By epoch do you mean the unix-style timestamp - the number of seconds since Thu Jan 1 00:00:00 1970 UTC? ("The epoch" refers to that exact date/time - see Unix Time.)

(

  perl -e "print scalar gmtime(0);"
  Thu Jan  1 00:00:00 1970
[download]

)

If so, and if (and that's an enormous if) the timestamps are always exactly 5 minutes apart, you could change the sub-select clause to

old.timestamp = new.timestamp - 300
[download]

Otherwise, the sub-select for the previous-in-time row will have to do.

If your tables had an ID column that automatically incremented by one for each row*, the sub-select could be replaced by 'old.id = new.id - 1' as in the MySQL Cookbook examples.

*Of course that assumes that rows are never inserted out of order, like when the network connection drops for a few hours and you have to go back later and add the missing data . . .