Nope, got another idea
gives the universal timestamp while
cat /proc/net/dev | grep ppp0 | cut -d":" -f2 | cut -d" " -f1
gives the number of bytes received.
For each round (BTW, this will be a Munin plugin) the current number of bytes received gets compared with the number of bytes received during the previous round (stored somewhere in file in /tmp/). If it's <, then the connection has been dropped and reestablished. Total uptime gets calculated during each round.