If you have a Tx-Rx pair, the link latency is added to the runtime of both nodes, while in reality, that should only be added to the Rx node. That is, the runtime calculation should be:
- Tx node runtime: data_size* BW
- Rx node runtime: data_size* BW + latency
Although it might be small for a Tx-Rx pair, this discrepancy adds up and can become non negligible.
If you have a Tx-Rx pair, the link latency is added to the runtime of both nodes, while in reality, that should only be added to the Rx node. That is, the runtime calculation should be:
Although it might be small for a Tx-Rx pair, this discrepancy adds up and can become non negligible.