NetFlow using single parallel multi-linked lists

The bottleneck of NetFlow is the memory access. To solve this problem, we propose continuous reading and writing method whose efficiency reaches 90% or more. The single parallel multi-linked lists method used to handle the inactive timeout flows is based on the continuous reading and writing method. Theoretical analysis and experimental results show that the single parallel multi-linked lists method can be applied in OC-768(40Gbps) to manage tens of millions of entries.