I don't think we have any measurements here. I'd just run some tests with different numbers of edges and vertexes. Then I would extrapolate. I would not be surprised if there were optimizations that could be done on the code though On Sun, May 24, 2020, 4:41 AM yanhc519 <yanhc519@163.com> wrote:

Hi all.

I am using ospfv3 routing in an embeded processor. The most time consuming part is the shortest path first(spf, Dijkstra) calculation in ospf6_spf.c. Since the spf calculation uses a priority queue, so the computational complexity is O(E*log2(V)) where E is the number of edges and V is the number of vertices. I have measured the running time of spf calculation in the embeded processor with small E and V, i.e. E=1 and V=2. The measured spf running time is 10 milliseconds. My goal is to evaluate the spf running time in a large network e.g. V=800 and E=1600. Since deploy such a large network in conceptual design phase is not practical, so is there an estimation method using the computational complexity O(E*log2(V)) and measurement in small network? Thanks in advance.

yan

_______________________________________________ frog mailing list frog@lists.frrouting.org https://lists.frrouting.org/listinfo/frog