Persistence

Well, as Nitass' alludes, even though source address persistence isn't being used, it still has to gather the data for when it is needed. If it weren't storing that data, then it wouldn't much use. The question then becomes, what is driving you to use source address as a fallback, or really any fallback with cookie as the primary. HTTP cookie persistence is, for any purely browser-based application, the most stable method of all of them. There are certainly edge cases where a fallback would make sense, like when a browser spawns separate agents to separate URLs, but that's not a typical sort of thing.