GB2249003A

GB2249003A - Data transmission in burst mode

Info

Publication number: GB2249003A
Application number: GB9022825A
Authority: GB
Inventors: Brian Michael Unitt; Paul Douglas Welton
Original assignee: STC PLC
Current assignee: STC PLC
Priority date: 1990-10-19
Filing date: 1990-10-19
Publication date: 1992-04-22
Also published as: GB9022825D0

Abstract

A method of identifying the start of a data burst in a burst mode digital data transmission system, wherein a data burst sent over the system is preceded by a preamble containing a marker bit pattern for the purpose of identifying the start of the data burst and a receiver includes means for correlating the received data signals with a local reference marker bit pattern, the method including the step of producing a data burst start signal when the correlation between the received data signals and the local reference marker bit pattern exceeds a predetermined value, e.g. the received data differs from the local reference marker bit pattern in, at most, two bit positions. The transmission medium is optical fibre. <IMAGE>

Description

Data Transmission in Burst Mode This invention relates to digital data transmission systems where the data is sent in "bursts" or "packets".

It is common practice in electronic information systems to convey the information in bursts or packets which are transmitted on demand when sufficient information is available to be sent. In such a system the timing of a burst may be asynchronous with respect to any previous burst and the receiver needs to be able to identify the start of the information burst in order to decode the incoming information correctly.

Where the transmission medium is of a binary nature (as is often the case) a message normally consists of groups of characters each of which is made up from several binary digits. It is common practice to prefix the information stream with one or more marker bits whose purpose is to identify to the receiver the start of the message and allow the receiver to divide the binary digits into the correct character groupings.

If the packet is totally asynchronous (with no fixed phase relationship to a local clock signal) the receiver will need to identify the phase (and possibly also the frequency) of the incoming data in order to sample the binary digits at the correct instant. To achieve this the marker bit pattern is often preceded by a bit pattern with high clock phase content.

A drawback of this approach is that errors occurring during transmission which cause corruption of the marker bit pattern prevent recognition of the start point causing loss of the information. Furthermore the marker bit pattern could later appear within the information itself resulting in incorrect infórmation capture.

The marker bit pattern must be chosen such that errors in any earlier part of the preamble are unlikely to cause apparent but false (early) detection of a pattern corresponding to the marker.

According to the present invention there is provided a method of identifying the start of a data burst in a burst mode digital data transmission system, wherein a data burst sent over the system is preceded by a preamble containing a marker bit pattern for the purpose of identifying the start of the data burst and a receiver includes means for correlating the received data signals with a local reference marker bit pattern, the method including the step of producing a data burst start signal when the correlation between the received data signals and the local reference marker bit pattern exceeds a predetermined value.

In a binary digital system this might be, for example, when the received data differs from the local reference marker bit pattern in, at most, two bit positions.

Embodiments of the invention will now be described with reference to the accompanying drawings in which: Fig. 1 illustrates a transmission format of digital data in a burst mode, Fig. 2 illustrates details of the make-up of part of a data burst, and Fig. 3 illustrates an implementation of a data burst marker identification arrangement.

In the detailed example which follows: - information characters consist of 8-bit bytes, whose most significant (leftmost) bit is transmitted first - "preamble" refers to the whole of the field preceding the information - "phase alignment" refers to that part of the preamble which precedes the marker and whose main purpose is to convey clock phase and frequency to the receiver Fig. 1 illustrates the format of a digital data burst 10 consisting of a phase alignment field 11, a marker bit pattern 12 and the information itself 13.

Data bursts are separated by a gap 14 whose size may vary. Previous and subsequent bursts lOa and lOb respectively are also shown.

The phase alignment field would normally be chosen for maximum clock content, appearing on the medium as alternate ones and zeroes (binary 10101010...). A marker bit pattern with a length of 8 bits (for example binary 10100100) can be chosen such that at least three errors must occur in the preamble before false (early) detection of the marker can occur.

Fig. 2 illustrates one such an event where phase alignment field 11 and marker bit pattern 12 contain errors 20 at the bit positions shown and result in apparent detection of the marker 21.

In the normal embodiment of such a system any error in the marker bit pattern would prevent the correct start position from being recognised. In this example, according to the present invention, the correct start position is deemed to have arrived if the bit pattern on the medium differs from the expected marker bit pattern in at most one bit position. With this arrangement the correct position of the start of the information field will be correctly detected for any single error within the preamble.

Marker fields of greater length than eight bits can be chosen to give greater error resilience.

The length of the marker field and the bit patterns appearing within the preamble would be chosen to optimise the performance of the system with respect to the expected error characteristics of the medium.

Whilst reference has been made to information transmission using data packets and exact matching of markers, it is recognised that such techniques are known and readily available to those skilled in the art.

In order to implement the invention an arrangement such as that shown in Fig. 3 can be used.

The incoming data stream 31 is fed into a serial-in/ parallel-out shift register 30 clocked at the data rate. The shift register outputs feed a number of exclusive-OR gates 40 whose other inputs are driven from logic levels representing the correct bit pattern 41 of the marker. The output of each exclusive-OR gate will be logic high when the logic states of the two inputs differ. The number of gates whose outputs are high therefore indicates the degree of mismatch between the pattern arriving on the medium and the marker. When the number of mismatching bits is below the desired threshold the output 51 of evaluation circuit 50 indicates that the point of alignment has been found.

The design of the individual elements of this circuit will be evident to those skilled in the art.

Claims

1. A method of identifying the start of a data burst in a burst mode digital data transmission system, wherein a data burst sent over the system is preceded by a preamble containing a marker bit pattern for the purpose of identifying the start of the data burst and a receiver includes means for correlating the received data signals with a local reference marker bit pattern, the method including the step of producing a data burst start signal when the correlation between the received data signals and the local reference marker bit pattern exceeds a predetermined value.

2. A burst mode data transmission system where a data burst is preceded by a preamble containing a marker bit pattern for the purpose of identifying the start of the data burst and where said marker bit pattern is identified at the receiver by a match differing from the expected value by at most a predetermined amount.

3. A system as in claim 1 where the preamble also includes a data pattern designed to identify to the receiver the optimum sampling point of the incoming data.

4. A system as in claim 2 where the marker bit pattern is chosen such that the effect of transmission errors in the part of the preamble designed to identify the optimum sampling point resulting in incorrect identification of the marker position are minimised.

5. A system as in any previous claim where information is transmitted as binary digits represented in a form applicable to the transmission medium used.

6. A system as in any previous claim where the transmission medium is optical fibre.

7. An arrangement for identifying the start of a data burst in a digital data burst mode transmission system substantially as described with reference to the accompanying drawings.

8. A method for identifying the start of a data burst in a digital data burst mode transmission system substantially as described with reference to the accompanying drawings.