Solectek SKYWAY-LM Series, User Manual, Page: 65 / 69

65
11. Native TDM/IP Transport
11.1 Overview
One of the key benefits of LM Series point-to-point microwave radio Indoor Unit (IDU) is the ability to
transport native TDM data along side Ethernet frame data with minimal latency and without degrading
TDM jitter performance. Latency is indirectly configurable as a function of bandwidth mode (symbol rate),
modulation type and FEC configuration, but is generally well under 1 ms per link. Jitter tolerance and
attenuation are compliant with ITU G.823.
11.2 Native TDM/IP Transport Architecture Description
For a functional block diagram depicting the LM IDU TDM/IP transport architecture, refer to Figure 11-1.
TDM data, in this case E1 bit streams, are first conditioned by E1 line interface transceivers. The data
from each separate active E1 port is separately buffered in individual FIFOs, and “stuff” bytes are added
to accommodate the lack of synchronicity between the symbol rate of the air link and the bit rate of each
incoming E1 bit streams. Incoming Ethernet frames are separately buffered and GFP framed to account
for the difference between the Ethernet frame and the LM over-the-air frame size. The individual E1 bit
streams are byte-wise multiplexed together along with GFP framed Ethernet data, assembled into an
over-the-air frame as indicated in Figure 11-2, then finally coded and modulated for transmission over the
link. On the far-side of the link, the received signal is demodulated and decoded, and the received data
stream is then de-multiplexed into its constituent E1 and Ethernet data streams. The stuff bytes are
removed from each stream, then each bit stream is separately buffered and rate smoothed prior to line
conditioning in the line interface transceiver. Likewise, the GFP framing is stripped from the Ethernet
data, and whole frames are assembled prior to ingressing the frame to the internal IDU Ethernet switch
for forwarding to the proper port based on the frames destination address.
In contrast, pseudo-wire techniques employed in other products require buffering E1 traffic into Ethernet
frames and transporting them over the link using an encapsulation scheme such as GFP. From hop to
hop, these frames accumulate jitter and latency due to Ethernet priority queues. At final destination, a
buffer is required to accommodate the large amount of accumulated jitter, adding considerable latency to
the end-to-end E1 traffic.
Figure 11-1: Block Diagram