Software-defined networking (SDN) promises to drastically simplify how transport networks deploy, operate and get serviced. Reducing OPEX remains a significant factor for implementing software-defined networking. Automating service creation, traffic and bandwidth control, and network management as well as reducing maintenance complexity of routing protocols remain areas where it will simplify backhaul and lower OPEX. The only questions seem, “When will this happen?” and “How much will it save?” And what about CAPEX? Can we expect reduction in purchase price of microwave backhaul based on such a migration?
To answer, we considered an all-outdoor Ethernet radio supporting IP/MPLS. And looked at parts we could remove in an SDN-only implementation. Figure 1 compares cost of IP-SDN radio supporting full IP/MPLS and SDN protocols against an SDN-only device without support for traditional routing and switching protocols. In parentheses, Figure 1 shows cost reduction for each set of components.
SDN Will Not Significantly Reduce Some Costs
Therefore, we can conclude that SDN will not significantly reduce the cost of RF-related components including RF-ASICs, modems and diplexers. Will still need all these components to support a radio link. With or without it. Will also need majority of mechanicals, power, connectors and miscellaneous components.
However, expect cost reductions in switch, FPGA and memory related to simpler protocols and centralization of control. Switch and memory functionality costs can significantly decrease due to reduction in local complexity to support routing and switching functions. Some switching and memory remains a requirement in SDN-only implementation. So we cannot eliminate them completely.
FPGAs, ICs and IEEE 1588
FPGA functionality would exist with less complexity due to less local management and control in a software-defined network. But will still need to support functionality such as IEEE 1588 synchronization, encryption and others. Because of switch, memory and FPGA simplification, reductions in power, passive, IC and other cost reductions will also occur.
In the End
In the end, when considering an SDN-only architecture move, expect microwave radio to come in around 13 percent lower CAPEX. A nice reduction. But in and of itself, probably not a reason to migrate.
That said, most strategies for migration of transport networks require support for existing protocols (e.g., Ethernet, IP/MPLS). Those alongside SDN protocols (e.g., OpenFlow, NETCONF/YANG, PCE) on the same platform. In these cases, very little cost reduction will be directly attributable to software-defined networking technology.
Reducing to the Primary Promise
In summary, despite little microwave CAPEX reduction, software defined networking promises to drastically simplify how transport networks deploy, operate and get serviced. OPEX reduction will remain the primary reason to deploy it.
Stay tuned. We’ll reveal more about OPEX savings in coming months. Meanwhile, we also have a software-defined networking presentation for the road to 5G. Just sign in then download.