 | Issue: | Asia-Pacific II 1999 |
| Article no.: | 4 | |
| Topic: | Defusing the Network Time-Bomb | |
| Author: | Phil Griffin | |
| Title: | Photonics Development Director | |
| Organisation: | Marconi Communications | |
| PDF size: | 24KB |
About author
Not available
Article abstract
Never have the worlds telecommunications operators and the telecommunications networks they own and run faced such a challenge – and never has their equipment been under so much pressure as today.
Full Article
Businesses, from multinational corporations to Small Office/Home Office (SOHO) workers increasingly rely totally on telecommunications: without the telephone and PC resources, they simply cannot do business. The number of people connecting to the Internet is rising exponentially, demanding access to on-line services 24 hours a day across much of the globe. When a customer makes a phone call, the average length of time they are connected over the network is only a few minutes. In contrast, the average connection by a user to the Internet is at least half an hour. So it is hardly surprising that telecommunications carriers resources are being stretched – not only at the telephone exchange as individual lines are engaged for longer periods, but on the actual fibre-optic cable in the operators core network. Only a few years ago, fibre was viewed – perfectly understandably – as practically limitless in its capacity. This is no longer the case: as capacity requirements soar, the amount of fibre in the ground is quickly revealed to be both finite and falling short of demand. Fortunately, optical technology – or photonics has been developed. Photonics effectively gives near-saturated or exhausted fibre, and the carriers that own it, a new lease of life. For at least the past two years dense Wavelength Division Multiplexing (DWDM) has been considered a viable commercial option – in many cases a no-argument decision for long?haul transport of telecommunications traffic. It is now evolving into a networking technology in its own right. DWDM technology involves dividing the optical transmission capacity of a single fibre strand into multiple channels,each recognised by its optical wavelength. It splits the single laser-generated lightwave carried by an individual optical fibre into a number of wavelengths. Each individual wavelength is able to carry the same amount of information as the whole original single fibre. Instead of carrying, say, 60,000 phone calls, the same fibre can now transport 250,000 calls or their equivalent in data. Systems with more than eight wavelengths are commonly described as Dense WDM. As the complexity of providing ever more wavelengths increases, the development focus is shifting towards increasing the total aggregate capacity that can be carried over the fibre as the data per wavelength channel rate rises ? 10 Gbps, 20 Gbps and even 40 Gbps per channel. Existing carriers are adopting photonics systems to extend network capacity; emerging carriers are now building DWDM into their strategic plans, effectively betting their business on optical technology and looking to photonics to deliver dramatic technology improvements and revenue opportunities. The vast majority of DWDM implementations today are designed to deliver cost?effective point?to?point long-distance transport by increasing backbone capacity, so reducing bottlenecks. Given the choice of deploying new fibre or increasing the capacity of existing fibre through DWDM, there is little argument among those building or enhancing fibre capacity in long?haul networks. DWDM delivers more capacity quickly and, by comparison with other choices – laying more fibre or upgrading existing fibre – at an attractively lower cost. As photonics continues to mature, it no longer focusing simply on wringing more capacity out of point – to -point connections. It is now the catalyst for redefining the entire telecommunications network architecture. The transport layer is evolving to accommodate increases in bandwidth demand by introducing a discrete photonic layer. Conventional Synchronous Digital Hierarchy (SDH) technology provides a precedent for the remarkable development of photonics. It, too, began in long?distance networks. The turning point that enabled SDH to be universally accepted was the emergence of ring?based ADMs, enabling data channels to be added or removed electronically at particular nodes in the network as the capacity needs dictate. Marconi Communications has long been an innovator in fibre technology and the mechanisms for transporting information over the fibre-optic cable. It enjoys 38% of the global market for SDH multiplexers – the foundation on which the worlds carrier networks is built and an essential, complementary technology to tomorrows photonics networks. Marconi has currently developed its flagship SmartPhotonix range of optical switching products, including an Optical ADM. SmartPhotonix systems offer a get out of jail card to hard-pressed operators. Moreover, it can support all types of network – from point-to-point systems to ring-based networks and deliver the benefits of optical transmission at national, regional or campus levels. The SmartPhotonix range includes optical multiplexers targeted at long-distance, point-to-point line systems, and optical amplifiers, possibly the best of their type in the world to augment the signals over long distances. SmartPhotonix systems also adhere to all relevant international standards as they evolve and are integrated with existing Marconi SDH management systems. SDHs inherent protection mechanisms, for example, are available at the optical layer through the SmartPhotonix family. The regional layer or metropolitan levels of a carrier network, moving towards the access layer,demands greater flexibility than the long-haul. Ring?based Optical Add/Drop Multiplexers (OADMs) enable individual wavelengths to be added/dropped in similar fashion – in the case of Marconis PMA?8 reconfigurable OADM, this can be carried out remotely via a Windows-based management terminal. The add/drop function, whether traditional Time Division Multiplexing (TDM) or emerging DWDM, is to provide access to part of the multiplexed traffic at an intermediate point along a route. The rationale for an ADM is that it is cheaper to provide an add/drop function than to go through the full multiplexing/demultiplexing process: this is as true for DWDM as for SDH. A conventional electronic ADM adds, drops or terminates time?slots in a TDM frame. Each frame is divided into a number of timeslots, the number depends on the size of the circuit. There is a direct analogy in the optical world: adding, dropping or terminating wavelengths in a wavelength spectrum, with specific advantages in using photonics at high speeds. By being able to select one or more specific wavelengths, an OADM provides cost-effective access to a small portion of the DWDM traffic passing along a fibre. Each wavelength may be carrying one of a range of payloads, such as Asynchronous Transfer Mode (ATM) or SDH. Typically, the payloads of the accessed wavelengths are terminated within a single fibre hop from the OADM. In the case of SDH, the timeslots of specific Virtual Containers (VCs) are then accessed in turn. There are two main classes of OADM: serial and parallel, though hybrids are also possible, and each can either be passive or dynamic, giving a range of technology options, each with its own advantages. For example: Each filter of a serial OADM is fitted, in turn, into the fibre as needed. An advantage of a serial OADM is that the operator fits only the number of devices necessary, so minimising through loss compared with the costs and loss of pre?fitted OADMs, with filters not being used or later needing to alter the configuration and having to cut the fibre anyway. Parallel OADMs can be manipulated independently of one another, so the decision to add/drop one wavelength can be taken without disturbing the remaining channels. Changes to individual channels of a passive OADM must be made manually by an engineer on?site Individual channels of a dynamic OADM can be switched remotely. Marconis PMA?8 is a dynamic parallel OADM. Dynamic OADMs necessarily involve complex engineering, not least because photonics relies on cascading analogue functions that become less clear every time they are interfered with, in contrast to working with digital signals. The emergence of truly flexible switching in OADMs will enable operators to create virtual networks in which connectivity is not defined solely by the physical topology of the network. Whatever the demands on the network ? hubs, Virtual Private Networks (VPNs), point? to?point, ring?based DWDM can support all types of service topology, enabling the operator to create multiple separate logical networks over the same physical fibre. In the optical domain, the fibre is divided into portions of the spectrum, wavelengths, instead of portions of time. The only difference is that a wavelength will contain the equivalent of a very large number of timeslots. Ultimately, both timeslots and wavelengths will become simply manageable entities, representing a load that a customer is paying for and one that can be moved about with equal alacrity as icons on the operators network management screen. Although the technology enabling Optical Cross-Connect (OXC) deployment has yet to mature, OADMs are commercially available, so carriers and larger enterprises can take their first steps in optical networking beyond proven point-to-point long-distance DWDM connections. The likelihood is that, longer-term, the core of these new optical infrastructures will comprise high-capacity OXCs, supported by collections of rings of OADMs. The rings will therefore provide vital flexibility as well as transport functionality. OADMs are passing through several evolutionary stages, taking them progressively towards the functionality of an OXC – just as rings of SDH ADMs have evolved to act as small DXCs. While some operator networks are better suited to rings and others to OXCs linked point?to?point, Marconi Communications approach is that, whatever the individual topology, the way forward is undoubtedly a migration path towards optical networking. New operators, for example, may build the network incrementally in line with business demand through adding successive OADM rings. They can generate a revenue stream without the significant investment necessary for an OXC, while enjoying a combination of high?volume transport, flexibility and resilience. Existing operators may prefer to introduce OXCs, exploiting the existing transport infrastructure by adding these high?capacity optical nodes to the core network: an OXC or SDH Cross?Connect is equivalent to a ring collapsed into a single sophisticated device. It is likely that, typically, an OXC will act as parent to a cluster of rings for ease of management. It can be thought of SDH thinking grafted on to the optical layer. The advent of fully featured, manageable OXCs will enable operators to carry out end?to?end wavelength routing using DWDM. And as the prospects grow for carrier-grade quality transmission of Internet Protocol (IP) on SDH and IP on DWDM – two examples of this anticipated simplification of layers in an operators network – photonics matrix switches will appear for high-capacity packet switching based on all-photonic switch architectures. Also, by managing and switching traffic carried in optical wavelengths, carriers can simplify their networks, one or more network layers can be by-passed, to reducing the number of elements, management and associated costs. The advent of photonics is a welcome addition to existing network transmission technologies, and most definitely a complementary, not competitive development that will operate alongside and integrate with SDH systems for many years to come. Generally, there are decided benefits for an operator adopting photonics technology when the data rate is high and large volumes of traffic have to be switched. It remains more advantageous to use SDH when the system granularity is, in comparison, far lower and traffic does not have to be broken down into yet slower data rates. Conclusion Photonics is still at a very early point in its evolution ? as much an art as a science, just as SDH was a decade ago. But it holds the key not only to huge capacity, but a capacity that can be manipulated and managed cost?effectively. Only then can its full potential be exploited, enabling new and existing carriers not only to defuse the telecommunications time-bomb, but also to provide a cost-effective migration path to tomorrows high-capacity carrier networks. Photonics looks set to hold a pivotal position in the pantheon of networking technologies.
