Application Placement Research Articles

Adaptive data placement in the Fog infrastructure of IoT applications with dynamic changes

Time-sensitive Internet of Things (IoT) applications need to provide instant responses to end users. Seeking to reduce the latency of IoT applications is challenging task. Due to the dynamic changes that can occur over time such as failure of IoT devices or integration of new services and devices, the IoT application constantly requires new storage resources. To ensure the high performance of an IoT application, it is common today to place its produced data in the Fog infrastructure while satisfying its requirements over time. The main challenge in this regard is to ensure optimal data placement in this heterogeneous and distributed infrastructure with low latency. In fact, with continuous changes, new data assignments to target storage nodes can lead to data migration and increased latency. In this paper, we propose an adaptive approach to place the IoT data with multiple replicas in the Fog infrastructure while estimating and optimizing the overall latency. To cope with the IoT application changes, we establish a ℓ-backtracking greedy algorithm called adaptive multiple data replicas placement of IoT applications with external event changes in the Fog infrastructure (iFogDPE). It is a stable many-to-many matching algorithm that minimizes the write, read and migration latency using a few number of data replicas. Then, it assigns each data consumer with its appropriate data replica. To model and to simulate the IoT application and Fog infrastructure enactment as well as to test and to evaluate our solution we used iFogSim simulator. Our experimental results highlight the efficiency of our approach in reducing the storage and access latency with fewer data replicas while maintaining low data migration latency over time.

Heterogeneous reconfigurable architectures for machine learning dataflows

AbstractThis work explores the placement and routing of machine learning applications' dataflow graphs on different heterogeneous coarse‐grained reconfigurable architectures (CGRA). We analyze three different types of processing element (PE) heterogeneity, the first concerning the interconnection pattern, the second being on the kind of operations a single PE can execute, and the last concerning the PE buffer resources. This analysis aim to propose a fair reduction to the overall cost in comparison to the homogeneous CGRA architecture. We compare our results with the homogeneous case and one of the state‐of‐the‐art tools for placement and routing (P&R). Our algorithm executed, on average, 52% faster than VPR 8.1 (Versatile Place and Route), which is an open‐source academic tool designed for the FPGA placement and routing phases, reaching better mapping in 66% of cases and achieving the same results in 26% of cases. Furthermore, a heterogeneous architecture reduces the cost without losing performance in 76% of the cases considering multiplier heterogeneity. We propose a novel heterogeneous buffer architecture that minimizes the buffer resources by 56.3% for K‐means dataflow patterns. We also show that a heterogeneous border chess architecture outperforms a homogeneous one. In addition, our mapping reaches optimal instances of single tree dataflows compared to classical Lee/Choi and H‐trees.

Evaluation of fog application placement algorithms: a survey

Recently, the concept of cloud computing has been extended towards the network edge. Devices near the network edge, called fog nodes, offer computing capabilities with low latency to nearby end devices. In the resulting fog computing paradigm (also called edge computing), application components can be deployed to a distributed infrastructure, comprising both cloud data centers and fog nodes. The decision which infrastructure nodes should host which application components has a large impact on important system parameters like performance and energy consumption. Several algorithms have been proposed to find a good placement of applications on a fog infrastructure. In most cases, the proposed algorithms were evaluated experimentally by the respective authors. In the absence of a theoretical analysis, a thorough and systematic empirical evaluation is of key importance for being able to make sound conclusions about the suitability of the algorithms. The aim of this paper is to survey how application placement algorithms for fog computing are evaluated in the literature. In particular, we identify good and bad practices that should be utilized respectively avoided when evaluating such algorithms.

Monte Carlo simulation of tilted contact plaque brachytherapy placement for juxtapapillary retinoblastoma

BackgroundThe 106-Ruthenium contact plaque applicator is utilized for the treatment of intraocular tumor within a thickness of less than 6 mm. If anything obstructs the placement of the plaque applicator, the treatment is generally difficult because the applicator has to be temporarily located just on the opposite side of the retinal tumor. Furthermore, the plaque applicator edge of approximately 1 mm does not contain 106Ru, estimating the delivered radiation dose for eccentric tumor is challenging because the lateral dose profile is inadequately provided by the manufacture’s certification. This study aims to simulate tumor coverage of the tilted applicator placement for treating an infant with juxtapapillary retinoblastoma and to achieve the effective treatment.Case presentationWe present an infant with retinoblastoma whose tumor involved macular and was invading just temporal side of the optic disc. Additionally, posterior staphyloma was induced by a series of previous treatments, making it more difficult to treat the standard plaque placement. Thus, the applicator type of CCA was intentionally tilted to the eyeball and the distance between the posterior edge of the applicator and the eyeball had to be then equal to or more than 2 mm based on the dose distribution of the applicator calculated using Monte Carlo simulation to minimize damage to surrounding tissues while covering the tumor. It was then comparable to the certification and previous reports. Based on the acquired dose distribution, the optimal placement of the applicator was derived from varying the distance between the applicator’s edge and the eyeball, and the distance was then determined to be 2 mm. In this case, the minimum dose rate in the tumor was 25.5 mGy/min, and the time required to deliver the prescribed dose was 26.2 h. Therefore, the tilted 106Ru plaque applicator placement could deliver the required dose for the treatment. The physical examination revealed no active tumor as a result of the treatment.ConclusionsOptimizing the placement of the 106Ru plaque applicator, it was possible to guarantee that the prescribed dose will be delivered to the tumor even if the standard placement is not possible for the juxtapapillary tumor.

QoS-aware placement of microservices-based IoT applications in Fog computing environments

The Fog computing paradigm, offering cloud-like services at the edge of the network, has become a feasible model to support computing and storage capabilities required by latency-sensitive and bandwidth-hungry Internet of Things (IoT) applications. As fog devices are distributed, heterogeneous and resource-constrained, efficient application scheduling mechanisms are required to harvest the full potential of such computing environments. Due to the rapid evolution in IoT ecosystems and also to better suit fog environment characteristics, IoT application development has moved from the monolithic architecture towards the microservices architecture that enhances scalability, maintainability and extensibility of the applications. This architecture improves the granularity of service decomposition, thus providing scope for improvement in QoS-aware placement policies. Existing application placement policies lack proper utilisation of these features of microservices architecture, thus failing to produce efficient placements. In this paper, we harvest the characteristics of microservice architecture to propose a scalable QoS-aware application scheduling policy for batch placement of microservices-based IoT applications within fog environments. Our proposed policy, QoS-aware Multi-objective Set-based Particle Swarm Optimisation (QMPSO), aims at maximising the satisfaction of multiple QoS parameters (makespan, budget and throughput) while focusing on the utilisation of limited fog resources. Besides, QMPSO adapts and improves the Set-based Comprehensive Learning Particle Swarm Optimisation (S-CLPSO) algorithm to achieve better convergence in the fog application placement problem. We evaluate our policy in a simulated fog environment. The results show that compared to the state-of-the-art solutions, our placement algorithm significantly improves QoS in terms of makespan satisfaction (up to 35% improvement) and budget satisfaction (up to 70% improvement) and ensures optimum usage of computing and network resources, thus providing a robust approach for QoS-aware placement of microservices-based heterogeneous applications within fog environments.

TRAPPY: a truthfulness and reliability aware application placement policy in fog computing

TRAPPY: a truthfulness and reliability aware application placement policy in fog computing

Crane Cloud: A resilient multi-cloud service abstraction layer for resource-constrained settings

Developers and users situated in low-resource settings are faced with unique contextual and infrastructure challenges when accessing and consuming cloud-based services. In low-resource settings, access to cloud services and platforms is usually characterized by low-end computing devices and often unreliable and slow mobile broadband Internet connections. In this paper, we discuss key challenges for developing for and accessing cloud services in resource constrained settings, namely, (1) Frequent Internet partitions and bandwidth constraints, (2) Data jurisdiction restrictions, (3) Vendor lock-in, and (4) Poor quality of service. Inspired by these challenges, we propose a set of important design considerations and properties for a resilient multi-cloud service layer, that includes: (1) Containerization and orchestration of applications, (2) Application placement and replication, (3) Portability and multi-cloud migration, (4) Resilience to network partitions and bandwidth constraints, (5) Automated service discovery and load balancing, (6) Localized image registry, and (7) Support for platform monitoring and management. We present an implementation and validation case study, Crane Cloud, an open source multi-cloud service abstraction layer built on-top of Kubernetes that is designed with inherent support for resilience to network partitions, microservice orchestration (deployment, scaling and management of containerized applications), a localized image registry, support for migration of services between private and public clouds to avoid vendor lock-in issues and platform monitoring. We evaluate the performance and user experience of Crane Cloud by implementing and deploying a computational and bandwidth intensive machine learning system. The results show lower response times of the system on Crane Cloud compared with hosting on other public clouds. The Crane Cloud platform is serving as a cloud-service for students and developers in low-resource settings and also as an education platform for cloud computing.

Evaluation of dose–Volume-based image-guided high-dose-rate brachytherapy in carcinoma uterine cervix: A prospective study

Background: In recent few decades, the evolution in imaging technology, especially computed tomography (CT) and magnetic resonance imaging, results in widespread availability and its use in high-dose-rate (HDR) intracavitary brachytherapy (ICBT) applications. Aim: The present study was aimed to analyze the cumulative dose–volume histogram of the tumor and organs at risk (OARs) in three-dimensional (3D) CT image-based brachytherapy planning and clinical outcomes of the treated patients. Materials and Methods: This prospective observational study included 40 patients with carcinoma cervix. After external beam radiotherapy (EBRT), a dose of 6 Gy per fraction of HDR ICBT in four fractions with a total dose to point “A” approximately 80–85 Gy was given. For planning, the tumor volumes (high-risk clinical target volume [HR-CTV]) and volume of OARs (bladder, rectum, and sigmoid colon) were contoured on each CT slice. The dose–volume parameters, i.e., minimum dose received to 90% and 100% by HR-CTV volume (D90 and D100) for target and the maximum dose received by minimum volume of 2CC (D2CC) for OARs, were calculated and assessed for clinical response in patients. Results: The mean D2CC dose was 18.24 ± 0.93 Gy, 16.44 ± 1.11 Gy, and 16.37 ± 0.67 Gy for bladder, rectum, and sigmoid colon, respectively. The combined (EBRT and HDR ICBT) mean equieffective dose in 2 Gy per fraction (EQD2) dose for bladder was 76.71 ± 2.05 Gy, for rectum was 72.82 ± 2.58 Gy, and for sigmoid colon was 72.71 ± 1.41 Gy, and its comparison with baseline values showing P < 0.01 for bladder, rectum, and sigmoid colon was considered statistically significant. The mean EQD2 dose of HR-CTV D90 was 151 ± 27.3 Gy. Patients who had received HR-CTV D90 of >90 Gy compared with <90 Gy had exceptionally better local control and complete response. Conclusion: The present study suggested that CT is a favorable modality for treatment planning in cervical cancer with limited resources setup in terms of improved tumor coverage, lesser toxicity, confirmation of applicator placement, and accounting dose to OARs.

A fuzzy approach for optimal placement of IoT applications in fog-cloud computing

A fuzzy approach for optimal placement of IoT applications in fog-cloud computing

Splitting and placement of data-intensive applications with machine learning for power system in cloud computing

Aiming to meet the growing demand for observation and analysis in power systems that based on Internet of Things (IoT), machine learning technology has been adopted to deal with the data-intensive power electronics applications in IoT. By feeding previous power electronic data into the learning model, accurate information is drawn, and the quality of IoT-based power services is improved. Generally, the data-intensive electronic applications with machine learning are split into numerous data/control constrained tasks by workflow technology. The efficient execution of this data-intensive Power Workflow (PW) needs massive computing resources, which are available in the cloud infrastructure. Nevertheless, the execution efficiency of PW decreases due to inappropriate sub-task and data placement. In addition, the power consumption explodes due to massive data acquisition. To address these challenges, a PW placement method named PWP is devised. Specifically, the Non-dominated Sorting Differential Evolution (NSDE) is used to generate placement strategies. The simulation experiments show that PWP achieves the best trade-off among data acquisition time, power consumption, load distribution and privacy preservation, confirming that PWP is effective for the placement problem.

An improved multi-objective genetic algorithm with heuristic initialization for service placement and load distribution in edge computing

Edge Computing (EC) is a promising concept to overcome some obstacles of traditional cloud data centers to support Internet of Things (IoT) applications, especially time-sensitive applications. However, EC faces some challenges, including the resource allocation for heterogeneous applications at a network edge composed of distributed and resource-restricted nodes. A relevant issue that needs to be addressed by a resource manager is the service placement problem, which is the decision-making process of determining where to place different services (or applications). A related issue of service placement is how to distribute workloads of an application placed on multiple locations. Hence, we jointly investigate the load distribution and placement of IoT applications to minimize Service Level Agreement (SLA) violations due to the limitations of EC resources and other conflicting objectives. In order to handle the computational complexity of the formulated problem, we propose a multi-objective genetic algorithm with the initial population based on random and heuristic solutions to obtain near-optimal solutions. Evaluation results show that our proposal outperforms other benchmark algorithms in terms of response deadline violation, as well as terms of other conflicting objectives, such as operational cost and service availability.

NFV-Enabled IoT Service Provisioning in Mobile Edge Clouds

Conventional Internet of Things (IoT) applications involve data capture from various sensors in environments, and the captured data then is processed in remote clouds. However, some critical IoT applications (e.g., autonomous vehicles) require a much lower response latency and more secure guarantees than those offered by remote clouds today. Mobile edge clouds (MEC) supported by the network function virtualization (NFV) technique have been envisioned as an ideal platform for supporting such IoT applications. Specifically, MECs enable to handle IoT applications in edge networks to shorten network latency, and NFV enables agile and low-cost network functions to run in low-cost commodity servers as virtual machines (VMs). One fundamental problem for the provisioning of IoT applications in an NFV-enabled MEC is where to place virtualized network functions (VNFs) for IoT applications in the MEC, such that the operational cost of provisioning IoT applications is minimized. In this paper, we first address this fundamental problem, by considering a special case of the IoT application placement problem, where the IoT application and VNFs of each service request are consolidated into a single location (gateway or cloudlet), for which we propose an exact solution and an approximation algorithm with a provable approximation ratio. We then develop a heuristic algorithm that controls the resource violation ratios of edge clouds in the network. For the IoT application placement problem for IoT applications where their VNFs can be placed to multiple locations, we propose an efficient heuristic that jointly places the IoT application and its VNFs. We finally study the performance of the proposed algorithms by simulations and implementations in a real test-bed, Experimental results show that the performance of the proposed algorithms outperform their counterparts by at least 10 percent.

PLEN01 Presentation Time: 1:15 PM: Virtual Reality-Based Simulation Improves Gynecologic Brachytherapy Proficiency, Engagement, and Trainee Self-Confidence

Purpose: Brachytherapy is an integral component in treatment of gynecologic malignancies. Intracavitary brachytherapy has the highest rates of local control and survival in cervical cancer. Classical brachytherapy training requires 1-on-1 instruction during patient encounters. This training is resource intensive and variable based on the type/frequency of the procedures performed by the attending physician/facility. Consequently, exposure to brachytherapy varies among residency programs. Even with Accreditation Council for Graduate Medical Education mandated requirements, only about 50% of graduating residents express confidence to develop a brachytherapy practice with caseload as the greatest barrier. The use of simulation-based education has been shown to increase confidence, applicator placement quality, and decrease implant time and errors.

Ultrasound-guided Brachytherapy for Cervix Cancer

Radiotherapy and brachytherapy are the definitive treatments for locally advanced cervix cancer. The use of soft-tissue imaging, particularly magnetic resonance imaging, has enhanced their effectiveness and improved clinical outcomes. However, the use of magnetic resonance imaging is largely restricted to well-resourced centres in both the first and developing world and remains elusive to many less advantaged centres, particularly those in areas with a high burden of cervix cancer. Ultrasound is an accessible, affordable and accurate imaging modality that can be used throughout the brachytherapy procedure. Ultrasound is primarily used to ensure safe insertion of the applicator but can also be used to guide planning. The methods used to utilise ultrasound images for planning are described. Ultrasound is particularly useful as a verification aid to confirm applicator placement after patients are moved and transferred around the radiotherapy department. It can also be used to verify the dimensions of treatment volumes over the course of brachytherapy. There is a crucial unmet need for an accessible economical soft-tissue imaging modality in cervical brachytherapy. Ultrasound has the potential to meet this need.

FONS: a fog orchestrator node selection model to improve application placement in fog computing

Fog computing not only executes applications in the vicinity of the IoT devices/users but also keeps transient data which removes the need for data transfer to the cloud on regular basis. For applications’ placement, an orchestrator considers the requirement of the application and the current fog status to place the application on suitable fog nodes. The orchestrator itself may be placed on fog or cloud nodes. In a centralized approach, a fixed entity that works as an orchestrator is prone to single point failure, less mobility support, etc. Therefore, the literature advises for a decentralized approach in which a nearby fog node is selected to act as an orchestrator. Poor selection of Fog Orchestrator Nodes (FONs) may result in the performance degradation of the system. None of the earlier work proposed the selection of FON for the placement of the applications on the fog nodes. Towards this, a brief but latest survey of the works, to understand the FON selection problem, is done along with the importance of decentralized approach in the FON selection problem. Further, few performance metrics have been identified which helps in the FON selection. A lightweight FON Selection model (FONS) is proposed so that even the least powerful IoT devices can select a FON. The proposed work is validated by incorporating the FON selection algorithm in one state of art to observe a remarkable improvement in application placement.

A Cost-Aware Management Framework for Placement of Data-Intensive Applications on Federated Cloud

Modern big data analysis and business intelligence applications come with high resource demands, along with a requirement for large data transfer between storage and compute nodes. Since the available resources of a single data center might not be sufficient to host these applications, federated cloud systems present a promising solution. The objective of each cloud service provider in a federation is to maximize its own profit and to minimize its total operating cost. To achieve these objectives, spatial variation in the energy cost and bandwidth cost could be leveraged while allocating the workload. We propose a hierarchical approach for resource management in a federated cloud, catering to the requirements of each provider in the federation. We formulate the placement of a data-intensive applications as an optimization problem to minimize the total operating cost, including the energy and communication costs. We propose an algorithm to allocate the virtual components of a data-intensive application, in two phases; partitioning and mapping. While partitioning creates clusters of correlated nodes, mapping allocates the created clusters to data centers that minimize the total cost. Through extensive experiments in different scenarios, we demonstrate that the proposed algorithm achieves a significant reduction in the total operating cost.

Bi-objective optimization of application placement in fog computing environments

Fog computing has been recently introduced to complement the cloud computing paradigm and offer application services at the edge of the network. The heterogeneity of fog computational nodes makes application placement in fog infrastructures a challenging task that requires proper management in order to satisfy application requirements. This paper proposes a bi-objective application placement algorithm for fog computing environments. The proposed algorithm seeks to optimally place application modules on the underlying fog devices considering applications criticality levels and security requirements. The placement problem has been formulated as a bi-objective knapsack problem and solved using the non-dominated sorting genetic algorithm II (NSGA-II). It has been implemented using a specialized fog computing simulation tool and compared against existing placement algorithms. Simulation results demonstrate the ability of the proposed algorithm to optimize application placement in fog computing environments in terms of application performance, power efficiency and security satisfaction rates.

Influence of biodegradable lubricant on the ecology, lubrication, and machining process in minimum quantity lubrication

In minimum quantity lubrication (MQL) processing, lubricating oil is passed through the air flow field in the form of oil droplets. The oil droplets penetrate the field and are adsorbed by it to form a film in the capillary tube at the workpiece–tool interface. Therefore, the type, atomization, penetration, and adsorption of the lubricating oil have crucial influences on the lubrication quality. Three types of lubricant—trimethylolpropane trioleate, vegetable oil, and polyethylene glycol—were used in this research. To protect the environment and meet application requirements, the biodegradability, oxidation stability, and placement stability of the aforementioned three oils were tested in accordance with the relevant national standard. The biodegradation rate and storage stability of vegetable oil 6000 were high; however, its oxidation stability was poor. A three-dimensional particle dynamic analyzer was used to measure the atomization of oil droplets, and the amount of oil penetrating the tool–workpiece contact area was quantitatively evaluated using a specially designed penetration sampling device. The contact angles of the droplets on AISI 1045 carbon steel and TiN coating were measured, and experiments were performed on the MQL machining of AISI 1045 carbon steel. According to the experimental results, the main cutting force, coefficient of friction, surface morphology, and roughness obtained using vegetable oil 6000 were superior to those obtained using polyethylene glycol 400 and trimethylolpropane trioleate, because a more focused oil drop distribution, higher amount of penetrated oil, and higher wettability were obtained with vegetable oil 6000.

A priority, power and traffic-aware virtual machine placement of IoT applications in cloud data centers

Recent telecommunication paradigms, such as big data, Internet of Things (IoT), ubiquitous edge computing (UEC), and machine learning, are encountering with a tremendous number of complex applications that require different priorities and resource demands. These applications usually consist of a set of virtual machines (VMs) with some predefined traffic load between them. The efficiency of a cloud data center (CDC) as prominent component in UEC significantly depends on the efficiency of its VM placement algorithm applied. However, VM placement is an NP-hard problem and thus there exist practically no optimal solution for this problem. In this paper, motivated by this, we propose a priority, power and traffic-aware approach for efficiently solving the VM placement problem in a CDC. Our approach aims to jointly minimize power consumption, network consumption and resource wastage in a multi-dimensional and heterogeneous CDC. To evaluate the performance of the proposed method, we compared it to the state-of-the-art on a fat-tree topology under various experiments. Results demonstrate that the proposed method is capable of reducing the total network consumption up to 29%, the consumption of power up to 18%, and the wastage of resources up to 68%, compared to the second-best results.

DECA: A Dynamic Energy Cost and Carbon Emission-Efficient Application Placement Method for Edge Clouds

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