Research On The Measurement Of IPv6 Active Address Behavior

The IPv6 protocol is the next generation Internet protocol to replace the current mainstream version of the IP protocol IPv4.After more than 20 years of development,it has become more mature.Its design not only considers the defects of the IPv4 protocol in the actual application of the Internet,but also It has many other features,such as hierarchical address allocation structure,more convergent routing,automatic address configuration mechanism,new broadcast address types,more effective support for mobility and security,etc.These new designs provide a richer use space and faster use efficiency for the next generation of the Internet.At the same time,since the date of the invention of IPv6,a large number of RFCs related to IPv6 address structure have been proposed,covering multiple fields such as address types,routing selection,IPv6 transition,and privacy protection.On the other hand,the design of the IPv6 protocol and these RFCs also make it much more complex than IPv4,leading to many new challenges in address space detection,address generation,address usage,network management,and network security.Nowadays,most researches at home and abroad are based on the IPv6 address structure,with the discovery of the inherent laws of IPv6 addresses as a breakthrough,and various researches are carried out.The complexity of the design of the IPv6 protocol makes the use of active addresses in the IPv6 network complex and changeable.This paper attempts to start from the perspective of the use of IPv6 active addresses,study the use of IPv6 active addresses in daily network activities,and design relevant measures with clear meaning to better manage IPv6 networks.The main work of this paper is as follows:1)Designed and implemented the measurement of the number of active users of the IPv6 network,which represents the number of active users of the client that have communication activities on the IPv6 network in a certain day.The IPv6 temporary address mechanism causes the client to have multiple IPv6 addresses,and there may be more than one address in communication,which makes it impossible to use the number of IPs in the network to represent the number of users.This paper takes advantage of the fact that the preferred temporary address of the IPv6 client changes every 24 hours and does not repeat,and proposes a measurement method that can effectively measure the number of users in the network.In this paper,the measurement is implemented with different forms of data sources,and a stream record compensation model is used to optimize the statistical results with sampled stream records as the data source.2)Designed and implemented IPv6 address sparsity,which reflects the density of addresses used in IPv6 networks over a period of time.The IPv6 address space is huge,but the address usage rate is very low,resulting in very sparse distribution of IPv6 addresses in the huge address space.At the same time,special attention should be paid to the address usage rate during the IPv6 address application process.Therefore,this paper designs IPv6 address sparsity,proposes a measurement method based on the RFC standard,and implements the measurement using full message flow as the data source.3)Designed and realized IPv6 address aggregation degree.Through the observation of actual network behavior,it is found that most active IPv6 addresses are clustered under a few specific subnet prefixes,indicating that IPv6 addresses have obvious aggregation characteristics.This measure abstractly expresses the number of addresses under the subnet prefix as the quality of a segment of arc,and abstractly expresses all the subnet prefixes as the central angle corresponding to this segment of the arc,and then calculates the circle centroid to represent the convergence of IPv6 addresses center.Then the centroid points at multiple time granularities are obtained,and the meanshift algorithm is used to calculate the number of the largest clusters formed by the centroid point sequence,and the ratio of the number of centroid points to the total number of centroid points is regarded as the IPv6 address aggregation degree in this period of time.Find out the subnet prefix corresponding to the area where the largest cluster is located,and then analyze the details of the allocation technology used in the IPv6 address plan,and compare with the recommendations,there can be a basis for adjusting the next step of the IPv6 address plan to achieve better results.This measurement is realized in a full message flow environment.Finally,the above three measurements are all based on the actual measured traffic of a campus network connected to CERNET2,and the results have reached expectations,indicating its availability and effectiveness.