Does a network's latency affect its availability?

I am calculating the up time percent per year (availability) for simple paths across pairs of nodes in a network of 40 nodes. I wanted to know if there exists a quantitative and theoretical relationship between the latency and availability such that I would need to penalize the network's availability if the latency on a given edge between two nodes is very high. Or are the availability and latency numbers independent of each other; in other words I could have the same availability even if the latency is very high.