Future-Proof Your High-Speed Data Access Network
Ensuring network flexibility is difficult because capital for layered networks is no longer freely available. But standards-based media such as GigE, DWDM, RPR and POS will add intelligence and resiliency to your net.
New revenue opportunities for MSOs are driving ever-increasing demand for network scalability and flexibility. Cable engineers must understand and account for services with highly complex, but not yet fully defined requirements, including possibilities such as voice over Internet protocol (VoIP), premium content distribution, video over IP and managed commercial services. Where then should engineers focus their efforts? How do you design for services that haven’t yet been developed? These are the questions ringing throughout the cable industry today. The challenge is that, while network flexibility is the obvious answer, achieving it can be difficult because capital for layered networks is no longer freely available.
The primary emphasis of this discussion is the core network infrastructure that transports data applications across the various hubs, headends and systems. Today’s flexibility demands center around platform choices that can grow with service demands, feature sets that can reliably support voice, video and data, and routing protocols and switching technologies that provide granular control. These are the tools that will help carry your services and systems into the future.
1. Choose a flexible platform
Platform choices abound—while one system may choose a Layer 2 optical solution such as synchronous optical network (SONET) or dense wavelength division multiplexing (DWDM), a similar system may deploy packet-optimized gigabit Ethernet (GigE) or packet over SONET (POS) technologies. Considering today’s service offerings, both approaches can fulfill the requirements.
However, looking into the very near future of rising Data Over Cable Service Interface Specification (DOCSIS) bandwidth demands, SONET and POS solutions tend to be prohibitive both monetarily and in terms of fiber consumption. GigE is the obvious cost leader, while DWDM has tremendous fiber utilization efficiency. The existing DOCSIS demands alone may drive an early decision in one direction or the other, but making that immediate decision may preclude flexibility in the future because of scaling limits in bandwidth and switch processors, or make introduction of new service connectivity media cost-prohibitive.
2. Plan for VOD
Introduce video-on-demand’s (VOD) bandwidth-hungry distribution, and a system may very well prefer Ethernet’s cost advantages. Distributed and cached VOD architectures demand gigabit speed throughput throughout the secondary hub rings interconnecting storage facilities with the distributed streaming servers.
Centralized servers make economic sense with a transport infrastructure that maximizes the asymmetric traffic patterns such as uni-directional GigE transmission from the primary hub to secondary hub quadrature amplitude modulation (QAM) modulators. As a standalone service, no return path is required thus halving the transmitter costs. Redundancy also is not a requirement in most systems. In the final design decision for VOD, operational costs and fiber utilization will tip the balance in favor of a layered multiservice network versus dedicated infrastructure.
3. Don’t overlook voice
VoIP’s quality of service (QoS) requirements drive a need for additional resilience and intelligence, thus bringing DWDM and core routers back into the forefront. Incorporating redundancy into your networks is easier if accounted for during the design phase. Ethernet-based architectures tend to provide complex redundancy schemes and slow convergence. Optical solutions, either SONET or DWDM, provide excellent recovery in sub 50 millisecond timeframes, but with a price tag that reflects that capability. Router-based solutions, such as POS or resilient packet ring (RPR), can provide a configurable and predictable architecture that lends itself to the QoS requirements of voice such as availability, deterministic latency, prioritization and admission control.
Residential users’ applications, such as peer-to-peer file sharing and gaming, demand edge intelligence, packet inspection and admission control. Access routing capabilities, whether incorporated into cable modem termination systems (CMTSs), aggregation switches or edge routers, must begin to include the deep intelligence to institute effective traffic engineering.
Today’s networks are in free-run with minimum traffic engineering mechanisms and no QoS. Systems designed for best-effort service are being pulled into the future of content delivery from the subscriber-side out. Operators can turn this train around with the addition of intelligent edge control. With the insight and capabilities available from new packet inspection engines and features, operators have the tools at hand to identify, classify and control network traffic. This engineering and analysis not only will manage traffic loads to subscriber requested services, but will, as a benefit, greatly extend the life of existing CMTSs, aggregation routers and Internet transit.
4. Consider business sector needs
Looking at the commercial sector, business services often require symmetric throughput, monitoring and reporting capabilities beyond the existing infrastructure’s capabilities. Enterprise usage more closely resembles the standard network paradigm of constant and equal utilization of bandwidth both upstream and downstream. This situation creates a demand on existing residential equipment, such as DOCSIS-based systems, that they are not designed to accommodate.
Commercial customers need flexibility to increase services beyond standard best-effort data services. Time division multiplexing (TDM)-based voice services such as T-1 based PBX trunks, cellular backhaul services and Centrex services all provide value-added revenue streams. Operators can efficiently provide these services over today’s and tomorrow’s networks with a little planning in the choice of scalable transport network platforms and utilization of aggregation devices in the hub that can accommodate a wide variety of input technologies such as 10/100/1,000/10,000 Ethernet, OC-3 to OC-192 POS and RPR.
5. Understand security requirements
Traffic partition and security also are typical requirements in commercial applications. These can be achieved using dedicated, but costly, optical access infrastructure or with Ethernet-based virtual local area network (VLAN) services. The barrier-to-entry for most systems is that many potential customers can’t economically justify fiber builds. Coax-based networks can ease an operator’s entry into this market. Ease of service provisioning, both new and modification of existing, is an important factor in commercial service offerings while deciding between these approaches.
Recognizing that different services demand different characteristics from the network is the key to planning for the introduction of new applications. If we take the previous scenarios, we find that our network demands have increased from our base digital video and DOCSIS data networks to a converging set of requirements that lend themselves to supporting multiple services.
6. Build redundancy
Multigigabit level transport networks capable of voice-level redundancy with packet differentiation and prioritization point to DWDM fiber management with core high-speed router intelligence, based on one- and ten-GigE and OC-48/OC-192 RPR and POS. The edge of the network has migrated from standard access rings to high-touch, aggregation switches and extended-feature CMTSs.
New access technologies, fiber- or coax-based, and new distribution technologies such as GigE QAM modulators add new capabilities within the confines of developing networks. In between, we have the ever-present hazard of operational costs that drive simplicity and integration of resources and skill sets of employees. The additional cost of common infrastructure can find balance with the operational savings. Accounting for these types of emerging technologies that support tomorrow’s IP-based applications is a significant portion of today’s network design philosophy.
High-speed core infrastructure with intelligence and resiliency is the foundation for all services, DWDM for fiber management, and RPR or POS for the intelligence and flexibility. Access and aggregation points distributed throughout secondary hubs are key to high-touch services. Be sure to ensure that the aggregation devices use standards-based media such as one- and ten-GigE, RPR and POS, which provide flexibility to future technologies such as fiber- or coax-based enterprise access. Focusing design efforts around this basic methodology will permit the network to grow and scale with bandwidth demands as well as increase in intelligence and features for future applications.
Ahmet Ozalp is vice president of marketing at Narad. Email him at [email protected].
Did this article help you? Send comments to [email protected].
Smart Planning for Your Net’s Future
New access and distribution technologies can add new capabilities to developing networks. When planning for your network, it’s essential that engineers account for these types of emerging technologies that support tomorrow’s IP-based applications.
A high-speed core infrastructure with intelligence and resiliency is the foundation for all services, DWDM for fiber management, and RPR or POS for the intelligence and flexibility. By ensuring that your network’s aggregation devices use standards-based media such as one- and ten-Gigabit Ethernet, RPR and POS, you’ll have the flexibility to deploy future technologies such as fiber- or coax-based enterprise access. Such a design focus will permit your network to scale as bandwidth demands grow, and increase in intelligence to support future applications.
Six Network Planning Tips
BOTTOMLINE
When planning your high-speed data access network, be sure to consider these six essential elements.
1.Choose a flexible platform.
2. Plan for video-on-demand.
3. Don’t overlook voice.
4. Consider business sector needs.
5. Understand security requirements.
6. Build redundancy.
