MMAPs generally consist of a shelf with a multiple-gigabit ATM and TDM backplane. Numerous plug-in cards are available for the access and MAN connections. The ATM and TDM buses on the backplane interconnect the system's plug-in cards. Figure 3 shows a schematic diagram of an MMAP architecture. For clarity of the figure, the TDM bus is not shown, but it runs in parallel with the ATM bus.

Figure 3. MMAP Platform Architecture
The line cards connect to downstream devices (IADs, DSLAMs, routers, private automatic branch exchanges [PABX], etc.) on the client side of the MMAP. Each line card accepts data from one or more downstream devices and, if necessary, converts them from frame-relay or TDM format to ATM cells. Those signals then travel over the ATM bus to the uplink cards and out to the MAN or to other line cards and back to downstream devices. TDMbased traffic can also be cross-connected to other line cards over the TDM backplane bus when no ATM features are required. This reduces processing load and eliminates ATM cell packetization delays.
Upon receiving ATM cell data from the line cards, the MAN card combines those signals into one ATM stream with multiple virtual paths and circuits. It then sends the data on those virtual circuits to the high-speed MAN.
In the receive direction, each MAN card retrieves the incoming data from the MAN. That data then travels over the ATM cell bus to the appropriate line cards for distribution to the IADs, DSLAMs, and other devices. If necessary, the line cards also convert the incoming cells from the uplink cards to frame-relay and/or TDM format before sending it to the downstream devices.
The MMAP is effectively a full ATM switch. By exploiting all the powerful quality of service (QoS) features of ATM, each service gets its guaranteed performance. This is key to integration of all kinds of different functions and services in one platform.


