One of the errors made in nearly every technology market is to think of trends in terms of the technologies they affect rather than the causes that drive them. So it is with optical and electrical (opto-electrical) convergence in the telecom network.
Certainly, optics is hot and Carrier Ethernet is becoming critical to planners. But it's the why behind these trends that will help you prepare for the changes they represent and link them to coherent business and infrastructure policies.
As technology improves in the optical transport of digital data, it reduces the unit cost of provisioning capacity. Given a stable or predictable level of traffic growth, the reduction in cost shifts the planning focus from managing aggregation for optimum utilization to managing capacity to control operations costs.
Since capacity gains are created at the optical network layer, optical planning is impacted first. But the lowest electrical layer is the bridge between traffic and transport, so it is impacted, as well. Optical enhancements have created increased interest in converging the optical and lowest electrical layers -- usually Ethernet -- into a single transport layer and a single set of devices.
IP convergence at the higher layers of the network is also driving opto-electrical convergence. While networks are segregated into packets (Ethernet/IP) and circuits (SONET/TDM), the value of having these electrical layers on a common backbone promotes optical independence.
But as TDM use declines and packet use explodes, there is less to be gained by maintaining independent optical networks, so TDM and packet can be converged with optical. That's particularly true when incremental capacity investment is almost completely in the packet realm. Driven from both above and below, converging optical and electrical layers is inevitable.
Convergence isn't an abstract concept if you expect tangible benefits. It means you'll be looking for devices that consolidate optical networking and electrical networking in a single box. The type of networking supported in both areas will depend on your current opto-electrical handling and your future plans and goals.
Network complexity
Unless you are installing a large amount of new greenfield gear, you can expect that your current operations practices will also have to be reviewed as a part of the convergence project.
Only a small part of the value of opto-electrical convergence comes from capital equipment economies; most of the benefit comes in operations savings. Network operators know that network complexity tends to be proportional to the number of devices, with the exponent depending on device connectivity.
That connectivity must be considered in normal geographic (or horizontal) terms, and in OSI-layer (or vertical) terms, so eliminating a layer of devices by consolidating two adjacent layers will significantly reduce complexity. It is important not to sacrifice capabilities in the consolidation, however, or you'll lose more than you gain in the operations expense (opex) area.
Converged opto-electrical devices must groom traffic connections at both the optical and electrical layers with the connection granularity your network requires.
If you currently route Gigabit Ethernet traffic in your Ethernet layer, you'll need to route it in an opto-electrical layer or add incremental Gigabit Ethernet grooming, which defeats the purpose of consolidating layers of devices. Similarly, if you route SONET Virtual Tributaries (VT), you'll either need to continue to groom or route them in your converged layer or restructure your network to avoid VT handling.
Another issue that can affect the value of opto-electrical convergence is the management interfaces and features available with the converged configuration.
At one level, it might be tempting to look for devices that perpetuate the independent management of optical and electrical paths because that would preserve the management connections into the current OSS/BSS elements. If you do that, however, you may limit your operations benefits.
Since virtual and physical device management is essentially the same, it is often better to think in terms of transport path management, where a transport path is a logical construct that can be mapped to an optical, electrical or combined route set.
Full operating scale
This is almost certain to require some modifications to network management systems (NMS) and even to some elements of OSS/BSS, but the result should be simplified provisioning and problem determination.
Problem determination and resolution is also an issue to consider in your opto-electrical convergence plan. To the extent that a converged device presents isolated interfaces to accept electrical and optical connections, it is inevitable that it perpetuates the management of not only those interfaces but also the transport layer they connect with (Ethernet, for example).
While you have "Ethernet transport" in an independent sense, you'll need to be able to manage converged devices through their Ethernet role, and the same will hold for optical-layer services. This illustrates that opto-electrical convergence will tend to lose its value if it's done on a small scale where the elements remain buried in independent optical and electrical sub-networks.
Because of this, projects to converge the layers should probably be considered only if the scope of the project is large and the project will quickly achieve its full operating scale.
Operators have reported that their success in converging the optical and electrical layers is greatest where they have selected vendors in both layers that provide both capabilities, even if the capabilities are not currently converged.
Independent optical vendors or Ethernet vendors have less to gain by convergence, so they may either not support it fully or even attempt to discourage it. Vendors that offer independent optical- and electrical-layer products, as well as converged products, are the ones most likely to be able to provide useful guidance and support for your convergence project.
Tom Nolle is president of CIMI Corporation
This article originally appeared on SearchTelecom.com