The concept of a slip is quite straightforward once we recognize that the frequency of clocks, albeit notionally equal, can be different. If a bit source generates traffic using a clock of frequency f1 and the recipient expects traffic at a clock rate f2, there is clearly a problem if f1  f2. If a buffering arrangement is employed, the buffer will either overflow or underflow. This is the notion of a slip. If the transmit clock is high then data is lost; if the receive clock is high then data must be inserted, usually by repeating the last block or an equivalent strategy. This situation is not uncommon and can in fact occur in every network element involved in the delivery of the bit-stream between the end-points.
The rate at which slips occur is determined by the frequency difference, Df. It is conventional to express this difference in fractional units, i.e., (Df/f), which are dimensionless. For example, in the case of speech digitization at 8 kHz, the telephony standard, if the analog-to-digital (A/D) and digital-to-analog (D/A) clocks differ by 125 parts-per-million (ppm), also written as 125x10-6, then in every one-second interval one of the clocks will have gone through one additional cycle, or period, nominally 125 msec. That is, there will be one slip every second. If the (fractional) frequency difference is
1x10-11, then a slip will occur every 1.25x107 seconds (about 144 days) which, for all practical purposes, is never. The damaging impact of slips is of course dependent on the type of traffic and a brief summary is provided in the following table.
| TRANSMISSION TYPE | IMPACT |
| VOICE | AUDIBLE CLICKS |
| FAX | MISSING LINES / DISTORTION |
| VOICE BAND DATA | CARRIER DROPOUT |
| VIDEO | PICTURE FREEZE / "BLUE" SCREEN |
| DIGITAL DATA | REDUCED THROUGHPUT |
| ENCRYPTED DATA | RETRANSMISSION |
| CELLULAR TRANSMISSION | POOR HANDOVER |
|
