Research On Intent Processing Technology Of Optical Network

In the B5G communication age,with the development of multi-region,multi-network,multi-service and holographic communication,the scale of data and scenarios of network services have greatly expanded,such as network services in the Internet of Vehicles,Internet of Things,and high reliability and low latency scenarios.These new network services that interact closely with dynamic environments focus more on user expectations and goals for the network,while neglecting specific network parameters.The user requirement that describes user expectations and goals through declarative language is called intent.Compared to traditional network services that typically provide specific network metrics such as bandwidth,latency,jitter,etc.,changes of intent in quantity,content,and expression pose higher challenges in obtaining accurate user requirement and traditional network architecture.The current research on intent focuses more on descriptions and classification of intent from a single perspective and a single dimension,and these research methods without multidimensional classification lack completeness in the classification of intent.Moreover,most existing research on intent parsing focuses more on how to express user intent through various methods,ignoring the impact of the network environment,making the accuracy of intent parsing relatively low.In summary,structuring precise user intent parsing mechanisms and flexible intelligent network architectures has become a trend.The Intent Driven Optical Network(IDON)architecture,which has the characteristics of self-learning,decision-making,and self-protection,has also become a research hotspot in intelligent network architecture.Therefore,from the perspective of intent,this article conducts in-depth research on intent processing technology and IDON architecture in optical networks.The main contents are as follows:For the problem of parsing intent without multi-dimensional perspective and interacting with the network environment,this paper proposes an intent joint-parsing mechanism based on different dimensional classifications.According to different dimensions,this mechanism firstly classifies the intent and network environment status.And the classification will make network status and historical data of users’ service requests part of intent parse.Then,this mechanism will parse the intent by the form of division and combination based on the classification results,which not only improves the accuracy of intent parse from the perspective of multidimensional aspects and network environment status,but also advances possible policy conflicts to the intent parse module.Therefore,the mechanism can save time and resources of IDON system.In the intent joint-parse mechanism,the classification of intent will lead to the possibility of conflicts between various intentions.In response to this problem,this paper proposes a game theory based intent conflict guarantee mechanism.The guarantee mechanism introduces the Stackelberg model and multi-agent game theory solution to ensure that conflicts in parsing intent are guaranteed.And the guarantee mechanism will contribute to improving the accuracy of intent parse.Finally,this paper proposes a new IDON architecture based on the intent joint-parse mechanism and the intent conflict guarantee mechanism.The simulation results show that the new IDON architecture can reach more than 90%in both intent parse and conflict resolution.Compared with conventional methods,it can effectively reduce the delay and blocking rate.