Cloud Computing provides an effective platform for executing large-scale and complex workflow applications with a pay-as-you-go model. Nevertheless, various challenges, especially its optimal scheduling for multiple conflicting objectives, are yet to be addressed properly. The existing multi-objective workflow scheduling approaches are still limited in many ways, e.g., encoding is restricted by prior experts' knowledge when handling a dynamic real-time problem, which strongly influences the performance of scheduling. In this paper, we apply a deep-Q-network model in a multi-agent reinforcement learning setting to guide the scheduling of multi-workflows over infrastructure-as-a-service clouds. To optimize multi-workflow completion time and user's cost, we consider a Markov game model, which takes the number of workflow applications and heterogeneous virtual machines as state input and the maximum completion time and cost as rewards. The game model is capable of seeking for correlated equilibrium between make-span and cost criteria without prior experts' knowledge and converges to the correlated equilibrium policy in a dynamic real-time environment. To validate our proposed approach, we conduct extensive case studies based on multiple well-known scientific workflow templates and Amazon EC2 cloud. The experimental results clearly suggest that our proposed approach outperforms traditional ones, e.g., non-dominated sorting genetic algorithm-II, multi-objective particle swarm optimization, and game-theoretic-based greedy algorithms, in terms of optimality of scheduling plans generated. INDEX TERMS Multi-objective workflow scheduling, deep-Q-network (DQN), multi-agent reinforcement learning (MARL), infrastructure-as-a-service (IaaS) cloud, quality-of-service (QoS).

No abstract

Service composition is a technology capable of combing a collection of existing services where many smaller services are coordinated together to form a larger one. Functionally similar services can often show different quality-of-service (QoS) properties. For a specific service composition request, how to choose from a bag of suitable services that fulfill the required functions under given quality-ofservice constraints is widely believed to be a great challenge. The traditional approach usually tackles this problem by assuming fixed, bounded, or statistic QoS and views the decision-making of service composition as a static process. Instead, we address this problem by considering time-varying and fluctuating QoS and presenting a predictive-trend-aware service composition method by using a time series prediction model and genetic algorithms. We conduct extensive case studies based on multiple randomly-generated service templates with varying process configurations and show that our method outperforms existing ones. INDEX TERMS Time series, Quality-of-Service (QoS), service composition, genetic algorithm (GA).

Virtual Machine (VM) consolidation technique plays an important role in energy management and load-balancing of cloud computing systems. Dynamic VM consolidation is a promising consolidation approach in this direction, which aims at using least active physical machines (PMs) through appropriately migrating VMs to reduce resource consumption. The resulting optimization problem is well-acknowledged to be NP-hard optimization problems. In this paper, we propose a novel merge-and-split-based coalitional game-theoretic approach for VM consolidation in heterogeneous clouds. The proposed approach first partitions PMs into different groups based on their workload levels, then employs a coalitional-game-based VM consolidation algorithm (CGMS) in choosing members from such groups to form effective coalitions, performs VM migrations among the coalition members to maximize the payoff of every coalition, and finally keeps PMs running in a high energy-efficiency state. The simulation results based on three scenarios clearly suggest that our proposed approach outperforms traditional ones in terms of energy-saving, and also achieve a fair level of load balance.

No abstract

No abstract

No abstract

No abstract