By Ronen Mikdashi
When it comes to winning the battle against competitors, cost optimization is an essential part of an operator’s strategy. Such a strategy must examine costs and requirements at both the macro and micro levels.
Winston Churchill once said that victory should be achieved at all costs. That might be true at times, but when it comes to telecom, it has long been known that expenses are a critical part of any competitive supremacy. No operator can win the market battle without a cost-effective business model.
With network infrastructure as one of the biggest budget items for operators, getting the right cost structure to support such a critical investment is not always a given. Even harder is choosing the best practice to optimize the financial investment in the network—the strategy that will reduce the total cost of ownership while maintaining, or even improving, network performance, both on a macro level (as converging network assets) and the micro level (as reducing the cost of components).
Optimizing the macro level
The design of any optical network affects its return on investment (ROI). This fact is crucial in any service provider’s approach to its business model. For optimum cost-effectiveness, operators have to factor in two main goals that will affect the network design and, eventually, its cost:
- Increase network flexibility: By making the network flexible enough to adapt to any new service request, at any capacity and any path, operators can reduce time-to-market. As a result, they can also minimize the cost involved in infrastructure expansion or network reengineering
- Ensure a smooth transition to an all-IP next-generation network (NGN) infrastructure: In addition to supporting current requirements, the network design also must support future NGN services. This capability vastly increases the long-term efficiency of the network and reduces costs by leveraging past investments
With this in mind, an operator needs to examine four building blocks that meet the demand for architectural optimization.
1. Converged packet optical networking platform (PONP; also known as a packet optical transport system, or P-OTS). To realize its full benefits, convergence should be implemented at various levels within the network:
- Hardware-based packet-optic convergence in a single platform provides both data (Carrier Ethernet) and wavelength (WDM, Optical Transport Network, reconfigurable optical add/drop multiplexing) services, along with traditional circuit services (SONET/SDH based), on a single space-efficient platform
- Management convergence eases operations and reduces operational expenses (opex). It is important to choose a unified management system that enables management across all network layers and technologies—optical, Ethernet-based, and SONET/SDH—and provides end-to-end access to long-haul provisioning with a single system
The value of convergence lies in its integration and in the flexibility it gives service providers, enabling them to implement any service almost seamlessly and any technology at any place and any time within a unified transport platform. The major benefits of simplified management and reduction in service-provisioning time and network complexity complement the more expected cost savings associated with reduced spare parts training and space requirements. Projections based on preliminary field data from users of PONPs indicate an average capital expenditure (capex) savings of approximately 25% and an estimated drop in total cost of ownership (TCO) of about 45% over a period of five years.
2. Higher-speed wavelengths—40G and 100G. The sustained high growth of IP-based traffic, as evidenced by the increased popularity of real-time video and high-speed Internet services, has important scaling implications for optical networks. The implications are twofold:
- As capacity grows, core router-to-router links are over-burdened, demanding increased wavelength capacity
- Capacity expansion also intensifies fiber capacity limitations. This issue is addressed by either adding on more 10G wavelengths per fiber or increasing the bit rate per wavelength
The cost reduction of implementing 40G and 100G links comes in various areas. The first area lies in the fact that deploying higher-speed channels often means deploying fewer wavelengths per link and, by that, easing operation of the network. A second area comes from the reduction in the cost per bit. The rule of thumb has been that a 4× data rate increase translates into a 2.5× price increase for transponder cards. The same rule is expected for the transition from 10G to 40G.
3. Edge-to-core ROADM portfolio. Whenever flexibility and fast presentation of new services is needed, reconfigurable optical add/drop multiplexers (ROADMs) have become the optimized building block. Provisioning any new service is an easier task in a ROADM-based network. New service deployment is done automatically and remotely without the need for technical staff at tandem sites.
Today, there are various technologies for implementing flexible ROADMs. Each technology is suited to a different topology (mesh, ring), with specific capacity and complexity requirements.
When flexibility is required in ring-based networks, the two-degree ROADM is considered the technology of choice. Two-degree ROADMs cost less than mesh-topology ROADMs. The preferred technology for two-degree ROADMs is the planar lightwave circuit (PLC), which optimally fits budget limitations while fulfilling the bandwidth and flexibility requirements within that segment.
On the other hand, when mesh topology is needed—for example, in core networks—wavelength-selective-switch (WSS)-based ROADMs are considered the preferred technology worldwide, adding flexibility and lowering TCO for various core and regional network scenarios. WSS ROADMs enable carriers to route any wavelength (or any combination of wavelengths) to any node without the need to predefine traffic demands or install additional devices. For an operator, this significantly reduces time-to-market for new services.
A recent deployment of an Asian ILEC is a perfect example. A regional WDM network was designed twice: The first time was based on WSS ROADMs, and the second was based on traditional fixed DWDM. The mesh network had a nodal degree from two to four, and a detailed capex comparison revealed a counterintuitive result: Although the initial cost of the ROADM-based network was 13% higher than a traditional fixed network, once it was populated with only six wavelengths per link, it became more cost effective.
4. Automated planning and design tools. One of the burning issues confronting network operators is how to maintain network efficiency not only at the deployment stage but also for the duration of the network lifecycle. For this purpose, operators have begun to adopt a suite of real-time automated planning tools to analyze network problems and provide solutions for network optimization. Single-point-of-failure analysis, elimination of bottlenecks, and designing a smooth transition to NGNs are only some of the capabilities that planning tools offer, reducing long-term capex and opex in a variety of ways.
To better understand the benefits of planning tools, let’s review the recent example from a large Asian mobile operator.
With more than 730 network elements and 15,000 different services, the network was based on multivendor platforms, incorporating Ethernet, SONET/SDH, and DWDM technologies. The network had reached its expansion limit with an average of more than 80% utilization per link. There was almost no capability for expansion. However, once planning and optimization tools were implemented, the operator was able to reduce network expansion costs dramatically—and ultimately managed to expand the network 50%, without any addition of platforms or capex.
Further, the use of planning tools reduced the number of single points of failure (where both the primary and the protected service go through a shared link) along the network by 80%. The tools also helped minimize the number of hops and reduced bandwidth utilization, as shown in the table.
The micro-level consideration
Operators invest considerable effort in finding the most cost-effective products for their networks. Several issues dramatically affect these decisions and the subsequent optimization process. Operators need to ask them themselves which products have the biggest impact on the total cost of the network. And more important, what is the ceiling of the product’s price point?
Before dealing with the cost reduction of individual components, operators must understand the overall structure of their network costs and focus on the major building blocks that bring the highest level of improvement. The figure shows the cost structure of a typical core/regional optical network based on an analysis of ECI Telecom customers’ behavior.
It is evident from this analysis that the majority of capex investments in regional networks is spent on amplifiers and service cards. Therefore, it is recommended to initially deal with these costs.
The cost of service cards can be optimized by using higher-density cards. The price per port is reduced while the number of assigned slots is diminished, thereby minimizing the number of platforms in the network.
So what can we conclude? To win the battle, cost is a determining factor. It is evident that cost reduction has long become mandatory to survive and become a leader in the telecom market. The important takeaway is that any optimized cost reduction can only be achieved by a combination of both micro- and macro-level actions.
By first recognizing the major “currency consumers” in the network, operators can efficiently reduce total network costs with minimal time and expenditure investments. Best practices have shown that an optimized cost reduction is achieved by analyzing network performance and that of the network components themselves.
Ronen Mikdashi is a senior product manager at ECI Telecom (www.ecitele.com).
LIGHTWAVE: What Features Should Packet Optical Transport Equipment Contain?
LIGHTWAVE ONLINE: Advances in WDM and ROADMs: The Key to Your Success
LIGHTWAVE ONLINE:: OIF Issues 100G Framework Document