Research On Defragmentation Technology Based On Spectrum Compactness In Multi-dimensional Optical Network

With the exponential growth of global IP data,the dynamic flexibility and ultra-large capacity requirements of have been increasing in optical networks.The transmission capacity in single-core optical fiber almost reaches the Shannon limit,and the frequency domain resources cannot meet the increasing bandwidth demand.Thus,multi-dimensional optical networks with multi-dimensional resources,such as time domain,frequency domain and space domain,become the potential future optical network.However,due to the shrinking of the transmission granularity and the dynamization of services,a large number of idle spectrum particles that are not utilized in the multi-dimensional optical network(ie,spectrum fragments)are present,and the problem of multi-dimensional resource fragmentation is very serious.In order to meet the above challenges,this paper is organized as follows,1)how to accurately describe the multi-dimensional resource state in multi-dimensional optical network,2)how to make full use of multi-dimensional spectrum fragments,3)how to defragment multi-dimensional spectrum resources.To improve network resource utilization.we proposed three innovative results accordingly as following.Firstly,to accurately describe the multi-dimensional resource state in multi-dimensional optical networks,this paper establishes a crosstalk-aware resource description model based on spectrum compactness.This model considers the physical crosstalk constraints of multi-dimensional optical networks in measuring the occupancy and idleness of network spectrum resources,and provides an important basis for studying the problem of spectrum defragmentation in multi-dimensional optical networks.Secondly,to effectively utilize multi-dimensional spectrum fragments in multi-dimensional optical networks,this paper proposes a crosstalk-aware virtual concatenation resource allocation scheme based on spectrum compactness(MCVC).The scheme considers spectrum compactness as the core selection priority,and implements the resource allocation scheme based on the virtual concatenation using super channel construction.The simulation results show that the network spectrum utilization can be effectively increased by about 10%.and the larger the number of spectrum slots for service requests,the better the effect,and the 12 fiber structure performs the best.Thirdly,to effectively defragment multi-dimensional spectrum resources in multi-dimensional optical networks,this paper proposes a crosstalk-aware spectrum defragmentation scheme based on spectrum compactness(CASD).The scheme considers spectrum compactness as the defragmentation decision mechanism,and constructs the network spectrum fragmentation steps.Moreover,we also respectively discuss the application of CASD in two-dimensional(frequency domain,airspace)and three-dimensional(time domain,frequency domain,airspace)scenarios.The simulation results show that the network spectrum utilization in two-dimensional scenary can be effectively improved by about 6%and can be effectively improved by about 5%in three-dimensional scenary.The higher the spectrum compactness threshold,the more obvious the defragmentation effect.