ANONYMOUS wrote: There's a past discussion on this here: [help3002]⬈ The unreliability of UDP needs to be addressed so that any lost datagrams do not prevent a query being answered. The use of sequence numbers could be used to address this but, as there's no natural *sequence* to the datagrams, you'd probably just employ a monotonically increasing (unique) value. Moreover, as the queries are independent of each other, it's unclear how a receiver should react to a 'missing' datagram, because it should still accept later datagrams before the missing one arrives.