Support Load Balancing Research Articles

Cost-based Virtual Machine Scheduling for Data-as-a-Service

Data-as-a-Service (DaaS) is a branch of cloud computing that supports “querying the Web”. Due to its ultrahigh scale, it is essential to establish rules when defining resources’ costs and guidelines for infrastructure investments. Those decisions should prioritize minimizing the incidence of agreement breaches that compromise the performance of cloud services and optimize resources’ usage and services’ cost. This article aims to address the cost problem of DaaS by developing a model that optimizes the costs of querying distributed data sources over virtual machines spread across multisite data centers. We have designed and analyzed a cost model for DaaS, besides implementing a scheduling system to perform a cost-based VM assignment. To validate our model, we have studied and characterized a real-world DaaS system’s network and processing workloads. On average, our cost-based scheduling performs at least twice as well as the traditional round-robin approach. Our model also supports load balancing and infrastructure scalability when combined with an adaptive cost scheme that prioritizes VM allocation within the underutilized data centers and avoids sending VMs to data centers in the eminence of becoming over-utilized.

Improving Urgency-Based Backlog Sequencing of Jobs: An Assessment by Simulation

When order release is applied, jobs are withheld in a backlog from where they are released to meet certain performance targets. The decision that selects jobs for release is typically preceded by a sequencing decision. It was traditionally assumed that backlog sequencing is only responsible for releasing jobs on time, whereas more recent literature has argued that it can also support load balancing. Although the new load-based rules outperform time-based rules, they can be criticized for requiring workload information from the shop floor and for delaying large jobs. While some jobs will inevitably be delayed during periods of high load, we argue that this delaying decision should be under control of management. A simulation study of a wafer fab environment shows that a time-based rule matches the performance of more complex load-based backlog sequencing rules that have recently emerged. The new rule realizes the lowest percentage of tardy jobs if the lower bound that distinguishes between early and urgent jobs is set appropriately. It provides a simpler means of improving release performance, allowing managers to delay jobs that have adjustable due dates.

Cheetah: A High-Speed Programmable Load-Balancer Framework With Guaranteed Per-Connection-Consistency

Large service providers use load balancers to dispatch millions of incoming connections per second towards thousands of servers. There are two basic yet critical requirements for a load balancer: <i>uniform load distribution</i> of the incoming connections across the servers, which requires to support advanced load balancing mechanisms, and <i>per-connection-consistency</i> (PCC), i.e, the ability to map packets belonging to the same connection to the same server even in the presence of changes in the number of active servers and load balancers. Yet, simultaneously meeting these requirements has been an elusive goal. Today&#x2019;s load balancers minimize PCC violations at the price of non-uniform load distribution. This paper presents Cheetah, a load balancer that supports advanced load balancing mechanisms <i>and</i> PCC while being scalable, memory efficient, fast at processing packets, and offers comparable resilience to clogging attacks as with today&#x2019;s load balancers. The Cheetah LB design guarantees PCC for <i>any</i> realizable server selection load balancing mechanism and can be deployed in both stateless and stateful manners, depending on operational needs. We implemented Cheetah on both a software and a Tofino-based hardware switch. Our evaluation shows that a stateless version of Cheetah guarantees PCC, has negligible packet processing overheads, and can support load balancing mechanisms that reduce the flow completion time by a factor of <inline-formula> <tex-math notation="LaTeX">$2-3 \times $ </tex-math></inline-formula>.

Optimized scheduling techniques focused on powerful heuristics leveraging cloud services soft computing

PurposeDetermining the roles of multiple CSPs is important because it affects job costs and time off. The primary objective of this work is to ensure an efficient and complex distribution of resources in cloud-based computing. Workflow study of various algorithms such as ant colony optimization (ACO), differential evolution algorithm, genetic algorithm, particle swarm optimization (PSO), hybridization of the above algorithms (ADGP). For research, CSP’s tools are put all over the world.Design/methodology/approachThe main objective of this study is to effectively introduce cloud-based computing in CSPs. The algorithm minimizes resource response time and overall workflow tasks. It seeks to improve load balancing by modifying the algorithm to support load balancing. In the proposed multipurpose scheduling methods, the ADGP algorithm performs better than any other proposed algorithm during the resource response. This algorithm was found to be superior to the selected 200 sources and thousands of tasks. It reduces resource response time by copying service nodes through several sites. As this algorithm moves faster to the best solution, the response time of the resource is reduced compared to other algorithms.FindingsHybrid ACOs perform best when it comes to resource management when workloads are uniformly spread across multiple virtual machines. However, hybrids PSOs are better suited to choosing the best options to minimize costs. Overall, an optimal cloud-based scheduling solution can be successfully simulated using CloudSim in CSP to share resources between end-users to support consumers and users effectively.Originality/valueHybrid ACOs perform best when it comes to resource management when workloads are uniformly spread across multiple virtual machines. However, hybrids PSOs are better suited to choosing the best options to minimize costs. Overall, an optimal cloud-based scheduling solution can be successfully simulated using CloudSim in CSP to share resources between end-users to support consumers and users effectively.

A Comprehensive Open-Source Simulation Framework for LiFi Communication

Light Fidelity (LiFi) is a new candidate for wireless networking that utilizes the visible light spectrum and exploits the existing lighting infrastructure in the form of light-emitting diodes (LEDs). It provides point-to-point and point-to-multipoint communication on a bidirectional channel at very high data rates. However, the LiFi has small coverage, and its optical gain is closely related to the receiver’s directionality vis-à-vis the transmitter, therefore it can experience frequent service outages. To provide reliable coverage, the LiFi is integrated with other networking technologies such as wireless fidelity (WiFi) thus forming a hybrid system. The hybrid LiFi/WiFi system faces many challenges including but not limited to seamless integration with the WiFi, support for mobility, handover management, resource sharing, and load balancing. The existing literature has addressed one or the other aspect of the issues facing LiFi systems. There are limited free source tools available to holistically address these challenges in a scalable manner. To this end, we have developed an open-source simulation framework based on the network simulator 3 (ns-3), which realizes critical aspects of the LiFi wireless network. Our developed ns-3 LiFi framework provides a fully functional AP equipped with the physical layer and medium access control (MAC), a mobility model for the user device, and integration between LiFi and WiFi with a handover facility. Simulation results are produced to demonstrate the mobility and handover capabilities, and the performance gains from the LiFi-WiFi hybrid system in terms of packet delay, throughput, packet drop ratio (PDR), and fairness between users. The source code of the framework is made available for the use of the research community.

ACC: Aerial Cell Cloning protocol for efficient session continuity in LAP replacement for 6G

Abstract Low Altitude Platform (LAP) based aerial cells improve the performance of the cellular network dynamically. However, the small form factor of LAPs constraints the battery size, limiting its hovering time. LAPs should be replaced when the available power or available capacity goes below a threshold, in addition to reasons like a mechanical malfunction. Replacement requires data sessions from all the connected user equipment (UEs) to transfer from the old LAP to the new LAP. Classical schemes like (i) UE handover and (ii) carrier link management support UE mobility and load balancing. The classical schemes have higher latencies, which scales if there are many connected UEs, leading to increased data session interruption if applied in LAP replacement. The interruption leads to a degraded user experience during gaming and multimedia consumption. In this paper, to the best of our knowledge for the first time, we propose a novel Aerial Cell Cloning (ACC) protocol for network management applied during LAP replacement. Through call flow analysis and mathematical modeling backed with extensive system-level simulation, we show that the proposed ACC protocol reduces the session transfer latency by over 68% and brings down the LAP replacement time to a negligible value (99% reduction) when compared to existing classical schemes.

Scheduling independent tasks in cloud environment based on modified differential evolution

SummaryCloud computing has been widely adopted in practical applications due to its strong calculating ability and high parallel feature. Although cloud computing can achieve significant cost reduction and flexibility enhancement, it results in a serious task scheduling problem. As one of the key techniques for automate management of cloud resources, task scheduling plays an important role in improving system utilization and supporting load balancing. In this article, we focus on the scheduling problem of independent tasks in cloud environment with heterogeneous and distributed resources. First, with models of resources and tasks, we present an exact formulation based on linear programming to fully search solution space and produce optimal allocation schemes for tasks. Then, inspired from the differential evolution method, we propose a population‐based approach to allocate tasks to their suitable resources such that the total time cost would be minimized. Experiments with multi‐task sets are conducted to show the convergence and efficiency of the proposed approach.

Dynamic Graph Partitioning Scheme for Supporting Load Balancing in Distributed Graph Environments

As dynamic graph data have been actively used, incremental graph partition schemes have been studied to efficiently store and manage large graphs. In this paper, we propose a vertex-cut based novel incremental graph partitioning scheme that supports load balancing in a distributed environment. The proposed scheme chooses the load of each node that considers its storage utilization and throughput as the partitioning criterion. The proposed scheme defines hot data that means a particular vertex frequently searched among graphs requested by queries. We manage and utilize hot data for graph partitioning. Finally, we perform vertex-cut based dynamic graph partitioning by using a vertex replication index, the load each node, and hot data to distribute the load evenly in a distributed environment. In order to verify the superiority of the proposed partitioning scheme, we compare it with the existing partitioning schemes through a variety of performance evaluations.

Distributed Self-Stabilizing Capacitated Maximal Independent Set Construction in Wireless Sensor Networks

Wireless sensor networks (WSNs) are composed of a large number of wireless self-organized sensor nodes connected through a wireless decentralized distributed network without the aid of a predefined infrastructure. Fault-tolerance and power management are fundamental challenges in WSNs. A WSN is self-stabilizing if it can initially start at any state and obtain a legitimate state in a finite time without any external intervention. Self-stabilization is an important method for providing fault-tolerance in WSNs. Maximal independent set (MIS) is an extensively used structure for many important applications such as clustering (Randhawa and Jain in Wirel Personal Commun 97(3):3355, 2017. https://doi.org/10.1007/s11277-017-4674-5 ) and routing (Attea et al. in Wirel Personal Commun 81(2):819, 2015. https://doi.org/10.1007/s11277-014-2159-3 ; Lipinski in Wirel Personal Commun 101(1):251, 2018. https://doi.org/10.1007/s11277-018-5686-5 ) in WSNs. The capacitated MIS (CapMIS) problem is an extension of MIS in that each node has a capacity that determines the number of nodes it may dominate. In this paper, we propose a distributed self-stabilizing capacitated maximal independent set algorithm (CapMIS) in order to reduce energy consumption and support load balancing in WSNs. To the best of our knowledge, this is the first algorithm in this manner. The algorithm is validated through theoretical analysis as well as testbed implementations and simulations.

A novel approach for energy- and memory-efficient data loss prevention to support Internet of Things networks

Internet of Things integrates various technologies, including wireless sensor networks, edge computing, and cloud computing, to support a wide range of applications such as environmental monitoring and disaster surveillance. In these types of applications, IoT devices operate using limited resources in terms of battery, communication bandwidth, processing, and memory capacities. In this context, load balancing, fault tolerance, and energy and memory efficiency are among the most important issues related to data dissemination in IoT networks. In order to successfully cope with the abovementioned issues, two main approaches—data-centric storage and distributed data storage—have been proposed in the literature. Both approaches suffer from data loss due to memory and/or energy depletion in the storage nodes. Even though several techniques have been proposed so far to overcome the abovementioned problems, the proposed solutions typically focus on one issue at a time. In this article, we propose a cross-layer optimization approach to increase memory and energy efficiency as well as support load balancing. The optimization problem is a mixed-integer nonlinear programming problem, and we solve it using a genetic algorithm. Moreover, we integrate the data-centric storage features into distributed data storage mechanisms and present a novel heuristic approach, denoted as Collaborative Memory and Energy Management, to solve the underlying optimization problem. We also propose analytical and simulation frameworks for performance evaluation. Our results show that the proposed method outperforms the existing approaches in various IoT scenarios.

Improving Switch-to-Controller Assignment with Load Balancing in Multi-controller Software Defined WAN (SD-WAN)

Software defined networking (SDN) gains a lot of interest from network operators due to its ability to offer flexibility, efficiency and fine-grained control over forwarding elements (FE) by decoupling control and data planes. In the control plane, a centralized node, denoted controller, receives requests from ingress switches and makes decisions on path forwarding. Unfortunately, requests processing may lead to controller performance degradation as the number of incoming requests goes up. This paper deals with the controller performance issue in Software Defined WAN (SD-WAN). Mainly, it proposes a new approach to optimize the switch-to-controller assignment problem with load balancing support. The issue is formulated as a Minimum Cost Bipartite Assignment optimization problem which is solved using an improvement of the Hungarian algorithm. The new algorithm is based on the introduction of the load-driven penalty concept which aims to achieve a trade-off between the round trip time and the controller load. Finally, a new protocol denoted Distributed Hungarian-based Assignment Protocol (DHAP) is described as an implementation of the proposed solution in multi-controller environments. As shown in results, the proposed solution outperforms parallel schemes in terms of flow setup time and load balancing.

Dynamic Partitioning Supporting Load Balancing for Distributed RDF Graph Stores

Various resource description framework (RDF) partitioning methods have been studied for the efficient distributed processing of a large RDF graph. The RDF graph has symmetrical characteristics because subject and object can be used interchangeably if predicate is changed. This paper proposes a dynamic partitioning method of RDF graphs to support load balancing in distributed environments where data insertion and change continue to occur. The proposed method generates clusters and subclusters by considering the usage frequency of the RDF graph that are used by queries as the criteria to perform graph partitioning. It creates a cluster by grouping RDF subgraphs with higher usage frequency while creating a subcluster with lower usage frequency. These clusters and subclusters conduct load balancing by using the mean frequency of queries for the distributed server and conduct graph data partitioning by considering the size of the data stored in each distributed server. It also minimizes the number of edge-cuts connected to clusters and subclusters to minimize communication costs between servers. This solves the problem of data concentration to specific servers due to ongoing data changes and additions and allows efficient load balancing among servers. The performance results show that the proposed method significantly outperforms the existing partitioning methods in terms of query performance time in a distributed server.

An efficient iterative graph data processing framework based on bulk synchronous parallel model

SummaryGraph data processing has been widely applied in a variety of domains such as industry, science, social network, and so on. It therefore has stimulated many efforts devoted to this area. To embrace the fast development trend of big graph data, graph data processing based on Pregel‐like systems has been regarded as one of the most promising ways and has widely attracted the attention of researchers. However, it still remains in its early stage and there still exist many challenges. In Pregel, the superstep synchronization is time consuming as the graph data iteration operation requires multiple synchronizations. Furthermore, the graph data partition strategy adopted by Pregel fails to support load balancing, therefore causing the increase of network I/O overhead as the scale of graph data grows. To address these issues, this paper presents an efficient computational framework for graph data processing based on the bulk synchronous parallel model. The global synchronization control mechanism is improved by determining the start time of the next round of superstep through counting the number of global message files. Furthermore, an improved graph data partition mechanism based on a balanced hash method is proposed to reduce the communication overhead between different partitions of sub‐graph computational tasks. We also re‐design the PageRank algorithm to verify the effectiveness of the proposed framework. Experimental results on different real‐world datasets verify the efficiency of our proposed framework as it outperforms Giraph (an open source Pregel‐like system) by 58%−69%, and achieves 10×−17× performance improvement over Hadoop.

A Novel On-Line Association Algorithm for Supporting Load Balancing in Multiple-AP Wireless LAN

Nowadays, wireless LAN has become the most widely deployed technology in mobile devices for providing Internet access. Operators and service providers remarkably increase the density of wireless access points in order to provide their subscribers with better connectivity and user experience. As a result, WLAN users usually find themselves covered by multiple access points and have to decide which one to associate with. In traditional implementations, most wireless stations would select the access point with the strongest signal, regardless of traffic load on that access point, which might result in heavy congestion and unfair load. In this paper, we propose a novel on-line association algorithm to deal with any sequence of STAs during a long-term time such as one day. the performance of our algorithm is evaluated through simulation and experiments. Simulation results show that our algorithm improves the overall WLAN throughput by up to 37%, compared with the conventional RSSI-based approach. Our algorithm also performs better than SSF (Strongest Signal First) and LAB (Largest Available Bandwidth) in the experiments.

A Novel Contribution to Control a Wind Turbine System for Power Quality Improvement in Electrical Networks

The typical control strategy for wind energy conversion systems (WECS) is the maximum power coefficient tracking method. However, this method limits the participation of wind turbines in power quality improvement in the network. To overcome this limitation, this paper presents, in the first step, a new methodology that derives the power coefficient that permits to expand the role of WECS based on doubly-fed induction generators, to load balancing and reactive power support to the grid. The proposed method allows the WECS to provide active power to the grid under better conditions compared to the typical control method. However, in case multilevel converters are used for grid connection of wind turbines, the application of the proposed method may imply the accentuation of the inherent problem of dc capacitor-voltage imbalance of these converters, leading to jeopardize their operation. For this reason, in the second step, a predictive control algorithm capable of simultaneously improving the power quality in power grids and ensuring reliable operation of the associated converter is proposed. The simulation and experimental results are presented.

The Future Promise of Vehicle-to-Grid (V2G) Integration: A Sociotechnical Review and Research Agenda

Vehicle-grid integration (VGI) describes various approaches to link the electric power system and the transportation system in ways that may benefit both. VGI includes systems that treat plug-in electric vehicles (PEVs) as controllable load with a unidirectional flow of electricity, such as “smart” or “controlled” charging or time-of-use (TOU) pricing. VGI typically encompasses vehicle-to-grid (V2G), a more technically advanced vision with bidirectional flow of electricity between the vehicle and power grid, in effect treating the PEV as a storage device. Such VGI systems could help decarbonize transportation, support load balancing, integrate renewable energy into the grid, increase revenues for electricity companies, and create new revenue streams for automobile owners. This review introduces various aspects and visions of VGI based on a comprehensive review. In doing so, it identifies the possible benefits, opportunities, and barriers relating to V2G, according to technical, financial, socio-environmental, and behavioral components. After summarizing our sociotechnical approach and the various opportunities and barriers indicated by existing literature, we construct a proposed research agenda to provide insights into previously understudied and unstudied research objectives. We find that the majority of VGI studies to date focus on technical aspects of VGI, notably on the potential of V2G systems to facilitate load balancing or to minimize electricity costs, in some cases including environmental goals as constraints. Only a few studies directly investigate the role of consumer acceptance and driver behavior within such systems, and barely any studies address the need for institutional capacity and cross-sectoral policy coordination. These gaps create promising opportunities for future research.

Wireless Interface Bonding Supporting In-Order Delivery and Automatic Load Balancing

Wireless interface bonding is an important technique in multiradio systems, where multiple radios can be combined to establish a fast link between nodes. However, existing interface bonding modules are not well-suited for wireless interfaces due to lack of functionalities to cope with varying channel capacity. Variance in throughput and delay of the individual radios can cause significant degradation of transmission control protocol throughput due to suboptimal scheduling and out-of-order delivery. Since condition of wireless channels is dynamic, capacity of each channel cannot be predicted offline which makes the packet scheduling challenging. Naive queue management for in-order delivery could also be hazardous to the throughput due to long dwell time of packets at the receiver queue. This paper presents an extension to Linux bonding driver called the new load balancing (NLB) mode, which is designed to combine multiple WLAN interfaces into a single virtual interface. The NLB mode supports automatic load balancing and in-order delivery, and together, they make a multichannel link that behaves like a single link with aggregated bandwidth. Automatic load balancing is done using receiver measurements on interarrival time of consecutive packets. In-order delivery is implemented using a receive queue and a fast loss detection scheme which prevents long dwell time of packets at the queue. The NLB mode does not require modifications to physical network device drivers which makes it usable with devices from any vendor. Performance evaluations using OpenWrt-based wireless devices show that the NLB mode can efficiently aggregate bandwidths of channels under various conditions.

Efficient Cellular Load Balancing Through Mobility-Enriched Vehicular Communications

Supporting effective load balancing is paramount for increasing network utilization efficiency and improving the perceivable user experience in emerging and future cellular networks. At the same time, it is becoming increasingly alarming that current communication practices lead to excessive energy wastes both at the infrastructure side and at the terminals. To address both these issues, this paper discusses an innovative communication approach enabled by the implementation of device-to-device (d2d) communication over cellular networks. The technique capitalizes on the delay tolerance of a significant portion of Internet applications and the inherent mobility of the nodes to achieve significant performance gains. For delay-tolerant messages, a mobile node can postpone message transmission—in a store–carry and forward manner—for a later time to allow the terminal to achieve communication over a shorter range or to postpone communication to when the terminal enters a cooler cell, before engaging in communication. Based on this framework, a theoretical model is introduced to study the generalized multihop d2d forwarding scheme where mobile nodes are allowed to buffer messages and carry them while in transit. Thus, a multiobjective optimization problem is introduced where both the communication cost and the varying load levels of multiple cells are to be minimized. We show that the mathematical programming model that arises can be efficiently solved in time. Furthermore, extensive numerical investigations reveal that the proposed scheme is an effective approach for both energy-efficient communication and offering significant gains in terms of load balancing in multicell topologies.

Analysis of mobile networks 4th generation and prospects for implementation in Ukraine

The article contains the analysis of the current state of 4G mobile networks in the world and the problems and prospects of their implementation in Ukraine. An extended LTE-Advanced and WiMAX-Advanced technologies functions have been researched, through which the developers could provide requirements for 4G.Due to the beginning of preparations for the introduction of 4G technologies in Ukraine it was advisable to highlight the main achievements in the development of these technologies and trends to overcoming the problems encountered in implementing them in our country.The results of research have revealed that LTE-Advanced extended functions, such as Carrier Aggregation, the introduction of 8x8 MIMO in the DL and 4x4 MIMO in the UL, support of heterogeneous (HetNet) networks, the convergence of radio access for paired and unpaired spectrum, along with retransmitting allow to increase transfer of data: up to 3Hbit/s in the downlink (DL) and to 1,5 Hbit/s in uplink (UL) and increase spectral efficiency up to 30 bps/Hz. The development of WiMAX-Advanced led to the appearance of versions of WiMAX-Advanced Release2.1 and WiMAX-Advanced Release2.2, which are aimed at convergence of WiMAX and TD-LTE: WiMAX-Advanced Release2.2 must support load balancing between LTE TDD and WiMAX, load balancing within a base station or between pico- and macro cells, it must provide Link aggregation etc.One of the main problems to the introduction of 4G technology in Ukraine is lack of necessary frequency bands occupied by 2G-communications, the need for reforming the frequency spectrum and introduction of «technological neutrality» principle. To solve these issues it is extremely important to fulfill the research of the modern technology trends and use the experience of leading telecommunication companies.

Various Dynamic Load Balancing Algorithms in Cloud Environment: A Survey

Cloud computing is emerging as a new standard model for enabling ubiquitous network access, computing resources, deploying, organizing, and accessing vast distributed computing applications over the network. In cloud computing, Load balancing is one of the main challenges which are required to distribute the workload equally across all the nodes. Load balancing uses services offered by many computer network service provider corporations. Load balancing can be different types like network load, storage capacity, memory capacity and CPU load. Load balancing helps to achieve a high user satisfaction and resource utilization ratio by confirming an efficient and fair allocation of every computing resource. Proper load balancing support in implementing failover, enabling scalability, overprovisioning, and decreases costs associated with document management systems and maximizes the availability of resources. This paper describes a survey of different dynamic load balancing algorithms in the cloud environment with their comparisons on the bases of different load balancing metrics.