The biggest advantage of employing virtualization is the ability to flexibly remap physical resources to virtual servers in order to handle the resource redistribution. So virtual machine is the fundamental unit in cloud data center. However, the load of virtual machine constantly changes owing to the needs of applications. In order to improve the resource utilization and reduce power energy, data center needs an automatic, quick and dynamic resource scheduling strategy which treats virtual machine as a scheduling unit to balance load and consolidate servers.In this paper, we present a two-steps dynamic resource scheduling strategy, named Smart-DRS, which fits cloud data center well and strikes a balance between efficiency, cost and instantaneity. Firstly, we employ a prediction technique based on Single Exponential Smoothing algorithm. Then a novel and efficient migration algorithm based on Vector Projection was applied.For evaluating the performance of Smart-DRS, we develop a complete resource management prototype system in which resource scheduling is just only a module. Then we build a cluster with 32 physical machines running with 3200 virtual machines to simulate datacenter environment. Experiment results tell us that Smart-DRS has a high forecast accuracy and also can deal well with load balancing and load consolidation.
In this paper, we consider a new variation of the classical online scheduling problem. In our model, an online scheduler is allowed to migrate the assigned jobs to different machines. Live migration is a powerful tool for load balancing. However, migration will incur additional cost in the destination machines. In this paper, we study the scheduling problem with migration cost model. Suppose that a job with processing time p which is already scheduled on the machine A is removed and transferred to the machine B, the load of the machine A will decrease p, but the load of the machine B will increase (1 + r)p, where 0 ≤ r ≤ 1 is a constant and it is called the migration factor.First, we propose an approximation algorithm for arbitrary machines. Then we give an improved algorithm for the case of two machines. Both algorithms are better than list scheduling algorithm [2] if the migration factor is smaller than a certain value. Finally, we implement our algorithms both in real data and random data. The experimental results indicate that the performances of algorithms are very close to the optimal algorithm.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2025 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.