original GNEP into a classical NEP and address it with semi-smooth Newton method with Armijo line search. The simulation results demonstrate the effectiveness of the proposed scheme.
Article
In this paper, we study sustainable resource allocation for cloud radio access networks (CRANs) powered by hybrid energy supplies (HES). Specifically, the central unit (CU) in the CRANs distributes data to a set of radio units (RUs) powered by both on-grid energy and energy harvested from green sources, and allocates channels to the selected RUs for downlink transmissions. We formulate an optimization problem to maximize the net gain of the system which is the difference between the user utility gain and on-grid energy costs, taking into consideration the stochastic nature of energy harvesting process, time-varying on-grid energy price, and dynamic wireless channel conditions. A resource allocation framework is developed to decompose the formulated problem into three subproblems, i.e., the hybrid energy management, data requesting, and channel and power allocation. Based on the solutions of the subproblems, we propose a net gain-optimal resource allocation (GRA) algorithm to maximize the net gain while stabilizing the data buffers and ensuring the sustainability of batteries. Performance analysis demonstrates that the GRA algorithm can achieve close-to-optimal net gain with bounded data buffer and battery capacity. Extensive simulations validate the analysis and demonstrate that GRA algorithm outperforms other algorithms in terms of the net gain and delay performance.
Article
Digital communication is an enormous and rapidly growing industry, roughly comparable in size to the computer industry. The objective of this text is to study those aspects of digital communication systems that are unique. That is, rather than focusing on hardware and software for these systems (which is much like that in many other fields), we focus on the fundamental system aspects of modern digital communication. Digital communication is a field in which theoretical ideas have had an unusually powerful impact on system design and practice. The basis of the theory was developed in 1948 by Claude Shannon, and is called information theory. For the first 25 years or so of its existence, information theory served as a rich source of academic research problems and as a tantalizing suggestion that communication systems could be made more efficient and more reliable by using these approaches. Other than small experiments and a few highly specialized military systems, the theory had little interaction with practice. By the mid 1970s, however, mainstream systems using information-theoretic ideas began to be widely implemented. The first reason for this was the increasing number of engineers who understood both information theory and communication system practice. The second reason was that the low cost and increasing processing power of digital hardware made it possible to implement the sophisticated algorithms suggested by information theory.
Article
Motivated by the increasing popularity of learning and predicting human user behavior in communication and computing systems, in this paper, we investigate the fundamental benefit of predictive scheduling, i.e., predicting and pre-serving arrivals, in controlled queueing systems. Based on a lookahead window prediction model, we first establish a novel equivalence between the predictive queueing system with a \emph{fully-efficient} scheduling scheme and an equivalent queueing system without prediction. This connection allows us to analytically demonstrate that predictive scheduling necessarily improves system delay performance and can drive it to zero with increasing prediction power. We then propose the \textsf{Predictive Backpressure (PBP)} algorithm for achieving optimal utility performance in such predictive systems. \textsf{PBP} efficiently incorporates prediction into stochastic system control and avoids the great complication due to the exponential state space growth in the prediction window size. We show that \textsf{PBP} can achieve a utility performance that is within $O(\epsilon)$ of the optimal, for any $\epsilon>0$, while guaranteeing that the system delay distribution is a \emph{shifted-to-the-left} version of that under the original Backpressure algorithm. Hence, the average packet delay under \textsf{PBP} is strictly better than that under Backpressure, and vanishes with increasing prediction window size. This implies that the resulting utility-delay tradeoff with predictive scheduling beats the known optimal $[O(\epsilon), O(\log(1/\epsilon))]$ tradeoff for systems without prediction.
Chapter
Convex optimization problems arise frequently in many different fields. A comprehensive introduction to the subject, this book shows in detail how such problems can be solved numerically with great efficiency. The focus is on recognizing convex optimization problems and then finding the most appropriate technique for solving them. The text contains many worked examples and homework exercises and will appeal to students, researchers and practitioners in fields such as engineering, computer science, mathematics, statistics, finance, and economics.
Conference Paper
Although there is tremendous interest in designing improved networks for data centers, very little is known about the network-level traffic characteristics of data centers today. In this paper, we conduct an empirical study of the network traffic in 10 data centers belonging to three different categories, including university, enterprise campus, and cloud data centers. Our definition of cloud data centers includes not only data centers employed by large online service providers offering Internet-facing applications but also data centers used to host data-intensive (MapReduce style) applications). We collect and analyze SNMP statistics, topology and packet-level traces. We examine the range of applications deployed in these data centers and their placement, the flow-level and packet-level transmission properties of these applications, and their impact on network and link utilizations, congestion and packet drops. We describe the implications of the observed traffic patterns for data center internal traffic engineering as well as for recently proposed architectures for data center networks.
Book
This text presents a modern theory of analysis, control, and optimization for dynamic networks. Mathematical techniques of Lyapunov drift and Lyapunov optimization are developed and shown to enable constrained optimization of time averages in general stochastic systems. The focus is on communication and queueing systems, including wireless networks with time-varying channels, mobility, and randomly arriving traffic. A simple drift-plus-penalty framework is used to optimize time averages such as throughput, throughput-utility, power, and distortion. Explicit performance-delay tradeoffs are provided to illustrate the cost of approaching optimality. This theory is also applicable to problems in operations research and economics, where energy-efficient and profit-maximizing decisions must be made without knowing the future. Topics in the text include the following: • Queue stability theory • Backpressure, max-weight, and virtual queue methods • Primal-dual methods for non-convex stochastic utility maximization • Universal scheduling theory for arbitrary sample paths • Approximate and randomized scheduling theory • Optimization of renewal systems and Markov decision systems Detailed examples and numerous problem set questions are provided to reinforce the main concepts.
Article
We consider the following natural model: customers arrive as a Poisson stream of rate λn, λ<1, at a collection of n servers. Each customer chooses some constant d servers independently and uniformly at random from the n servers and waits for service at the one with the fewest customers. Customers are served according to the first-in first-out (FIFO) protocol and the service time for a customer is exponentially distributed with mean 1. We call this problem the supermarket model. We wish to know how the system behaves and in particular we are interested in the effect that the parameter d has on the expected time a customer spends in the system in equilibrium. Our approach uses a limiting, deterministic model representing the behavior as n→∞ to approximate the behavior of finite systems. The analysis of the deterministic model is interesting in its own right. Along with a theoretical justification of this approach, we provide simulations that demonstrate that the method accurately predicts system behavior, even for relatively small systems. Our analysis provides surprising implications. Having d=2 choices leads to exponential improvements in the expected time a customer spends in the system over d=1, whereas having d=3 choices is only a constant factor better than d=2. We discuss the possible implications for system design