■ Higher symmetrical bandwidth capability, which could become an even more significant advantage if demand for P2P and Web 2.0 applications increases rapidly (see 4.2.3).
■ Greater bandwidth scalability. The ability to swap Ethernet line cards in the central office to increase bandwidth capability makes AONs far more flexible than PONs. By virtue of an active Ethernet switch, AONs have more sophisticated remote management tools than PONs to deliver pre-agreed bandwidth rates to specific customers as and when they are required.
■ To run and manage a PON, more expensive IP Layer 3 CPEs are required for the customer to access the services he or she has subscribed to; these CPEs also need to have Layer 3 security functionality. By contrast, an AON operator can move Layer 3 management into the distribution layer (Layer 2) and take advantage of cheaper Layer 2 CPEs. This also makes network management easier as the AON operator doesn’t have to monitor every CPE.
■ Easier to isolate faults remotely, which reduces the number of expensive truck rollouts.
■ The way MDUs (multi-dwelling units) are clustered in a city might make it easier to build ring-designs – where the buildings are daisy-chained (which is possible with an AON) – than rolling out a PON-based tree architecture.
■ Higher opex costs to maintain a multitude of remote active Ethernet switches.
■ Potential difficulty of finding suitable locations for the Ethernet switches (which need to be fed with power). This is particularly true for alternative operators with no existing active elements in their access network.
■ Higher capex costs associated with point-to-point architectures (see 2.3.5).
■ Cannot support RF overlay.