In 5G networks,optimization of antenna beam weights of base stations has become the key application of AI for network optimization.For 6G,higher frequency bands and much denser cells are expected,and the importance of...In 5G networks,optimization of antenna beam weights of base stations has become the key application of AI for network optimization.For 6G,higher frequency bands and much denser cells are expected,and the importance of automatic and accurate beamforming assisted by AI will become more prominent.In existing network,servers are“patched”to network equipment to act as a centralized brain for model training and inference leading to high transmission overhead,large inference latency and potential risks of data security.Decentralized architectures have been proposed to achieve flexible parameter configuration and fast local response,but it is inefficient in collecting and sharing global information among base stations.In this paper,we propose a novel solution based on a collaborative cloud edge architecture for multi-cell joint beamforming optimization.We analyze the performance and costs of the proposed solution with two other architectural solutions by simulation.Compared with the centralized solution,our solution improves prediction accuracy by 24.66%,and reduces storage cost by 83.82%.Compared with the decentralized solution,our solution improves prediction accuracy by 68.26%,and improves coverage performance by 0.4 dB.At last,the future research work is prospected.展开更多
Cloud computing technology is the culmination of technical advancements in computer networks,hardware and software capabilities that collectively gave rise to computing as a utility.It offers a plethora of utilities t...Cloud computing technology is the culmination of technical advancements in computer networks,hardware and software capabilities that collectively gave rise to computing as a utility.It offers a plethora of utilities to its clients worldwide in a very cost-effective way and this feature is enticing users/companies to migrate their infrastructure to cloud platform.Swayed by its gigantic capacity and easy access clients are uploading replicated data on cloud resulting in an unnecessary crunch of storage in datacenters.Many data compression techniques came to rescue but none could serve the purpose for the capacity as large as a cloud,hence,researches were made to de-duplicate the data and harvest the space from exiting storage capacity which was going in vain due to duplicacy of data.For providing better cloud services through scalable provisioning of resources,interoperability has brought many Cloud Service Providers(CSPs)under one umbrella and termed it as Cloud Federation.Many policies have been devised for private and public cloud deployment models for searching/eradicating replicated copies using hashing techniques.Whereas the exploration for duplicate copies is not restricted to any one type of CSP but to a set of public or private CSPs contributing to the federation.It was found that even in advanced deduplication techniques for federated clouds,due to the different nature of CSPs,a single file is stored at private as well as public group in the same cloud federation which can be handled if an optimized deduplication strategy be rendered for addressing this issue.Therefore,this study has been aimed to further optimize a deduplication strategy for federated cloud environment and suggested a central management agent for the federation.It was perceived that work relevant to this is not existing,hence,in this paper,the concept of federation agent has been implemented and deduplication technique following file level has been used for the accomplishment of this approach.展开更多
Snow cover plays an important role in the fields of climatology and cryospheric science. Remotely-sensed data have been proven to be effective in monitoring snow covers. Improved methods to process the 8-day snow-cove...Snow cover plays an important role in the fields of climatology and cryospheric science. Remotely-sensed data have been proven to be effective in monitoring snow covers. Improved methods to process the 8-day snow-cover products derived from MODIS Terra/Aqua data can dramatically increase the data quality and reduce noise. A five-step algorithm for removing cloud effects was designed to improve the quality of MODIS snow products, and the overall accuracy of the MODIS snow data without cloud（defined as cloud-free snow-cover dataset） was enhanced by more than 90% based on direct and indirect validation methods. The snow-cover frequency（SCF） and snow-cover rate（SCR） of Central Asia were analyzed from 2000 to 2015 using trend analysis and empirical orthogonal functions（EOFs）. Over the plain regions, the SCF displayed a significant north-south declining trend with a rate of 0.03 per degree of latitude, and the SCR showed a similar north-south gradient. In the mountainous areas, the SCF significantly increased with altitude by 0.12 per kilometer. Within the study area, the SCF in 65% of the study area experienced an increasing trend, but only 4.3% of the SCF-increasing pixels passed a significance test. The remaining 35% of the area underwent a decreasing trend of SCF, but only 5.2% of the SCF-decreasing pixels passed a significance test. For the entire Central Asia, the inter-annual variations of snow-cover presented a slight and insignificant increase trend from 2000 to 2015. However, the change trends of snow cover are different between the plain and mountainous regions. That is, the annual mean SCR in the plain areas displayed an increasing trend, but a decreasing trend was found in the mountainous areas.展开更多
基金supported by the National Key Research and Development Program of China(2020YFB1806800)funded by Beijing University of Posts and Telecommuns(BUPT)China Mobile Research Institute Joint Innoviation Center。
文摘In 5G networks,optimization of antenna beam weights of base stations has become the key application of AI for network optimization.For 6G,higher frequency bands and much denser cells are expected,and the importance of automatic and accurate beamforming assisted by AI will become more prominent.In existing network,servers are“patched”to network equipment to act as a centralized brain for model training and inference leading to high transmission overhead,large inference latency and potential risks of data security.Decentralized architectures have been proposed to achieve flexible parameter configuration and fast local response,but it is inefficient in collecting and sharing global information among base stations.In this paper,we propose a novel solution based on a collaborative cloud edge architecture for multi-cell joint beamforming optimization.We analyze the performance and costs of the proposed solution with two other architectural solutions by simulation.Compared with the centralized solution,our solution improves prediction accuracy by 24.66%,and reduces storage cost by 83.82%.Compared with the decentralized solution,our solution improves prediction accuracy by 68.26%,and improves coverage performance by 0.4 dB.At last,the future research work is prospected.
文摘Cloud computing technology is the culmination of technical advancements in computer networks,hardware and software capabilities that collectively gave rise to computing as a utility.It offers a plethora of utilities to its clients worldwide in a very cost-effective way and this feature is enticing users/companies to migrate their infrastructure to cloud platform.Swayed by its gigantic capacity and easy access clients are uploading replicated data on cloud resulting in an unnecessary crunch of storage in datacenters.Many data compression techniques came to rescue but none could serve the purpose for the capacity as large as a cloud,hence,researches were made to de-duplicate the data and harvest the space from exiting storage capacity which was going in vain due to duplicacy of data.For providing better cloud services through scalable provisioning of resources,interoperability has brought many Cloud Service Providers(CSPs)under one umbrella and termed it as Cloud Federation.Many policies have been devised for private and public cloud deployment models for searching/eradicating replicated copies using hashing techniques.Whereas the exploration for duplicate copies is not restricted to any one type of CSP but to a set of public or private CSPs contributing to the federation.It was found that even in advanced deduplication techniques for federated clouds,due to the different nature of CSPs,a single file is stored at private as well as public group in the same cloud federation which can be handled if an optimized deduplication strategy be rendered for addressing this issue.Therefore,this study has been aimed to further optimize a deduplication strategy for federated cloud environment and suggested a central management agent for the federation.It was perceived that work relevant to this is not existing,hence,in this paper,the concept of federation agent has been implemented and deduplication technique following file level has been used for the accomplishment of this approach.
基金funded by the National Key Research and Development Program of China (2016YFA0602302,2016YFB0502502)
文摘Snow cover plays an important role in the fields of climatology and cryospheric science. Remotely-sensed data have been proven to be effective in monitoring snow covers. Improved methods to process the 8-day snow-cover products derived from MODIS Terra/Aqua data can dramatically increase the data quality and reduce noise. A five-step algorithm for removing cloud effects was designed to improve the quality of MODIS snow products, and the overall accuracy of the MODIS snow data without cloud（defined as cloud-free snow-cover dataset） was enhanced by more than 90% based on direct and indirect validation methods. The snow-cover frequency（SCF） and snow-cover rate（SCR） of Central Asia were analyzed from 2000 to 2015 using trend analysis and empirical orthogonal functions（EOFs）. Over the plain regions, the SCF displayed a significant north-south declining trend with a rate of 0.03 per degree of latitude, and the SCR showed a similar north-south gradient. In the mountainous areas, the SCF significantly increased with altitude by 0.12 per kilometer. Within the study area, the SCF in 65% of the study area experienced an increasing trend, but only 4.3% of the SCF-increasing pixels passed a significance test. The remaining 35% of the area underwent a decreasing trend of SCF, but only 5.2% of the SCF-decreasing pixels passed a significance test. For the entire Central Asia, the inter-annual variations of snow-cover presented a slight and insignificant increase trend from 2000 to 2015. However, the change trends of snow cover are different between the plain and mountainous regions. That is, the annual mean SCR in the plain areas displayed an increasing trend, but a decreasing trend was found in the mountainous areas.
