Doc. # 1-00000781 | |||
---|---|---|---|
Date Updated | 10-26-2018 | Date Created | 09-12-2017 |
Document Type | Knowledge Base | Related OS | |
Related Product | ZP9D-115RM-LR/ ZXT9-IO-222R2/ ZXT9-RM/ ZZ9D-NA-LR/ ZZ9D-NB-LR/ ZZ9D-NC-LR/ ZZ9D-ND-LR |
What is the effect of Repeaters on throughput. | |||
---|---|---|---|
Solution:
Repeaters Network bandwidth considerations Using mesh repeaters in a network reduces the overall network data throughput as each repeater must buffer an entire packet before retransmitting it. For example: if the destination is within range of the transmitter and the packet is 32 bytes long, the transmission will take 72ms on a 9600 baud XStream modem (much faster modems are available). If that same packet has to propagate through two repeaters, it will take 72ms to arrive at the first repeater, another 72 ms to get to the second and a final 72ms to get to the destination for a total of 216ms. Taking into account UART transfer times (~1ms/byte at 9600 baud), a server to send a 32 byte query and receive a 32 byte response in about 200ms, allowing for 5 polls per second. With the two repeaters in the path, the same query/response sequence would take about 500ms for 2 polls per second. To summarize, this system is sending and receiving 64 bytes 5 times per second for a throughput of 320 bytes per second with no repeaters and 128 bytes per second with 2 repeaters. Generally, the network throughput will decrease by a factor of 1/(R+1), with R representing the number of repeaters between the source and destination. Note that these numbers are absolutely worst case to illustrate how the system would perform in a typical, low bandwidth system. As a counter example the 115kbps Long Ranged radio can transfer the same 32 byte packet in 12 ms for a round trip with UART transfer times of ~30ms or 33 polls per second (1066 bytes per second) with no repeaters. With two repeaters the time would be ~100ms round trip time for 10 polls per second or 320 bytes per second network throughput with two repeaters.
|