Abstract: Combining the theory and method of complex network， Internet IPlevel topology measured by CAIDA was used to study characteristics of birth and death nodes. Characteristics of birth and death nodes’ degree distribution and powerlaw distribution were obtained. In addition， statistic and evolution analysis of birth attachments were achieved. The results show that both birth and death nodes accord with powerlaw distribution approximately， and birth nodes do not tend to connect with birth nodes， but connect the nodes which have existed already in the network. The number of attachment is likely to be 1. Because birth attachments do not accord with preferential attachment， the theory of the rich becoming richer has little significance on dynamic topology and nodes with high degree do not bring great impact on birth nodes. Generating mechanism of attachment is obtained， and the number of attachments is stable and universal， which provides a valid basis for Internet modeling.

Key words: complex network, Internet IPlevel, birth node, death node, powerlaw distribution, preferential attachment, evolution analysis