Integrating WDM & Sonet/SDH

Reducing complexity in transport networks * Why? * What? * How?

February 19, 2004

21 Min Read
Integrating WDM & Sonet/SDH

The arguments for integrating the different layers of transport networks read like a recommendation for Apple Pie and Motherhood:

  • Argument No. 1: It reduces capital costs and improves reliability.
    A single transport layer simplifies networks, resulting in fewer network elements and higher networkwide reliability. (Three cheers for a great theory; let's see whether it works out in practice.)

  • Argument No. 2: It creates opportunities to generate new revenues and reduces operating costs.
    A single management system with automated provisioning software simplifies network provisioning and allows more opportunity for rolling out next-generation services. (Thunderous ovation.)

  • Argument No. 3: It reduces churn by keeping customers happy.
    It makes it easier to offer service-level agreements (SLAs) and improves the operations, administration, and maintenance (OAM) functions that underpin customer service. (Mass delerium!)

But, of course, it's not quite that simple. There’s a bunch of questions surrounding the issue of how integration should be tackled:

  • Question No. 1: What technologies to integrate?
    The choices include Sonet/SDH, the new-generation data-friendly Sonet/SDH, Gigabit and 10-Gigabit Ethernet, and the raw optics of WDM.

  • Question No. 2: Where to start?
    The options include adding WDM to existing Sonet/SDH gear, adding Sonet/SDH to WDM gear, or deploying completely new integrated equipment.

  • Question No. 3: How to avoid disrupting existing customers?
    Most carriers also have multiple-vendor equipment at each layer, and typically the Sonet/SDH and WDM layers have different vendors, anyway, so one-shot integration of existing kit is pretty much a fantasy. That leads to some form of gradual integration as the only realistic option.

And this being telecom, vendors have created a new set of acronyms to cover the issue of integration. The Multiservice Provisioning Platform (MSPP) has been around for some years to describe (largely Sonet/SDH-based) devices that provide efficient low-speed service provisioning and aggregation. But now carriers need to come to terms with the MSxP series, whose members include the MSSP (multiservice switching platform) and the MSTP (multiservice transport platform) (see The Money's on MSPPs).

Fortunately, the underlying functions of MSxPs are closely aligned with classic telecom functions such as aggregation, switching, and transport, but reinterpreted for the integrated environment. However, different vendors, of course, always see things in slightly different ways, so in practice carriers have quite a choice in how they can approach transport integration.

That’s where this report comes in. After setting the scene on what is driving transport integration and what it means for network architectures, it looks at how four vendors are implementing transport integration in different ways into their metro product families, and how they believe carriers can benefit from their approaches. So to separate MSSPs from MSTPs, and to get a feel for the practical problems and deployment options that carriers have for these new technologies, read on.

Here’s a hyperlinked contents list:

  • Integration Drivers
    Carriers need smart capex to drive revenue growth

  • Delayering the Network
    Three approaches to simplifying the network – and three types of platform

  • Approach 1: Add WDM to Existing Sonet/SDH Kit
    How Cisco and Photonic Bridges do it

  • Approach 2: Add Sonet/SDH to WDM Kit
    The converse approach, with a novel twist by Internet Photonics

  • Approach 3: Deploy New Integrated Equipment
    WDM is not like Sonet/SDH, so Photuris plumps for integrated optical transport

  • Integrated Management
    Where everything has to converge, but often doesn't yet

To view the Light Reading Webinar on which this report is based, click here.

There's interesting background to Sonet/SDH and WDM issues in the Light Reading archives...

  • Metro WDM: What Carriers Think

  • Metro DWDM

  • Metro WDM Economics

  • Optical Gateways: The Portal to Profitability

  • Metro Multiservices Evolution

...as well as the Beginner's Guides...

  • Wavelength Division Multiplexing (WDM)

  • Sonet (Synchronous Optical NETwork) and SDH (Synchronous Digital Hierarchy)

...and recent reports from Heavy Reading: The Future of Sonet/SDH and Multiservice Provisioning Platforms: Empowering the Metro Edge.

– Michael Howard is Principal Analyst and Cofounder, Infonetics Research Inc.

The combined market for metro WDM and metro Sonet/SDH equipment is pretty substantial – nearly $1.9 billion in 2004, according to Infonetics Research Inc. (see Figure 1). Of course, most of this is not yet sold as integrated kit, but Infonetics believes that the total addressable market for metro integrated WDM and Sonet/SDH would not be far short of this, should carriers take the plunge.

45305_1.gifAnd there are compelling business reasons why they should, summed up in one word: revenues.

No business has ever returned to growth through cost cutting alone, so incumbents have to find new revenues. There are a number of areas where new revenues are available to carriers, both in retail and wholesale. These include:

Retail services for enterprises and small/medium businesses (SMBs): Large enterprises are looking for a means to move to higher-rate gigabit data services, and are tending to turn to competitive service providers or to dark fiber and self provisioning. For incumbent carriers this is both an opportunity and a threat to billons of dollars of forgone service revenues. And SMBs are looking to migrate from Frame Relay rates to higher-rate megabit data services. Again, competitive service providers, municipalities, cooperatives and MSOs are moving in to the fill the void.

And there is no denying the spectacular rise in carrier Ethernet offerings. Infonetics Research is currently doing a worldwide study of carriers and is finding that just about everybody is, or will be, offering Ethernet. Most of the major carriers will certainly be offering it by early 2004, and it has become an absolute must for metro optical equipment.

Wholesale: Carriers’ carrier and backhaul opportunities abound for higher-bandwidth services if certain technical problems can be solved in the transport of Sonet/SDH as services to other carriers.

A basic point of integrating WDM and Sonet/SDH is that it promises both to support such new-revenue opportunities and to help carriers control costs. But there is a twist. Carriers are already used to the idea of success-based capex – cutting capital expenditure (capex) to the bone (around 14 to 18 percent of revenues), and then tying network growth plans to actual revenue and subscriber growth by making incremental expenditures as required by service growth. A further argument for WDM and Sonet/SDH integration is that it is also a form of smart capex – capex that also cuts operational expenditure (opex) through automation, while allowing the carrier to offer new services quickly.

Since opex is typically about 55 to 60 percent of total revenues, it is a juicy target for profit-starved carriers, and cutting it works wonders for total cost of ownership and other metrics beloved of accountants – but most especially profitability.WDM and Sonet/SDH integration has such a big effect in lowering opex because it results in a fundamental simplification of the network architecture: fewer separate networks, fewer network layers, and (therefore gratis) fewer network elements to provision, maintain, and manage.

Currently, most carrier networks are layered along the lines of Figure 2. This is the legacy situation of most public networks, where a complex succession of different network technologies supports the range of end-user services. It suffers from inflexibility, inefficiency for modern data services, and high opex. Not surprisingly, carriers are looking for something better.

45305_2.gifMany think that they have found it in the structure of Figure 3. This is the next-generation two-layer network, basically a single optical transport layer over which runs an IP/MPLS multiservice layer. From the perspective of this report, the key fact about a single optical transport layer is that it has to combine the elements of both WDM and Sonet/SDH.

45305_3.gifAn ad hoc poll conducted during the Light Reading Webinar on which this report is based showed that about a third of the carriers in the audience had already started combining WDM and Sonet/SDH transport layers in the same equipment. And quite a few more are looking to do so in the next few years.

But few believe that carriers are going to reach this network nirvana in one easy step, as there are huge practical and financial issues involved. A difficulty in deploying a new integrated optical transport layer – considerations of cost aside – is that carriers don’t want to disturb their existing customer revenue services. Most carriers also have multiple-vendor equipment at each layer, and typically the Sonet/SDH and WDM layers have different vendors anyway, so one-shot integration of extant kit is pretty much a fantasy. That leaves some form of gradual integration as the only realistic option.But this can still be done in several ways, namely:

  • Adding WDM to existing Sonet/SDH gear

  • Adding Sonet/SDH to WDM gear

  • Deploying new integrated equipment

And there are general questions that any specific approach must address. How best to stop the expansion of the current multilayer network? And where should the carrier start the integration? For example, is it better to start with collector rings or the metro core?

Adding WDM to Sonet/SDH kit is espoused by vendors such as Cisco Systems Inc. (Nasdaq: CSCO), Nortel Networks Corp. (NYSE/Toronto: NT), and Photonic Bridges Inc. Naturally, it’s a popular approach with vendors like these that have built a multiservice strategy on Sonet/SDH. This section looks at how two contrasting vendors – industry bellwether Cisco Systems and the younger, China-based Photonic Bridges – are enhancing Sonet/SDH with WDM.

Cisco Systems

“The ONS 15454 has been a very successful product for us over many years, and this is really the vehicle that we are using to put DWDM into metro and regional applications,” says Greg Nehib, Manager, Optical Product Management Group, Cisco Systems. “Our experience over many years has been that the evolutionary practices usually go out over revolutionary approaches. Greenfields are somewhat a thing of the past, and the evolutionary approaches seem to be more easily introduced into large-service-provider environments.”

The ONS 15454 architecture (Figure 4) provides three core functions:

  • Sonet/SDH transport with TDM aggregation

  • Native Ethernet, IP switching, and aggregation

  • Intelligent DWDM transmission and wavelength transport

45305_4.gifAccording to Nehib, a key feature of the approach is the way routing and switching are directly loaded onto the WDM layer via ITU Grid optics from some of the high-end routers and switches. And, by building WDM on the existing Sonet/SDH platform, it preserves the same sparing, the same electrical system, and the same user interface.

The evolutionary approach causes Cisco to talk in terms of a development of a series of MSxP network elements (Figure 5), where x began as provisioning (MSPP), then embraced switching (MSSP), and finally transport (MSTP). While all have the common capabilities of multiservice support (voice, data, and Ethernet); support for linear, mesh, and ring topologies; and integrated element management, there are fundamental differences in network application.

45305_5.gifSo the ONS 15600 MSSP is focused on the carrier core network. Effectively, it is an ONS 15454 with 10 times the port capacity and no electrical interfaces. Instead, it has a very large STS fabric (and soon a very large VT fabric), and integrates ADM and broadband digital crossconnect (BBDXC) functions, providing a multiservice switch fabric that is strictly nonblocking – and also fast (<25ms protection switching).

The most recent element is the ONS 15454 MSTP, which integrates DWDM into the platform and maps multiple services/interfaces in individual wavelengths to provide scaleable optical transport from tens to hundreds of kilometers.

Says Nehib: “Some of the characteristics of an MSPP and MSTP can both reside in the same shelf in a 15454, so you could have OC3/12 cards going through an STS or VT matrix, and then launched directly into a 32-channel hub, with a lot of that functionality residing in the same shelf.”

Photonic Bridges

Photonic Bridges is also following an evolutionary approach, but stresses that the eventual addition of WDM was designed into its MSPP architecture from the start.

Says Lucas Hsu, VP Product Management, Photonic Bridges: “Our particular focus is the metro market, which we see as the bottleneck between the large backbone network capacity and the large-enterprise needs. This is where bandwidth demand and traffic patterns change dynamically, which drives the need to make networks more flexible. By this we mean the ability to provision WDM wavelengths as well as traditional TDM circuits and data services.”

However, Hsu does concede that, in China at least, end-user demand for wavelengths in the metro market is still low. But, in the Chinese market, MSPPs are displacing traditional SDH gear at the big service providers, such as China Telecommunications Corp. (NYSE: CHA), which Hsu says has required the use of MSPPs; and these service providers are vigorously testing and demanding advanced interworking of MSPP capabilities.

The company’s MSPP architecture (Figure 6) uses multiservice interface planes (TDM plane, IP/Ethernet plane, ATM plane, and lambda plane) and dual switching planes (TDM plane and data plane). Data streams are processed with Layer 2 switching, GFP encapsulation, and VCAT and LCAS for improved transport efficiency.

45305_6.gifThese technologies are now well-established. Photonic Bridges and 16 other equipment providers, including Alcatel SA (NYSE: ALA; Paris: CGEP:PA), Huawei Technologies Co. Ltd., and Lucent Technologies Inc. (NYSE: LU), have passed their interoperability testing for GFP, LCAS, and VCAT in a recently completed MSPP bake-off at China Telecom.

The WDM interfaces use pluggable optics that allow lasers to be swapped out without removing the card, a capability that reduces inventory costs, and also allows a card to support a variety of capabilities concurrently, such as a 1310nm Sonet/SDH laser and a 15xxnm WDM laser.

“But the real feature of this design is the application of dynamically reconfigurable optical add/drop muxing,” says Hsu. “This moves the network from a static, engineered network to a provisionable, dynamic network, allowing a single-seat operator to change the network to meet the unpredictable nature of traffic growth and service demand. This is the ideal complement to the crossconnect capabilities of Sonet/SDH, allowing control from low-speed circuits all the way to high-speed wavelengths.”

Adding Sonet/SDH to WDM kit is the flip side of the coin, and a natural approach for vendors with a strong interest in WDM, such as Alcatel, Ciena Corp. (Nasdaq: CIEN), and Internet Photonics Inc. This report focuses on just one company in this class – Internet Photonics – which has an interesting twist to this approach.

“We are articulating a slightly different value proposition from what you typically hear from most Sonet/SDH vendors,” says Gary Southwell, VP Marketing, Internet Photonics. “We went in to try to solve the revenue problem – that is, to provide high-speed services that generate new revenue. In particular we focused on the gigabit-level services. And we wanted to do one thing that was done by the existing Sonet/SDH services very well, and that is to provide service-level agreements.”

He points out that, typically, services with SLAs can command a tariff premium of up to 60 percent or even 100 percent over best-effort transport – and this would go a long way to helping service providers’ currently distressed finances.

However, the company argues that mapping Ethernet and other packed data direct into Sonet/SDH payloads via schemes such as GFP becomes progressively less efficient as the bandwidth of the packet streams increases much above 100 Mbit/s, because of the increasing mismatch with the size of Sonet/SDH payloads available. Gigabit Ethernet, in particular, doesn’t fit very well. So, while GFP, VCAT, and LCAS solve the current megabit packet-service issues, they don’t really solve the high-speed data services that most carriers are looking forward to offering in the next couple of years.

Also, GFP, VCAT, and LCAS do not, of themselves, create an end-to-end solution, as this still requires an end-to-end OSS capability for provisioning.

Another issue with a pure Sonet/SDH architecture is that jitter and timing (clock-recovery issues) prevent one Sonet/SDH signal from being encapsulated within another Sonet/SDH signal for transmission across an intermediate network operated by a second carrier. So wholesale carriers' carrier applications require another solution beyond Sonet/SDH.

The company’s solution to these problem is to use what it terms intelligent wavelengths to unify transport (Figure 7). This allows new services to ride on separate wavelengths alongside the existing Sonet/SDH services, which remain unchanged, and allows the reuse of lit as well as dark fiber.

45305_7.gifThe basis of the approach is the ITU-T G.709 Digital Wrapper (Figure 8). Effectively, high-speed data services and carriers' carrier traffic is wrapped into large, managed transport frames that are carried on different wavelengths and maintained hop by hop. Wavelengths can be added/dropped and switched end-to-end.

45305_8.gifSays Southwell: “This allows us to take existing Sonet/SDH services from various carriers and map them into a wavelength payload. And we have added some intelligence here. There’s a header and a tailer added to the wavelengths, basically chopping them into payloads, and this allows us to manage each of these wavelengths with our own management channel, and not have it disrupt the existing framed management channels that are inside the first-party carriers’ Sonet/SDH networks.”

The result is that the first-party carrier gets the managed service end-to-end transparently through the third-party carrier’s network, and the third-party carrier is also able to do its own level of management, with no interaction between the two.

Overall, Internet Photonics claims, this use of managed wavelengths gives scaleable service aggregation for carriers’ carrier applications, solves the Sonet/SDH transport issue, provides management visibility per customer, and gives a simple build-out as service demand dictates.

Interest in the use of G.709 Digital Wrappers generally is growing worldwide, and larger vendors (such as Lucent) are also promoting it. Internet Photonics says that it has two operators currently deploying these types of services; and though it is still early days in terms of demonstrating such aspects as interoperability, it believes that this would be fairly straightforward technically.

An additional attraction of G.709 Digital Wrapper is that its forward error correction effectively improves the link transmission budget by 3 to 5dB, which can translate into sizable engineering savings.

Deploying new integrated equipment tends to be startup territory. Examples of vendors following this approach are Lumentis AB, Photuris Inc., and Tropic Networks Inc.

Photuris provides a fully automated reconfigurable OADM-based optical platform (the V32000) with carrier-selectable integration of Sonet/SDH ADM functionality, mainly targeted at metro–regional interoffice (IOF) applications. The company says that it has taken this approach in response to carriers’ challenges in deploying Sonet/SDH and WDM in the same networks. These challenges included:

  • WDM systems are extremely manual intensive and costly to operate.
    Physical fiber jumpers are needed for each add/drop; and manual power tuning per wavelength is needed each time a channel is added/deleted (per-wavelength engineering).

  • WDM systems can strand fiber capacity at each add/drop point.
    Banded systems force three or four wavelengths to be dropped at each node, and can result in as little as 25 percent or 33 percent fiber capacity utilization (neatly known as banded strandwidth).

  • Stacked Sonet/SDH rings and overlaid networks are capital intensive.
    Each Sonet/SDH ring requires a fiber pair, leading to conduit exhaust, while Gigabit Ethernet and other nontraditional services need overlay networks.

“WDM is not like Sonet/SDH. The engineering, maintenance, and provisioning of WDM networks today is a significant departure from Sonet/SDH, where a network was engineered once at installation time, and you only visited the equipment either to add capacity to the system or to plug in a customer service,” says Paul Bonenfant, Chief Architect, Photuris. “If you translate this feedback into carrier requirements, what they were telling us was that they needed a Sonet/SDH-like provisioning and maintenance of the optical layer. In other words, make WDM systems a close as possible to Sonet/SDH.”

Like Cisco Systems, Photuris advocates a new class of multiservice platform – the multiservice transport platform (MSTP), which the company describes as a third-generation solution for metro/regional IOF needs. This provides an optical transport ADM (OTADM), comprising a WDM reconfigurable OADM (WDM-ROADM) integrated with a Sonet/SDH-ADM. It will handle the dynamic and unpredictable traffic patterns that characterize these networks, and is optimized for flexibility and pay-as-you-grow service strategies.

These new network elements provide, via full automation of the optical layer, efficient high-capacity service multiplexing and flexible transport for interfaces such as Gigabit Ethernet, 10-Gigabit Ethernet, and OC12/48/192.

Photuris argues that the key financial/performance metric in this environment is cost per second wavelength. Says Bonenfant: “This [approach] allows carriers to address dynamic and unpredictable traffic patterns, it provides a solution that is optimized for flexibility, it enables them to pay as they grow in a success-based fashion, and the metric now with the appropriate figure of merit for this portion of the network is the dollar per second wavelength – where you are eliminating the passthrough tax that you would otherwise have to pay for excessive electro-optic regeneration, while providing a fully automated provisioning solution.”

This is very different from first-generation solutions, which were essentially modified long-haul systems aimed at the core and optimized for point-to-point traffic, with the key metric being cost per gigabit–kilometer. And different also from second-gen systems, which were MSPPs with WDM bolted on and were aimed at the access network and optimized for lowest first costs, so the key metric was cost per port.

Such new integrated systems are now carrying live traffic in carrier networks. Figure 9 shows what Photuris believes represents the first commercial deployment of a reconfigurable OADM-based optical transport ADM system.

45305_9.gif Deployed by Verizon Communications Inc. (NYSE: VZ) in September 2003 for Texas A&M University and City of College Station, it uses the V32000 WDM-based transport system to carry protected Sonet and Ethernet services for a variety of applications – with the ability to grow for future capacity expansion for statewide applications and to interconnect to offcampus national research networks. Key requirements were the ability to manage WDM, Sonet, and data services; to route traffic without any re-engineering; to manage the uncertainty, not only of future locations, but of future offsite locations; and to manage the unpredictability and diversity of traffic.

Whatever approaches are taken to transport integration, there is one area at least where they have to converge – the management plane. This is where the issues of a multi-technology, multi-vendor, multi-carrier environment really come to the fore, become entangled in regulation, and will take a huge amount of detailed and painstaking effort to resolve.

Take GMPLS, for instance. This is widely touted as a future universal method of integrating optical wavelength switching and MPLS packet switching, and would, if adopted, have a massive influence on network-management models and systems. Yet vendors covered in this report have different attitudes and approaches to GMPLS:

  • Says Cisco’s Nehib: “From a Cisco perspective, we are not currently using GMPLS on the MSPP platform. We are using G.709 and have an optical service channel, and applications that include MSPPs obviously have the DCC as well. So there are multiple avenues for us to get there, but that’s not a focus currently for Cisco in the North American market.”

  • Says Photonic Bridges’ Hsu: “We are in the process of implementing GMPLS. We are moving towards a dynamic automatic network – ASON based – and obviously GMPLS is going to play a big role there. However, the end-user demand for that is going to take a while before it really develops. Wavelength services, although very useful from the service provider’s perspective, involve little end-user demand as we presently see it. So implementation is underway and standards are being worked on in the ITU, but until that is finalized, I think the deployment is going to take a while.”

  • Says Internet Photonics’ Southwell: “We are in the process of rolling it out. In the first steps we see that it makes more sense to do an OEO at every single hub because it gives you a little more flexibility, but as you get into the backbone and start to aggregate a lot of traffic types, it starts to make sense to be able to provision the wavelengths. So we have licensed technology from AT&T, and are moving along those lines, but it’s a phased approach. We believe you have to build the onramps first before you solve the interstate problem or the highway problem that connects those onramps.”

And in the U.S., at least, there is the ongoing complication of the RBOCs’ Osmine management process, which requires product-by-product approval of management compatibility. Questions have been asked over whether the Osmine environment can handle reconfigurability and whether it can handle the GMPLS control plane.

Fortunately, it doesn’t look as if optical reconfigurability will thus bust the OSS model for the RBOCs.

Says Photuris’s Bonenfant: “The real difference between provisioning a static network and provisioning a reconfigurable OADM-based platform with a GMPLS control plane is the use of single transaction language – one command, a Tier 1 command, for example – rather than having to go to multiple network elements, allowing the features and functions of the optical layer working with the GMPLS control plane to enact the provisioning.

“The short answer is that these systems can and will be used in RBOC networks in the context of an Osmine environment.”

Nevertheless, there is still a large amount of effort within Telcordia to address such immediate issues as the provisioning of Ethernet services across Sonet/SDH VCAT and LCAS capabilities; and some of this is complicated by the regulatory restrictions on RBOC activities in Layer 3 of the network.

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like