Packet and message can be the same meaning but of course, the number of bits per packet can vary.
- e.g. one packet can carry 1518 bytes * 8 bits/byte as per IEEE 802.3 ethernet standard.
Packet Error Rate (PER) is defined as ratio of number of packets received in error to total number of transmitted packets.
The relationship between BER and PER for n bits per packet is;
The words Message & Bits are not used in the OSI model, instead it depends on the procotol used for each layer and detection counting method, shown here;
7 Application Data High-level APIs, including resource sharing, remote file access
6 Presentation Translation of data between a networking service and an application; including character encoding, data compression and encryption/decryption
5 Session Managing communication sessions, i.e. continuous exchange of information in the form of multiple back-and-forth transmissions between two nodes
4 Transport Segment, Datagram Reliable transmission of data segments between points on a network, including segmentation, acknowledgement and multiplexing
3 Network Packet Structuring and managing a multi-node network, including addressing, routing and traffic control
2 Data link Frame Reliable transmission of data frames between two nodes connected by a physical layer
1 Physical Symbol
Since Forward Error Correction must be disabled, the lowest levels must be used with FEC disabled for the Data or Physical Layer whichever can be used without FEC.
Generally the quality of radio is determined by the threshold for fixed error rate near threshold and the slope of error rate vs SNR or Eb/No
A higher threshold may occur with discriminator error or multi-path Fading cancellation or Rx noise or adjacent channel interference or poor LO rejection etc. or Inter-symbol-Interference (ISI) or filter Group Delay distortion etc
The slope depends on noise characteristics being random, spectral shape, repetition rate etc, modulation method with BW compression affected by Comm. Theory such as (Shannon-Hartley's Law.)
It becomes much more steep, above some threshold using FEC depending on overhead and method used.
It is possible to have a lower threshold at 1% error yet higher threshold at 0.1% or 1e-9 threshold BER due to this slope and above interference.
Back in the 70's,80's I found worst-case & random data patterns to measure contributions of the BER curve due to my path margin budget for contributors like;
- Group Delay distortion, comparator asymmetry, unequal delays and rise time, pre-compensation and channel periodic defects.
- By isolating each root-cause contribution to BER using different patterns or by fault of noise, ripple or jitter injection, radiated impulse rejection, impedance control, shielding, one can optimize the design including the filter characteristics of Bandpass vs Bandstop.
- But 1st you must define an error budget from all sources( ie. a design spec)
I learned this the hard way, as my SCADA controlled robot over by a long distance coax cable fell to the ground from an arc-welder's noise, far away during a major client presentation. This was followed by 3 layers of error detection, correction implemented the following week.