Agreement Violations Research Articles

Dynamic Optimization of Energy Efficiency and Service Level Agreement violation in Cloud Data Centers via Virtual Machine Consolidation

With the escalating proliferation of cloud computing services, cloud data center providers are actively in pursuit of solutions not only to curtail energy consumption but also to efficiently cater to the unique demands of their users, thereby enhancing user satisfaction. The process of consolidating virtual machines emerges as a viable strategy for striking a balance between energy consumption and Service Level Agreement violation (SLAv), potentially surmounting this predicament faced by cloud service providers. This study aims to minimize energy consumption and SLAv to their utmost feasible extent by utilizing the First Median Threshold (FMT) technique for physical host selection and virtual machine allocation, in conjunction with the Maximum Ratio of CPU utilization to memory Utilization (MRCU) method for virtual machine selection within the framework of virtual machine consolidation. This dual-pronged approach not only translates into cost savings for cloud providers but also ensures the delivery of optimal services to users. Furthermore, a novel method has been introduced for selecting virtual machines for migration and appropriately choosing physical hosts to achieve equilibrium between two critical parameters: energy consumption and SLAv. The proposed methodology is implemented and evaluated using the CloudSim simulator, utilizing real-world PlanetLab data. The comprehensive analysis of the simulation results reveals that the proposed approach can significantly reduce energy consumption while effectively mitigating SLAv for users.

Spatial and thermal aware methods for efficient workload management in distributed data centers

Geographically distributed data centers provide facilities for users to fulfill the demand of storage and computations, where most of the operational cost is due to electricity consumption. In this study, we address the problem of energy consumption of cloud data centers and identify key characteristics of techniques proposed for reducing operational costs, carbon emissions, and financial penalties due to service level agreement (SLA) violations. By considering computer room air condition (CRAC) units that utilize outside air for cooling purposes as well as temperature and space-varying properties, we propose the energy cost model which takes into account temperature ranges for cooling purposes and operations of CRAC units. Then, we propose spatio-thermal-aware algorithms to manage workload using the variation of electricity price, locational outside and within the data center temperature, where the aim is to schedule the incoming workload requests with minimum SLA violations, cooling cost, and energy consumption. We analyzed the performance of our proposed algorithms and compared the experimental results with the benchmark algorithms for metrics of interest including SLA violations, cooling cost, and overall operations cost. Modeling, experiments, and verification conducted on CloudSim with realistic data center scenarios and workload traces show that the proposed algorithms result in reduced SLA violations, save between 15% to 75% of cooling cost and between 3.89% to 39% of the overall operational cost compared to the existing solutions.

Load balancing in cloud computing via intelligent PSO-based feedback controller

Load balancing effectively distributes network load and balances the load during the scheduling and allocation process. Hence various load balancing techniques in task scheduling and resource allocation along with VM migration has been presented previously but they have a heavy load on some VM and violate cloud service level agreement with a single point of failure. Therefore, a novel Intelligent PSO-based Feedback Controller has been proposed with regulated Scheduling, Allocation, and VM migration to perform optimal load balancing. In this proposed technique, a novel Intelligent Weighted filtering based PSO Approach is used to reduce computation time during task scheduling and resource allocation. This approach uses a multi-objective PSO algorithm with Pareto dominance to achieve high quality of service, throughput, scalability, low response time, and optimal bilateral transposed conv filtering. Moreover, during VM migration existing techniques result in service level agreement violations owing to inefficient VM placement among PMs. To overcome these issues, a Double Deep Q proximal model with a feedback controller has been proposed. The double weight set in the offline and online updating process in the decision model maintains a smooth service level agreement with the cloud. Also, centralized and decentralized controller algorithm fails with a single point of failure and coordination issue in complicated situations with instruction mixing of processes. Finally, the conditional GAN feedback controller has been used to eliminate a single point of failure with high fault tolerance, low energy consumption and migration time.

Towards virtual machine scheduling research based on multi-decision AHP method in the cloud computing platform

Virtual machine scheduling and resource allocation mechanism in the process of dynamic virtual machine consolidation is a promising access to alleviate the cloud data centers of prominent energy consumption and service level agreement violations with improvement in quality of service (QoS). In this article, we propose an efficient algorithm (AESVMP) based on the Analytic Hierarchy Process (AHP) for the virtual machine scheduling in accordance with the measure. Firstly, we take into consideration three key criteria including the host of power consumption, available resource and resource allocation balance ratio, in which the ratio can be calculated by the balance value between overall three-dimensional resource (CPU, RAM, BW) flat surface and resource allocation flat surface (when new migrated virtual machine (VM) consumed the targeted host’s resource). Then, virtual machine placement decision is determined by the application of multi-criteria decision making techniques AHP embedded with the above-mentioned three criteria. Extensive experimental results based on the CloudSim emulator using 10 PlanetLab workloads demonstrate that the proposed approach can reduce the cloud data center of number of migration, service level agreement violation (SLAV), aggregate indicators of energy comsumption (ESV) by an average of 51.76%, 67.4%, 67.6% compared with the cutting-edge method LBVMP, which validates the effectiveness.

Exploring Grammatical Gender Agreement in Russian Learners of Greek: An Eye-Tracking Study

This study investigates the acquisition of grammatical gender in Russian learners of Greek. Agreement in Determiner-Noun (Det-N) and Adjective-Noun (Adj-N) dependencies is explored through eye-tracking registration in a reading-based design. Twenty-four intermediate learners read short texts embedded with agreement violations, and then responded to a comprehension task. The study implemented a two-level triangulation by drawing its stimuli from the Greek Learner Corpus II (GLCII) and contrasting, at a first level, the findings with comparable offline data that were previously obtained from the same corpus. The second level entailed a contrast between the online evidence and the offline data that were collected through a post-reading questionnaire right after the online eye-tracking session. This questionnaire explored whether longer fixations on agreement violations are associated with explicit awareness of the study’s focus. To anticipate the outcome of the study, the gender agreement data suggests that the abstract gender feature is present in the developing grammar of Russian learners of Greek. Moreover, the participants seem to effectively deal with the syntactic computations underlying nominal agreement, though efficacy varies across the structures that have been examined. Apart from this, certain suggestions are made considering the research paradigm followed.

Design of an energy efficient dynamic virtual machine consolidation model for smart cities in urban areas

The growing smart cities in urban areas are becoming more intelligent day by day. Massive storage and high computational resources are required to provide smart services in urban areas. It can be provided through intelligence cloud computing. The establishment of large-scale cloud data centres is rapidly increasing to provide utility-based services in urban areas. Enormous energy consumption of data centres has a destructive effect on the environment. Due to the enormous energy consumption of data centres, a massive amount of greenhouse gases (GHG) are emitted into the environment. Virtual Machine (VM) consolidation can enable energy efficiency to reduce energy consumption of cloud data centres. The reduce energy consumption can increase the Service Level Agreement (SLA) violation. Therefore, in this research, an energy-efficient dynamic VM consolidation model has been proposed to reduce the energy consumption of cloud data centres and curb SLA violations. Novel algorithms have been proposed to accomplish the VM consolidation. A new status of any host called an almost overload host has been introduce, and determined by a novel algorithm based on the Naive Bayes Classifier Machine Learning (ML) model. A new algorithm based on the exponential binary search is proposed to perform the VM selection. Finally, a new Modified Power-Aware Best Fit Decreasing (MPABFD) VM allocation policy is proposed to allocate all VMs. The proposed model has been compared with certain well-known baseline algorithms. The comparison exhibits that the proposed model improves the energy consumption by 25% and SLA violation by 87%.

Efficient Auto‐scaling for Host Load Prediction through VM migration in Cloud

SummaryThe expeditious deployment of Cloud applications and services on wide‐ranging Cloud Data Centres (CDC) gives rise to the utilization of many resources. Moreover, by the increase in resource utilization, virtualization also greatly impacts achieving desired performance. The major challenges in virtualization are detecting over‐utilized or under‐utilized hosts at the right time and the proper scaling of Virtual Machines (VM) on the accurate host. Auto‐scaling in Cloud Computing allows the service providers to scale up or down the resources automatically and provides on‐demand computing power and storage capacities. Effective utilization and autonomous scaling of resources eventually reduce the load, energy consumption, and operating costs. In this paper, an efficient auto‐scaling approach for predicting host load through VM migration has been proposed. The ensemble method using different time‐series forecasting models has been proposed to forecast the approaching workload on the host. Based on this predicted load, different algorithms have been devised to detect over‐utilized and under‐utilized hosts and VMs can be migrated. The designed approach has been validated by experimentation on a real‐time Google cluster dataset. The proposed technique significantly improves average CPU utilization and reduces over‐utilization and under‐utilization. It also minimizes response time, service level agreement violations, and the slighter number of migrations and scaling overhead.

Myth or coverage: experience and expectation of policyholders from health insurance during COVID-19 in the NCR region of India

This study intends to determine whether and to what extent policyholders admitted to the hospital for COVID-19 treatment receive financial support from their health insurance. The researcher interviewed 67 health insurance claimants in the NCR region's several cities—some in Uttar Pradesh, Delhi, Haryana, and Rajasthan. They also interviewed various insurance system participants, including branch managers, agents, and hospital staff, to obtain an overview of policyholder experiences and insurer challenges. For analysis, qualitative research techniques have been used. This research is helpful in various situations when people are hospitalised for planned and unanticipated reasons by recommending what insurance companies should do to protect the interests of their customers. For the Indian insurance industry and regulator/supervisory authorities, the COVID-19 pandemic incidents can be a lesson about their strengths and weaknesses. According to this research, service-level agreement violations, procedural bottlenecks, and lack of clarity and awareness are the problems policyholders face when pursuing claims. The challenges insurers confront in serving their client's needs are maintaining operations in the face of financial stress and changing regulatory compliances, on the one hand, and helping policyholders by resolving their claims and raising awareness, on the other.

Cross-linguistic influence in the comprehension of reflexive possessive pronouns in L3 and L4 Swedish

Aims and Objectives: The study investigates the extent to which offline and online comprehension of reflexive possessive pronouns in L3 and L4 is modulated by previously acquired native and non-native grammars. Design/Methodology/Approach: Thirty-one intermediate to advanced L1 Polish learners of Swedish, who have also acquired German and English (L4 Swedish learners) or only English (L3 Swedish learners), completed a self-paced reading task and a grammaticality judgement task. The tasks targeted learners’ sensitivity to possessor and possessee agreement violations. Data and Analysis: Reading times and judgement accuracy were collected and analysed using generalised linear models. Findings/Conclusions: In online comprehension, L3/L4 Swedish learners showed sensitivity to possessor and possessee agreement, which was affected neither by their proficiency, nor by their prior native and non-native grammars. In offline comprehension, however, they had difficulty deploying possessor and possessee agreement. Performance on possessee but not possessor agreement increased with advancing proficiency, indicating the persistent influence of L1 Polish grammar. Furthermore, L4 Swedish learners achieved an overall higher judgement accuracy than L3 Swedish learners. At the same time, L4 Swedish learners read the relevant regions longer than L3 Swedish learners, suggesting a speed–accuracy trade-off for L4 Swedish learners. Originality: The study investigates cross-linguistic influence in the comprehension of reflexive possessive pronouns in multilingual learners, which has not been explored before. It is also the first to provide evidence for a multilingual advantage in the accuracy of reflexive possessive agreement, albeit at the cost of reading speed. Significance/Implications: The study demonstrates that native and non-native grammars differently affect the offline comprehension of reflexive possessive pronouns in L3 and L4. The native language seems to have a more direct effect that consists in the employment of agreement mechanisms, while the non-native language has a more indirect effect, as it can increase learners’ sensitivity to morphosyntactic violations.

Multiobjective Prioritized Workflow Scheduling in Cloud Computing Using Cuckoo Search Algorithm

Effective workflow scheduling in cloud computing is still a challenging problem as incoming workflows to cloud console having variable task processing capacities and dependencies as they will arise from various heterogeneous resources. Ineffective scheduling of workflows to virtual resources in cloud environment leads to violations in service level agreements and high energy consumption, which impacts the quality of service of cloud provider. Many existing authors developed workflow scheduling algorithms addressing operational costs and makespan, but still, there is a provision to improve the scheduling process in cloud paradigm as it is an nondeterministic polynomial-hard problem. Therefore, in this research, a task-prioritized multiobjective workflow scheduling algorithm was developed by using cuckoo search algorithm to precisely map incoming workflows onto corresponding virtual resources. Extensive simulations were carried out on workflowsim using randomly generated workflows from simulator. For evaluating the efficacy of our proposed approach, we compared our proposed scheduling algorithm with existing approaches, i.e., Max–Min, first come first serve, minimum completion time, Min–Min, resource allocation security with efficient task scheduling in cloud computing-hybrid machine learning, and Round Robin. Our proposed approach is outperformed by minimizing energy consumption by 15% and reducing service level agreement violations by 22%.

Algorithmic Approach to Virtual Machine Migration in Cloud Computing with Updated SESA Algorithm.

Cloud computing plays an important role in every IT sector. Many tech giants such as Google, Microsoft, and Facebook as deploying their data centres around the world to provide computation and storage services. The customers either submit their job directly or they take the help of the brokers for the submission of the jobs to the cloud centres. The preliminary aim is to reduce the overall power consumption which was ignored in the early days of cloud development. This was due to the performance expectations from cloud servers as they were supposed to provide all the services through their services layers IaaS, PaaS, and SaaS. As time passed and researchers came up with new terminologies and algorithmic architecture for the reduction of power consumption and sustainability, other algorithmic anarchies were also introduced, such as statistical oriented learning and bioinspired algorithms. In this paper, an indepth focus has been done on multiple approaches for migration among virtual machines and find out various issues among existing approaches. The proposed work utilizes elastic scheduling inspired by the smart elastic scheduling algorithm (SESA) to develop a more energy-efficient VM allocation and migration algorithm. The proposed work uses cosine similarity and bandwidth utilization as additional utilities to improve the current performance in terms of QoS. The proposed work is evaluated for overall power consumption and service level agreement violation (SLA-V) and is compared with related state of art techniques. A proposed algorithm is also presented in order to solve problems found during the survey.

A resource scheduling method for cloud data centers based on thermal management

With the rapid growth of cloud computing services, the high energy consumption of cloud data centers has become a critical concern of the cloud computing society. While virtual machine (VM) consolidation is often used to reduce energy consumption, excessive VM consolidation may lead to local hot spots and increase the risk of equipment failure. One possible solution to this problem is to utilize thermal-aware scheduling, but existing approaches have trouble realizing the balance between SLA and energy consumption. This paper proposes a novel method to manage cloud data center resources based on thermal management (TM-VMC), which optimizes total energy consumption and proactively prevents hot spots from a global perspective. Its VM consolidation process includes four phases where the VMs scheduler uses an improved ant colony algorithm (UACO) to find appropriate target hosts for VMs based on server temperature and utilization status obtained in real-time. Experimental results show that the TM-VMC approach can proactively avoid data center hot spots and significantly reduce energy consumption while maintaining low Service Level Agreement (SLA) violation rates compared to existing mainstream VM consolidation algorithms with workloads from real-world data centers.

Multi agent deep reinforcement learning for resource allocation in container‐based clouds environments

AbstractVirtualization enables the deployment of several virtual servers on the same physical layer, critical component of the cloud. As cloud services advance, more apps that use repositories are developed, which adds to the overburden. Containers have evolved into the most reliable and lightweight virtualization technology for cloud services thanks to their flexible sorting, mobility, and scalability. In container‐based clouds, containers can potentially cut data centre energy usage more than virtual machines (VMs) do. Containers are less energy intensive than VMs. Resource allocation is the most prevalent method in cloud systems. However, resource allocation in container‐based clouds (RAC) is innovative and complicated due to its two‐level architecture. This includes the pairing of virtual machines and physical computers with containers. In cloud container services, planner components are essential. This lowers expenses while improving the performance and variety of workloads using cloud resources. The cloud infrastructure resource allocation framework is gaining popularity since it is energy‐efficient and focuses on cloud data management to maximize income and minimize costs. In this paper, we proposed a deep learning‐based architecture capable of achieving high data centre energy efficiency and preventing Service Level Agreement (SLA) violations from deploying green cloud resources. This research describes a hybrid optimum and multi‐agent deep reinforcement learning (MADRL) technique for dynamic task scheduling (DTS) in a container cloud environment. The MADRL‐DTS model for the RAC problem considers VM overheads, VM types, and an affinity restriction. Then, to address the RAC issue, we develop a DTS hyper‐heuristic technique. MADRL‐RAC may give allocation rules by recognizing workload trends and VM types from previous workload traces. Compared to modern procedures, the results demonstrate a significant reduction in energy consumption. The evaluation for energy‐efficient resource allocation is tested in several virtualized environments to get a high power usage effectiveness and CPU usage.

Data-driven Cross-lingual Syntax: An Agreement Study with Massively Multilingual Models

Abstract Massively multilingual models such as mBERT and XLM-R are increasingly valued in Natural Language Processing research and applications, due to their ability to tackle the uneven distribution of resources available for different languages. The models’ ability to process multiple languages relying on a shared set of parameters raises the question of whether the grammatical knowledge they extracted during pre-training can be considered as a data-driven cross-lingual grammar. The present work studies the inner workings of mBERT and XLM-R in order to test the cross-lingual consistency of the individual neural units that respond to a precise syntactic phenomenon, that is, number agreement, in five languages (English, German, French, Hebrew, Russian). We found that there is a significant overlap in the latent dimensions that encode agreement across the languages we considered. This overlap is larger (a) for long- vis-à-vis short-distance agreement and (b) when considering XLM-R as compared to mBERT, and peaks in the intermediate layers of the network. We further show that a small set of syntax-sensitive neurons can capture agreement violations across languages; however, their contribution is not decisive in agreement processing.

The Role of OSCE and EU Cooperation in Resolving the Russia-Ukraine Conflict

The ongoing conflict between Russia and Ukraine has had significant consequences not only for the two states but also for the international community. The Organization for Security and Cooperation in Europe (OSCE) and the European Union (EU) have played crucial roles in resolving the conflict, with their cooperation being essential for creating a peaceful and stable outcome. This paper aims to analyze the role of OSCE and EU cooperation in resolving the Russia-Ukraine conflict and identify the challenges and areas for improvement in their coordination based on a qualitative approach. The OSCE Special Monitoring Mission to Ukraine (SMM) and the European Union Satellite Centre (SatCen) were particularly instrumental in providing unbiased monitoring and reporting of violations of the ceasefire agreement. Despite the challenges faced in coordinating their efforts, this paper presents that SMM and SatCen have proven to be valuable tools in promoting peace, security, and respect for human rights in the conflict-affected region. Thus, through this paper scholars could use it as a reference to start to study more the SatCen relevance and SMM role, to explain relations between EU and OSCE through their influential tools. The OSCE and EU have been actively engaged in conflict resolution efforts, with a focus on diplomatic and economic sanctions against Russia. In reviewing the OSCE’s cooperation with the EU, this paper suggests that they need to find other ways to minimize their limitations. The lack of political will and agreement among member states has weakened the effectiveness of the OSCE, while the EU's sanctions policy has been questioned as a means of resolving the conflict. The paper suggests that the OSCE should focus on strengthening its capacity for conflict prevention, management, and resolution as well as monitoring and reporting on violations of human rights. The EU should increase economic assistance to Ukraine and find other ways to cooperate with the OSCE to address human rights concerns and support a peaceful resolution to the conflict. Ultimately, the paper argues that a comprehensive and inclusive approach is needed to effectively address the conflict and promote sustainable peace and stability in the region.

Sensitivity to syntactic dependency formation in child second language processing: a study of numeral quantifiers in Korean

AbstractThis study investigates whether child second language (L2) learners can use syntactic information during the processing of sentences involving unbounded dependencies and how their processing patterns compare to those of child monolinguals and adult L2 learners. Through a self-paced reading experiment involving the numeral quantifier (NQ) construction in Korean, we tested participants’ sensitivity to agreement violations between a noun phrase (NP) and an NQ in local and nonlocal conditions. The results showed that a subset of child L2 learners who demonstrated target-like knowledge of NP-NQ agreement in an offline task spent a longer processing time in the NP-NQ mismatch than in the NP-NQ match condition, in both local and nonlocal contexts. These child L2 learners’ processing patterns were comparable to those observed in child monolinguals and adult L2 learners. These findings suggest that child and adult L2 learners rely on the same system of syntactic representations and processing mechanisms that guide first language processing.

Is beta in agreement with the relatives? Using relative clause sentences to investigate MEG beta power dynamics during sentence comprehension.

There remains some debate about whether beta power effects observed during sentence comprehension reflect ongoing syntactic unification operations (beta-syntax hypothesis), or instead reflect maintenance or updating of the sentence-level representation (beta-maintenance hypothesis). In this study, we used magnetoencephalography to investigate beta power neural dynamics while participants read relative clause sentences that were initially ambiguous between a subject- or an object-relative reading. An additional condition included a grammatical violation at the disambiguation point in the relative clause sentences. The beta-maintenance hypothesis predicts a decrease in beta power at the disambiguation point for unexpected (and less preferred) object-relative clause sentences and grammatical violations, as both signal a need to update the sentence-level representation. While the beta-syntax hypothesis also predicts a beta power decrease for grammatical violations due to a disruption of syntactic unification operations, it instead predicts an increase in beta power for the object-relative clause condition because syntactic unification at the point of disambiguation becomes more demanding. We observed decreased beta power for both the agreement violation and object-relative clause conditions in typical left hemisphere language regions, which provides compelling support for the beta-maintenance hypothesis. Mid-frontal theta power effects were also present for grammatical violations and object-relative clause sentences, suggesting that violations and unexpected sentence interpretations are registered as conflicts by the brain's domain-general error detection system.

Resource optimization using predictive virtual machine consolidation approach in cloud environment

The Proliferation of on-demand usage-based IT services, as well as the diverse range of cloud users, have led to the establishment of energy-hungry hefty cloud data centers. Therefore, cloud service providers are striving to reduce energy consumption for cost-saving and environmental sustainability issues of data centers. In this direction, Virtual Machine (VM) consolidated is a widely used approach to optimize hardware resources at cost of performance degradation due to unnecessary migrations. Hence, the motivation of the proposed approach is to minimize energy consumption while maintaining the performance of cloud data centers. This leads to a reduction in the overall cost and an increase in the reliability of cloud service providers. To achieve this goal Predictive Virtual Machine Consolidation (PVMC) algorithm is proposed using the exponential smoothing moving average (ESMA) method. In the proposed algorithm, the ratio of deviation to utilization is calculated for VM selection and placement. migrating the high CPU using VMs or we can restrict steady resource-consuming VMs from migration. The outcomes of the proposed algorithm are validated on computer-based simulation under a dynamic workload and a variable number of VMs (1–290). The experimental results show an improvement in the mean threshing index (40%, 45%) and instruction energy ratio (15%, 17%) over the existing policies. Hence, the proposed algorithm could be used in real-world data centers for reducing energy consumption while maintaining low service level agreement violations.

An energy-efficient load balance strategy based on virtual machine consolidation in cloud environment

In the cloud computing environment, unbalanced utilization of multi-dimensional resources in physical servers generates resource fragmentation, leading to inefficient resource utilization and energy wastage in data centers. The inefficient utilization of resources makes high energy consumption and low quality of service (QoS) in resource management an urgent problem to be solved. We propose a load balancing strategy based on virtual machine consolidation (LBVMC), which aims to reduce the energy consumption and service level agreement (SLA) violation of data centers by balancing the multi-dimensional resource utilization in physical machines (PMs). First, we present a load state classification algorithm for PMs with load abnormality considering current and future loads to reduce unnecessary virtual machine (VM) migrations caused by occasional load fluctuations. Then, we propose a resource-weight based selection model for migratable VMs, which selects appropriate VMs to be migrated based on multi-dimensional resource utilization, and reduces resource fragmentation caused by load imbalance. Finally, we design a VM placement algorithm based on resource fitness and load correlation to deploy VMs on the optimal destination PMs to assure load balancing of the destination PMs after VM placement. We perform simulated experiments under homogeneous, heterogeneous, and bottleneck resource environments. Experimental results show that LBVMC is superior to other strategies in reducing energy consumption and SLA violations and achieves better overall performance.

Learning to Dynamically Select Cost Optimal Schedulers in Cloud Computing Environments

The operational cost of a cloud computing platform is one of the most significant Quality of Service (QoS) criteria for schedulers, crucial to keep up with the growing computational demands. Several data-driven deep neural network (DNN)-based schedulers have been proposed in recent years that outperform alternative approaches by providing scalable and effective resource management for dynamic workloads. However, state-of-the-art schedulers rely on advanced DNNs with high computational requirements, implying high scheduling costs. In non-stationary contexts, the most sophisticated schedulers may not always be required, and it may be sufficient to rely on low-cost schedulers to temporarily save operational costs. In this work, we propose MetaNet, a surrogate model that predicts the operational costs and scheduling overheads of a large number of DNNbased schedulers and chooses one on-the-fly to jointly optimize job scheduling and execution costs. This facilitates improvements in execution costs, energy usage and service level agreement violations of up to 11%, 43% and 13% compared to the state-of-the-art methods.