| With the deepening of network informatization and digitization,the wide application of intelligent technologies such as cloud computing,artificial intelligence,big data and Internet of Things has led to the rapid growth of various types of business traffic.As the basic support of cloudbased technologies,the research related to data centers has become a popular topic for scholars to explore.The information construction of government and industry has led to a surge in the demand for information and network resources,and the significance and importance of data centers has become increasingly prominent.As a hub for information exchange,data centers store large-scale user privacy data.Data center is based on optical network for data bearing,which features high transmission rate,long communication distance and large network capacity.As an emerging technology of quantum communication,quantum key distribution technology based on the basic principles of quantum mechanics relies on optical network for bearing,with theoretically unconditional security characteristics,providing a new idea for network security of data center.The quantum key distribution based quantum confidential communication network is relatively mature at this stage and has the potential of industrial application.Based on the quantum key distribution network,this thesis focuses on the key provisioning strategies for different types of services in data centers.The paper proposes a quantum security-based solution to improve the security of data center services in two typical services,namely "real-time data services" and "asynchronous backup services",and completes This paper proposes a solution based on quantum security to improve the security of data center services,and completes three core researches:"Key provisioning strategy for real-time data services based on resource dynamic awareness","Key provisioning strategy for asynchronous backup services based on joint resource scheduling",and "Resource allocation strategy under key provisioning failure based on buffer pool"." three core research contents.The main research work and innovative achievements of this paper include the following three aspects:(1)For real-time data business scenarios in data centers,the data center real-time data business model and multi-dimensional resource dynamic-aware resource allocation method are proposed to solve the business bearing and key provisioning problems of real-time data scenarios.When providing quantum key services for data center real-time services within the metropolitan area,it is necessary to consider not only the reasonable utilization of key resources,but also how to allocate computing resources.The scheme establishes network and service models by abstracting real-time service characteristics,and cooperatively considers key resources in the data center network and computational resources inside the data center,and proposes a key-aware algorithm for real-time service resource allocation in arithmetic networks,which effectively alleviates the limitation of the key bottleneck on the network performance and makes the network resources more fully utilized at the same time.Simulations are performed with metro network topology as the base,and the performance of the proposed algorithm in terms of network blocking probability,key and computing resource utilization is verified from multiple perspectives,resulting in a reduction of network blocking rate from 7.15%to 1.61%(77.5%reduction ratio).(2)For the data center asynchronous backup business scenario,the data center asynchronous backup business model and the scheduling method based on the flexible characteristics of the time dimension are proposed to solve the data bearing and key supply problems of the asynchronous backup business.The asynchronous backup service allows a certain initial delay for the transmission start time,but the data transmission must be completed before the deadline,which introduces more constraints in the time dimension;at the same time,the key resources have the characteristic of "gradual accumulation and instantaneous consumption",and the coarser granularity of the asynchronous backup service causes the slow accumulation and fast consumption of keys to have a more significant impact.The impact of slow accumulation and fast consumption of keys is more significant due to coarser granularity of asynchronous backup service.The scheme takes advantage of the partially predictable service scheduling feature,applies the modeling of key and spectrum schedules,and proposes a joint time,spectrum and key resource scheduling method.The simulation program verifies the effectiveness of the algorithm and compares and analyzes the impact of different factors such as service attributes and initial keys on the network performance such as network blocking probability and key resource utilization status from three perspectives of service initial delay,channel dynamic adjustment and spectrum time block scheduling,resulting in a 40.2%performance reduction ratio of network blocking rate.(3)For the data center network key generation failure scenario,a resource buffering method with the main objective of minimizing the impact of failure failure is proposed to solve the problem of network key provisioning and alternative node selection under the failure situation.Inter-data center communication involves almost all user data services,and either internal data center resource failure or key provisioning interruption will cause unpredictable losses.Through the establishment of the key provisioning failure model,this solution designs a buffer pool-based resource allocation strategy in the case of network node failure from the perspectives of both key usage and replenishment of the failed road,and ensures redundancy through a backup system to ensure that even in the case of node failure,a balanced distribution of resources can be maintained through the reasonable utilization of resources,thus reducing the impact of failure on network blocking rate and other performance indicators The simulation results show that the proposed algorithm can be used in the network.The simulation results show that the proposed algorithm reduces the network blocking probability by up to 35%in the case of failure,which is 19.2%higher than that in the case of no failure,and effectively improves the fault tolerance of the network. |
