it should also be noted that the quantization window isn't the same thing as a cache TTL - it's a window in which all nodes responding to a request will select the same revision so that the response to a query is consistent with a single revision