Locals are, in my opinion, perfectly acceptable with parallel loops. They are rather tidy, and typically won't cause race conditions. If you wanted to be really careful about it, you could use either LabVIEW native quese, or a storage VI (a subVI with a single loop using unitialized shift registers - also known as LabVIEW 2 style global, among other things.) If this is a simple program, with no scalability, then I would suggest the use of locals is your best bet.
Good luck, and please let us know how you do.