DYNAMIC RESOURCE PROVISIONING FOR ENERGY EFFICIENT CLOUD RADIO ACCESS NETWORKS

 

ABSTRACT

Energy saving is critical for the cloud radio access networks (C-RANs), which are composed by massive radio access units(RAUs) and energy-intensive computing units (CUs) that host numerous virtual machines (VMs). We attempt to minimize the energyconsumption of C-RANs, by leveraging the RAU sleep scheduling and VM consolidation strategies. We formulate the energy savingproblem in C-RANs as a joint resource provisioning (JRP) problem of the RAUs and CUs. Since the active RAU selection is coupledwith the VM consolidation, the JRP problem shares some similarities with a special bin-packing problem. In this problem, the number ofitems and the sizes of items are correlated and are both adjustable. No existing method can be used to solve this problem directly.Therefore, we propose an efficient low-complexity algorithm along with a context-aware strategy to dynamically select active RAUs andconsolidate VMs to CUs. In this way, we can significantly reduce the energy consumption of C-RANs, while do not incur too muchoverhead due to VM migrations. Our proposed scheme is practical for a large-size network, and its effectiveness is demonstrated bythe simulation results.

EXISTING SYSTEM:

The concept of C-RAN was first proposed in  and furtherelaborated in. After that, an increasing number ofresearch has been carried out in both industry and academia. These works have demonstrated that theC-RAN can enable energy efficient network operations.In a C-RAN, the main network resources are the computationresources in CU pool and the spectrum resourcesin RAUs. To fully explore the potentials of C-RAN in reducingenergy consumption of the mobile networks, somestrategies were proposed to achieve right-sizing of networkresources, according to the dynamic change of network load.For the CU pool, the authors of clustered the neighbouringRAUs along with their corresponding VMs firstly.Then a coarse-grained approach was used to resize the VMs,based on the network traffic patterns at different time of aday. In, the computation resource allocation problemin C-RANs was formulated as a bin-packing problem, anda heuristic simulated annealing algorithm was provided toreduce power consumption of the CU pool. In considerationof the capacity constraints of fronthaul links, the authors of formulated the virtualized resource allocation problemas an integer linear programming problem, and presented aheuristic algorithm to minimize the energy consumption ofCUs. The authors of aimed to minimize the number ofCUs that are required to serve the UEs, with the considerationof multi-resource allocation in the CUs. However, theseworks did not consider the energy saving in RAUs.

PROPOSED SYSTEM:

The main contributions of this paper are summarized asfollows:1) We intend to optimize the energy cost of C-RANsin a global manner. We formulate the JRP problem in CRANs,by jointly considering RAU sleep scheduling andVM consolidation. The VM consolidation problem is furtherformulated as a special bin-packing problem. In this problem,the number of items and the sizes of items are coupled,and both of them can be adjusted for packing. This makethe JRP problem be different from other problems that areconsidered in existing works.2) The JRP problem involves not only the on-off problemsof VMs in the CU pool but also the on-off problemsof RAUs. The on-off states of RAUs are correlated with theon-off states of VMs, which makes the JRP problem difficultto be solved. We propose an efficient low-complexity algorithmto solve the JRP problem, by dynamically selectingthe active RAUs and consolidating the VMs to CUs. Thesimulation results demonstrate that the performance of thisalgorithm is close to that of a high-complexity exhaustivesearching algorithm. In addition, our proposed algorithm ismore practical for the time-critical reconfigurations of RAUsand CUs in a real network.3) Since the VM migrations that are induced by the VMconsolidation may degrade the quality of network services,we propose a context-aware strategy to make the trade-offbetween saving energy and reducing the number of VMmigrations. When the UEs’ distributions and required datarates change over time and space, our proposed methodcan significantly reduce the number of VM migrations withonly a slight increase of the network energy consumption.In this way, the objective of minimizing energy consumptionof C-RANs can be achieved without sacrificing the networkperformance.the active RAUs and consolidating the VMs to CUs. Thesimulation results demonstrate that the performance of thisalgorithm is close to that of a high-complexity exhaustivesearching algorithm. In addition, our proposed algorithm ismore practical for the time-critical reconfigurations of RAUsand CUs in a real network.3) Since the VM migrations that are induced by the VMconsolidation may degrade the quality of network services,we propose a context-aware strategy to make the trade-offbetween saving energy and reducing the number of VMmigrations. When the UEs’ distributions and required datarates change over time and space, our proposed methodcan significantly reduce the number of VM migrations withonly a slight increase of the network energy consumption.In this way, the objective of minimizing energy consumptionof C-RANs can be achieved without sacrificing the networkperformance

CONCLUSION

In this paper, we studied a minimum energy consumptionproblem of C-RANs, where the RAUs and CUs can beswitched off to save energy. We formulated this problemas a joint resource provisioning problem of the RAUs andthe CUs. To minimize the energy consumption of C-RANs,we proposed a dynamic resource provisioning algorithm toselect the active RAUs and consolidate the VMs to CUs. Toensure the quality of service for the C-RANs, we proposeda context-aware strategy to reduce the number of VM migrations.In this way, we can achieve the objective of minimizingenergy consumption of C-RANs without sacrificingthe quality of networks services. Simulation results havedemonstrated the effectiveness of our proposed scheme

REFERENCES

[1] A. Checko, H. L. Christiansen, Y. Yan, L. Scolari, G. Kardaras, M. S.Berger, and L. Dittmann, “Cloud RAN for mobile networks – atechnology overview,” IEEE Commun. Surveys Tut., vol. 17, no. 1,pp. 405–426, First Quarter 2015.

[2] M. Peng, Y. Sun, X. Li, Z. Mao, and C. Wang, “Recent advances incloud radio access networks: System architectures, key techniques,and open issues,” IEEE Commun. Surveys Tut., vol. 18, no. 3, pp.2282–2308, Third Quarter 2016.

[3] P. Rost, C. J. Bernardos, A. D. Domenico, M. D. Girolamo,M. Lalam, A. Maeder, D. Sabella, and D. WÃijbben, “Cloudtechnologies for flexible 5G radio access networks,” IEEE Commun.Mag., vol. 52, no. 5, pp. 68–76, May 2014.

[4] N. Saxena, A. Roy, and H. Kim, “Traffic-aware cloud RAN: A keyfor green 5G networks,” IEEE J. Sel. Areas Commun., vol. 34, no. 4,pp. 1010–1021, Apr. 2016.

[5] T. Werthmann, H. Grob-Lipski, and M. Proebster, “Multiplexinggains achieved in pools of baseband computation units in 4Gcellular networks,” in Proc. IEEE PIMRC ’13, Sep. 2013, pp. 3328–3333.

[6] D. Wubben, P. Rost, J. S. Bartelt, M. Lalam, V. Savin, M. Gorgoglione,A. Dekorsy, and G. Fettweis, “Benefits and impact ofcloud computing on 5G signal processing: Flexible centralizationthrough cloud-RAN,” IEEE Signal Process. Mag., vol. 31, no. 6, pp.35–44, Nov. 2014.

[7] D. Pompili, A. Hajisami, and H. Viswanathan, “Dynamic provisioningand allocation in cloud radio access networks (C-RANs),”Ad Hoc Netw., vol. 30, pp. 128–143, 2015.

[8] Y. Shi, J. Zhang, and K. B. Letaief, “Group sparse beamforming forgreen Cloud-RAN,” IEEE Trans. Wireless Commun., vol. 13, no. 5,pp. 2809–2823, May 2014.

[9] Y. Shi, J. Zhang, K. B. Letaief, B. Bai, and W. Chen, “Large-scaleconvex optimization for ultra-dense Cloud-RAN,” IEEE WirelessCommun., vol. 22, no. 3, pp. 84–91, Jun. 2015.

[10] M. Qian, W. Hardjawana, J. Shi, and B. Vucetic, “Baseband processingunits virtualization for cloud radio access networks,” IEEEWireless Commun. Lett., vol. 4, no. 2, pp. 189–192, April 2015.