Issue link: https://www.e-digitaleditions.com/i/1255991
July/August/September 2020 I 13 possible, but a genuine challenge arises collectively. What happens when packets get lost or arrive in the wrong order? What mechanism can be used to still bene- fit from the connectionless protocol layer and protect traffic, as much as possible, from arrival order issues and packet loss? This is where forward error correction (FEC) technology becomes critical. Before diving into FEC and how it is used, it is first necessary to understand what causes the packet drop in the first place. IP data packets are susceptible to packet drops, packet losses and packet arrival order misalignment as the facili- ty's network scales, more network nodes are added, and traffic load increases. High-density, complex network topologies with multiple nodes and routes can yield three main packet drop scenarios: • In scenario one, the network switch may drop pack- ets due to buffering issues. Switches often build up a packet queue, which are waiting to be forwarded or delivered to the next hop. In some scenarios, these packets spend too much time in the queue and the buffers start to fill, especially when the bandwidth used is high and traffic rates peak. The switch will then start to drop packets, disregarding sensitivities such as latency. • In the second scenario, each IP datagram packet carries a checksum value that is used to verify and detect errors that may have been introduced during data transmission. Checksum bit errors are translated into packet losses or packet drops by the receiver side. • In the third scenario, packets may drop due to a diver- gent routing path. This results in rerouting and pack- ets arriving in the wrong order at the destination. Highly complex networks will offer different routes to move signals to destinations, just as drivers can take the highway or back roads as they drive around their neighborhoods. Error Correction To fix errors and loss packets, the AV/IT administrator or system integrator needs to measure error rates. Software tools that measure error rates are available, as are dedicated hardware appliances that can run traffic in and out and measure packet loss statistics over time. Once the error rate and level are identified, the next step is to activate and configure the FEC mechanism. Audiovisual, IT administrators and systems integra- tors are strongly advised to choose AVoIP encoder and decoder devices that support FEC mechanisms. An FEC- enabled encoder places video packets in an internal vir- tual table with L columns and D rows. The packets are aligned in columns and rows with a mathematical XOR calculation for each row and for each column. The video packets, noted as 1-9 in Figure 6, together with the FEC packets, are sent over the network to the decoder. In case of packet loss, the reverse XOR operation (XNOR) is used to calculate back the missing packet. As shown in Figure 7, packet 5 was dropped due to net- work issues. By using the XNOR operation, the decoder can restore the missing packet and recover the data with- out the need to retransmit the data again. FIGURE 7: With XNOR, packet 5 that was dropped can be restored and the data recovered. FIGURE 6: FEC diagram. 1 2 3 FEC 4 5 6 FEC 7 8 9 FEC FEC FEC FEC 1 2 3 FEC 4 5 6 7 9 FEC FEC FEC 8 FEC FEC