Optical Transport Network Switching

Optical Transport Network Switching: Creating efficient and cost-effective optical transport networks

White Paper

1.0 Executive summary: Building for the future with optical transport networks

Almost every day, headlines in the communications media highlight the phenomenal growth in data traffic that networks across the world are experiencing. The vast majority of this growth is being driven by bandwidthhungry IP/Ethernet applications in both private and corporate users. These applications include VPN services, SAN networks, Internet browsing, peer-to-peer video distribution, IPTV and video on demand. As opposed to the Internet and data applications of the past which required only best effort traffic, these new real-time services require not only high bandwidth, but also high availability, low latency, no jitter and high Quality of Service (QoS). While fixed lines will continue to carry the majority of this traffic, mobile traffic is set for a period of massive growth, far exceeding anything that has gone before. This represents a major opportunity for communications service providers (CSP). The convergence of fixed and mobile networks means that this growth will drive traffic in every network, placing new demands on the

transport core networks serving both fixed and mobile customers. These requirements will dramatically change the structure of tomorrow’s networks as their architecture shifts from being PDH/SDH-based, originally designed to support only voice, to being packet based. Core networks will have to cope with added traffic demand, while metro and access networks will need enhanced capacity and changed interfaces to cope with the new mix of data and legacy traffic. The most effective solution for meeting these challenges is implementing a Packet Optical Transport Network (POTN) with DWDM technology for transport and cross connects for traffic switching at the level of ODUs (Optical Data Units). This OTN switching concept forms a vital part of converged optical networks of the future. OTN switches provide efficient grooming of the optical signal on a sub-wavelength level. This increases the network efficiency by enabling

more effective use of bandwidth. In a mixed network that carries TDM and Ethernet traffic. OTN switches can switch data of any format. OTN switches also provide the most cost-effective way to offload traffic from the IP network layer, thus minimizing the amount of traffic handled by routers and enabling smaller and less costly routers to be used. There are many other benefits available from deploying OTN technology. These include support of Operations, Administration and Maintenance (OAM) features such as fast end-to-end service provisioning and rapid restoration. Furthermore, a converged optical network is open to being operated under a single management system to achieve the lowest overall costs. Together with the adoption of mesh network topology for the highest network availability, extreme scalability and easy implementation of new traffic types, it effectively makes network investments future proof.

2.0 Trends in optical networking 2.1 Network evolution Forecasters predict that consumer Internet traffic will grow from over 10 Petabytes per month in 2010 to more than 40 PB/month in 2014. What’s more, this estimate could be blown out

2

of the water by disruptive changes in services or customer behavior, such as the rapid rise of 3DTV. Meanwhile, enterprise customers are pushing up traffic volumes by relying increasingly on cloud services to deliver their business support systems.

OTN switching: Creating efficient and cost-effective optical transport networks

The nature of traffic is changing too, from merely browsing web pages to real-time, high-bandwidth and interactive applications that require minimum loading and response times. IP services are increasingly dominant. Consumers and business users

demand fast access to services, with high QoS. Yet intense competitive market pressures mean that they need pay little extra, or even nothing extra, for ever-greater performance and enhanced service capabilities. It all adds up to a challenge for CSPs, who face growing demand for network capacity and quality at the same time that the price they can typically charge for each bit is flat or dropping. Driving down the cost per bit is, in the majority of cases, the most critical issue, encouraging many CSPs to seek out new ways to maximize the flexibility and efficiency of their networks. Another key priority is to deliver a great customer experience, and that demands the network functionality to provide end-to-end QoS and high network availability.

Over recent years, many CSPs have addressed the need for greater network efficiency, for example, with the introduction of multi-service provisioning platform (MSPP) technology around a decade ago and the convergence of IP and DWDM roughly five years later. More recent network evolution has removed complexity and has broadly taken place in two distinct steps. First, the reduction and removal of ATM and SDH traffic was achieved by transporting traffic directly over the DWDM layer. Then, these DWDM networks were upgraded with the introduction of multi-degree ROADM nodes for switching optical traffic, together with more transmission line capacity and colored OTN interfaces for traffic from the IP layer.

However, this process creates some issues. For instance, there’s the question of multi-vendor compatibility during interworking, as well as the complexity and time required to provide network resilience in the optical layer and the increased complexity and cost when scaling the IP layer. The focus for the most recent step in this evolution is therefore a shift to promote greater efficiency, scalability and functionality. This can be achieved by introducing the OTN concept as an intermediate layer between the IP and the DWDM layers. The key benefits arise from implementing OTN switching at the cross-connects. This adds new value to the network and promotes significant CAPEX and OPEX savings. OTN switching enables traffic in transit to bypass many of the network’s IP layer

main volume expected in 2012 and further Addressing IP over DWDM approach challenges

• Multi-vendor compatibility at DWDM layer (transponder interworking, optical performance) • High optical restoration switching time • IP layer scalability (complexity and cost grows exponentially)

IP

DWDM (PXC)

10G / 40G / 100G

OTN

Addressing IP over DWDM approach challenges

• Sub-lambda grooming for efficient pipe filling • IP traffic offload / optical bypassing • End-to-end networking (high-capacity & multi-service switch, with simple OAM) • Intelligent management and carrier class network protection • Reduced power consumption and space requirement

Figure 1: Enhancing the network by introducing the OTN layer


3

routers and thus reduces the amount of expensive router capacity required. It also provides efficient grooming of the optical signal on a sub-wavelength level, which enables traffic to fill the available DWDM bandwidth more efficiently. This effectively minimizes the capacity needed to carry a given volume of traffic.

2.2 What is OTN switching? OTN provides service-agnostic switching that maps different client services into ODU frames and then switches them at that level. The actual process of OTN switching handles one or more ODU frames by bundling them together in a new ODU packet.

This “digital wrapper” approach encapsulates diverse data frames from different sources together in a single entity, regardless of their native protocol, so they can be managed more easily. The concept was first described in ITU-T G.709. The ODU concept has frames with fixed size and uses bit rates ranging from 1G to 100G to match interfaces for a range of standards, including Ethernet, SDH/ SONET and others. The OTN switches are also simple to manage, with SDHlike operations and maintenance. OTN delivers the key network functions needed to support high-quality end-

user services as flexibly and efficiently as possible throughout the different network domains. In the core network it offers high-capacity networking and rapid restoration, while the metro network benefits from service-agnostic aggregation and grooming. Users can enjoy cost-efficient delivery of multiple services in the access network, as well as benefiting from the service quality that can only be provided with end-toend networking and provisioning. The following chapters discuss the technology and architecture that underpins OTN switching. We’ll also consider what potential benefits OTN switching promises to deliver to CSPs.

3.0 OTN technology overview

3.1 The principles of OTN switching OTN was developed by the ITU’s Telecommunication Standardization Sector (ITU-T) as a way of optimizing traffic efficiently while simultaneously coping with a new traffic mix. The ITU-T defines OTN as follows.

4

“An Optical Transport Network (OTN) is composed of a set of Optical Network Elements connected by optical fiber links, able to provide functionality of transport, multiplexing, routing, management, supervision and survivability of optical channels carrying client signals.” (http://www.networkworld.com/details/4521.html?def)


OTN provides digital wrappers that contain multiple data frames from different client services together in a common ODU. This enables the network to switch high volumes of any type of traffic efficiently, including Ethernet and legacy SDH traffic. OTN ensures that all 40 Gbps and 100 Gbps digital wrappers are fully packed to make maximum use of the network’s available bandwidth. A simple analogy of the principle behind OTN is filling buses with passengers. A bus may leave the bus station only partly full. As more passengers arrive at the bus station, more half-empty buses leave on their journeys in different directions, which is clearly inefficient. OTN switching is akin to ensuring that every bus is filled to capacity before it leaves and before passengers are allowed to start boarding the next bus, even if some will later change or get off at intermediate stops. This increased fill rate also applies to optical channels, which are now using their large available transport capacity better. This limits the need to deploy new and costly DWDM channels. According to a study in February 2009, using this kind of intermediate traffic grooming with an ODU switch can reduce wavelength usage by 40% (Thomas Engel, AchimAutenrieth, Jean-Claude Bishoff, “Packet Layer Topologies of Cost Optimized Transport Networks”, ONDM, Braunschweig, Germany). The OTN concept is simple yet powerful, because a diverse range of client signals can be managed

together. It’s also more flexible than the previously dominant architecture (SONET/SDH).

ODU and OTU signals is the forward error correction (FEC) header contained in the OTU format.

OTN is currently offered in the following line rates:

ODUk data generally follows the same approach as OTUk, but some new service formats can’t fit into any existing ODUk without wasting bandwidth. Also, defining a new ODU container each time a new client is added would require upgrading ODUk switch fabrics. ODUflex was therefore introduced as a flexible lower order container that can be “right sized” to fit any client rate and overcome the problem.

• OTU1 has a line rate of approximately 2.66 Gbit/s and was designed to transport a SONET OC48 or synchronous digital hierarchy (SDH) STM-16 signal. • OTU2 has a line rate of approximately 10.70 Gbit/s and was designed to transport an OC-192, STM-64 or WAN PHY (10GBASE-W). • OTU2e has a line rate of approximately 11.09 Gbit/s and was designed to transport a 10 Gbit Ethernet LAN PHY coming from IP/Ethernet switches and routers at full line rate (10.3 Gbit/s). This is specified in G.Sup43. • OTU3 has a line rate of approximately 43.01 Gbit/s and was designed to transport an OC-768 or STM-256 signal or a 40 Gbit Ethernet signal. • OTU3e2 has a line rate of approximately 44.58 Gbit/s and was designed to transport up to four OTU2e signals. • OTU4 has a line rate of approximately 112 Gbit/s and was designed to transport a 100 Gbit Ethernet signal. The OTUk (k=1/2/2e/3/3e2/4) is a signal format into which another information structure called ODUk (k=1/2/2e/3/3e2/4) is mapped. ODUk is the server layer signal for client data signals. The main difference between

ODU data packets form the basis for flexible mapping and multiplexing within the OTN switching process. Different services are converted into ODU packets and these are directed to their destination ports and converted into OTU for optical transmission. It’s useful to distinguish between single-stage and double-stage ODU multiplexing, which is typically used for mapping several small ODUs into one large ODU. The flexible multiplexing scheme also allows mapping, in one or more steps, of small ODU data packets into a larger ODU container, maintaining a small data granularity while using a large transport capacity. Note that lots of services and service types can be processed by a single OTN switch simultaneously. All that’s required is the right mix of interface cards to support the appropriate client formats, for example GbE, and sufficient port capacity. Figure 2 illustrates which service types can be mapped into an ODU packet and finally into the required transport (OTU) format.


5

GbE/DVB/FE FC 1G STM-1/OC-3 STM 4/OC-12

ODU0 (L)

STM-16/OC-48 FC 2G

ODU1 (L)

STM-64/OC-192 10 GBASE-W

ODU1 (H)

OTU 1 ODU2 (H)

ODU2 (L)

OTU 2

10 GBASE-R FC 10G

ODU2e (L)

OTU 2e ODU3 (H)

STM-256/OC-768 40 GBASE_R FC 4G FC 8G

100 GBASE_R

ODU3 (L)

OTU 3

ODU flex (L)

ODU4 (H)

ODU4 (L)

OTU 4

Not specified in G.709, but in G.sup43

Figure 2: Flexible mapping and multiplexing scheme highlighting the ODU concept

Figure 3 highlights a scenario with vast mapping options in an ODU4, combining single ODUk and dual-stage multiplexed ODUk, as well as different ODUflex pipes.

3.2 The architecture of OTN switches The substantial cost saving benefit of OTN switching is complemented by

increased network resilience and greater flexibility. The way in which OTN is implemented by the design of the hardware has a significant impact on the scale of these benefits. The architecture of an OTN switch should maximise flexibility and availability. OTN switches are protocol-independent and operate transparently both on TDM and packet-based traffic. They break traffic down into ODU packets in

n x MPLS-TP tunnel n x ODU flex (various size)

ODU4 (L) (100G)

dual stage multiplexing (ODUflex via ODU2 / ODU3) dual stage multiplexing (all ODUk) ODU0 / ODU1 / ODU2(e) / ODU3

Efficient bandwidth utilization

Built on flexible client service mapping and ODU frame multiplexing acc. to ITU-T G.709

Figure 3: Network capacity optimization – example of ODU mapping and multiplexing for large capacity 100G channels

6


a suitable interface card, then perform the switching and grooming function in a centralized switch fabric and forward the ODU packets to the switch’s relevant output port via the respective interface card. This switch fabric is the heart of the OTN switch, connecting the interface cards via an electrical backplane. The ODU packets are processed in the electrical switch matrix with agnostic cell switching. The total traffic load from all the interfaces in the switch can be shared among several switching cards that operate simultaneously. This approach should provide CSPs with excellent scalability in their switching capacity, thanks to multiple switch fabric modules (SFMs) in a single chassis. It opens up an opportunity to benefit from a pay-as-you-grow approach to investment. To achieve maximum redundancy and protection in the system, all the SFMs in the OTN switch should be able to share the traffic load. As well as achieving effective load balancing, this architecture ensures that, should one card fail, the remaining cards can redistribute the extra traffic between them to avoid service interruptions. This structure also enables CSPs to carry out upgrades while the OTN switch remains in service. Similarly, redundancy should be provided in the power supply cooling fans and system controllers. Combined with service protection, this results in the highest, carrier-grade, availability. Figure 5 summarizes the switch architecture from the perspective of the signal format. It illustrates ODU switching functionality, which is used for all different formats. Alternative approaches might include packet switching for MPLS/Ethernet signals using an MPLS interface card, or SONET interfaces for ODU or VC4-based switching, all handled by the same switch fabric. Therefore, the OTN switch becomes the only switching element for all the formats present in the network.

Switch Fabric Module

#1

OTN interface card

multi-vendor networks where the OTN standard ensures interoperability between many different platforms.

System controller card (w) System controller card (p)

MPLS-TP interface card

#2

Ethernet L2 interface card

#3

Switch Fabric Module

If the network is newly deployed and the DWDM transport system and the switch are both from a single vendor, a single NE can be defined at the network nodes of the mesh network, incorporating the OTN switch and the optical transport equipment logically into one NE. The advantage of this approach is that the composite element is managed by the network management system (NMS) as one NE featuring OTN switching, MPLS-TP switching, SONET switching, WDM switching and WDM transmission.

Peripheral controller card (w) Peripheral controller card (p)

Flow Sensor Card (CFSU) SDH/SONET interface card

#4

SDH/SONET Bridge card

#5

2x / 3x (1+1) Power Supply Unit

8 / 12 Fan Trays

#6 OTN interface card

#15 / #30 #1

Figure 4: Example for the architecture of an OTN switch showing interface cards, the switch fabric and additional controller cards

switching at the wavelength level has existed previously, for example, using multi-degree ROADMs. The wavelengths of the DWDM system are optically de-multiplexed and individually connected to the OTN switch interface cards. The standalone NE could also be connected to other equipment such as IP routers. This kind of deployment is typical in

OTN switches can be deployed in a network in several ways. In an existing DWDM network with mesh topology, OTN switches can be deployed as a standalone network element (NE) operating as a crossconnect at a network node. This is the typical way in which OTN functionality is introduced where only optical

OTN interface card Client signal

Mapping

In yet another scenario, the OTN switch can be deployed as a standalone NE together with extension shelves. This enhances the capabilities of MSPP platforms with additional OTN switching functionality. The next issue is where in the network to deploy OTN switches. Typically, most switching and aggregation takes place in the metro and core parts of the network, which define the capacity and configuration requirements for OTN switches.

Switching Fabric Module

OTUi Ethernet SDH/SONET

-> ODUj

Other (FC, etc.)

local node traffic OTU signal Ethernet signal SDH / SONET signal other signals ODU frames Ethernet / MPLS signal virtual switching domain

ODU switching

OTN interface card packet switching

Multiplexing

Line signal

…-> ODUk

…-> ODUk

VC-4/STS-1 switching

Client signal

Line signal

Figure 5: OTN traffic model based on ODU switched client formats


7

Services

Access

Metro

Core

NMS

Figure 6: OTN switches are typically placed in the metro and core networks

Generally, all OTN switches should support the same overall functionality. However, in metro networks the focus is on the aggregation of different services to fill the wavelengths efficiently with various traffic formats while ensuring service transparency across the network.

For the core network, the focus is on achieving switching capacity in the Terabit range, practically distributed over several chassis by capacity expansion. Core-based switches handle enormous traffic loads and can switch any service from one wavelength to another. They must also

support networking functions such as protection and restoration. Across the network domains, OTN switches perform other advanced networking functions such as end-to end provisioning and different resilience schemes.

4.0 The major benefits of OTN switching

The introduction of OTN switching will bring major benefits for CSPs in terms of lower costs, higher efficiency and greater functionality in their transport networks. As networks grow and need to accommodate new types of traffic alongside existing traffic, it becomes vitally important to fill the available transmission bandwidth efficiently. This calls for techniques such as grooming and aggregation. Beyond the task of transporting and switching traffic, a unified management system can add value to the network by providing simplified operations and functions such as the end-to-end provisioning of services and resilience schemes

8

throughout all the network layers and segments. It’s ultimately about aiming for the most efficient and cost-effective end-to-end networking. Such a converged optical network shall fulfil the following requirements. • Optimize and simplify the network structure with new OTN switches. • Simplify network operations. • Migrate any installed base. OTN switches are embedded in the converged packet optical transport system (P-OTS) and hence play a vital role in the metro and core of such a network. The choice of interface cards


enables these switches to handle the dominant packet traffic as well as any existing legacy formats, from MSPP or Carrier Ethernet Transport (CET), for example. Typical packet based services are FE/GE/10GbE/40GbE and 100GbE, which are mapped into ODU0,2,2e,3,4,flex as already shown. Moreover, advanced OAM capabilities provide CSPs with a unified platform to manage the network and services effectively end-to-end, with common service, fault and performance management. This helps deliver a high-quality end-user experience.

OTN

Figure 7: The functional view of OTN switching in the P-OTS architecture

Figure 7 illustrates how OTN switches fit into this overall network architecture. In the core network they connect the IP layer and the DWDM-based optical layer. They also provide traffic grooming and aggregation throughout the rings that make up the metro networks. From there they direct the traffic to the correct carrier in the access network, whether that’s Ethernet, DSL or a radio cell, for example. The following sections highlight what OTN switching can deliver in terms of capacity improvements, reduced investment, increased availability and reliability and improved support for service management. This white paper presents a broad network view, while

Sub-lambda grooming Channel 1 Channel 2

specific OTN topics are discussed in more detail in complementary white papers.

4.1 Filling the pipe with sub-lambda grooming In DWDM transport systems, line rates are approaching 100 Gb/s with 80 or more optical channels transmitted simultaneously. This enormous capacity can be handled purely on the optical layer when using multi-degree ROADMs for switching at network nodes. However, with this technology the smallest unit for switching is a single optical wavelength. Given the trend to high line rates, this translates

Up to 40% less wavelengths in real life networks reduces number of deployed transponders

Channel 3

Figure 8: Sub-lambda grooming packs more streams into each wavelength and therefore effectively reduces the number of required distinct wavelengths

into an equally high granularity. On the other hand, services demand comparatively low data rates in the Mb/s or low Gb/s range, which results in inefficient fill rates in the optical channels. The ideal solution for CSPs would therefore be to create a high-capacity network over DWDM combined with low switching granularity for efficiency. Sub-lambda switching and grooming is the answer. In this process, multiple data streams carrying different services are electrically multiplexed into larger units that can be processed and transmitted as single entities. In the same way, such streams can also be de-multiplexed at a switching node to access and extract specific data. The primary aim is to lower the cost of handling traffic in the network by making better use of the available capacity. In this set-up, a DWDM wavelength can be used as a pipe for lots of different traffic. In particular, adopting the ODU-based approach in OTN switching can reduce the wavelength used to carry a given volume of traffic by 40% by enabling CSPs to pack more streams into each wavelength.


9

combination allows CSPs to optimize the transport layer to handle the dynamic behavior of IP/MPLS.

Core router capacity [Tbps ] 10 Tbps

Required core router capacity

5 Tbps

Max single shelf router capacity

2005

2010

2015

2020

Figure 9: The evolution of core router capacity over time due to increasing IP traffic

MSPP platforms carrying traditional SDH/PDH services also support this concept of switching. However, with the ODU layer, the switching concept is expanded into a fully service-agnostic platform, incorporating new, IP/ Ethernet services in addition to legacy services and combining impressive scalability with a unified network management system. This provides a future-proof networking architecture for CSPs to manage and evolve their service portfolio easily.

4.2 Traffic offloading reduces CAPEX Routers are closely integrated into the overall network where they provide switching and routing capability. But, in general, router capacity remains expensive, with costs increasing as more and more IP traffic is generated. This increasing IP traffic can be divided into multipoint-to-multipoint traffic, like classical L3 VPN and VPLS and pointto-point traffic like Internet, which is directed from a PE router like a BNG or

10

GGSN/SAE-GW to the Internet peering points. This point-to point traffic does not necessarily need to be switched by the large core routers, so the IP core router can be saved for VPN traffic. Alternatively, upgrades from single-chassis routers to more costly multichassis routing systems can be avoided. Put simply, routers alone will not be able to keep up with the increasing bandwidth demand in a technoeconomical environment and only a combination of routers and optical transport will be economically viable for a traffic mix dominated by IP/MPLS. MPLS is a technology for labeling IP packets so they can be directed around the network. The particular MPLS Transport Profile (MPLS-TP) supports carrier-grade OAM, as well as performance and fault management. There are some additional benefits that the combination of MPLS-TP and ODU brings over and above those offered by the ODU concept alone. Generally, this


All IP/MPLS traffic and label-switched paths (LSPs) can share the full 10G/40G/100G line interface capacity between nodes that are acting as MPLS-TP switches. This delivers a statistical multiplexing gain and increases efficiency in the network even further. Since no fixed bandwidth is assigned, this is a close approximation of the dynamic nature of the IP/MPLS traffic. In the metro network, MPLS-TP helps provide more efficient packet aggregation from MSPP/CET and DWDM sources. MPLS-TP supports OAM functions so the network operator can configure traffic profiles, QoS, protection and restoration to optimize the transport layer and meet the quality requirements of IP/MPLS. A key advantage of combining MPLSTP functionality and OTN switching is the ability to offload traffic from the IP layer. This enables transit traffic to bypass intermediate routers entirely, significantly reducing the required router capacity and saving capital expenditure (CAPEX) and operational expenditure (OPEX). In the traditional set-up, the IP and transport networks are separated entities. DWDM transport networks offer plain connectivity to the IP core network. Features such as resilience are based in the IP layer. Such networks are relatively slow scaling, with increased capacity creating extra cost, space and power challenges, even though the IP routers only perform simple tasks such as label switching for the majority of traffic.

Figure 10: Combining MPLS-TP functionality and OTN switching gives IP traffic the ability to remain in the transport layer, thus effectively offloading it from the IP layer

Incorporating MPLS-TP in the OTN switches and transport layer - with the DWDM transport and OTN switching provides the following enhancements. • IP offload: Shifting traffic to the ODU layer frees up core router interfaces and saves CAPEX by requiring less router hardware and lowers OPEX by reducing power, footprint and cooling • Transit traffic: Label-switched routing can be carried out in the transport domain, so transit traffic can bypass the IP layer as it traverses the OTN platform.

• Resilience: MPLS-TP and ODUlevel mechanisms IP service layer also offer fast service recovery • DWDM-layer optimization: Combined with the benefit of subElectrical switching layerit yields a lambda traffic grooming, dramatic improvement in wavelength utilization and network efficiency. DWDM layer

For example, to assess the scale of the potential savings, consider the case of a European greenfield CSP experiencing a 50% annual traffic growth rate and using MPLS-TP to offload IP traffic in order to limit the need for large capital investment in

IP service layer

Electrical switching layer

DWDM layer

Figure 11: Incorporating MPLS-TP in the transport layer means that intermediate sites have reduced or no router traffic

router capacity. Using OTN switches to carry all peering traffic results in offloading 70% of all traffic from routers, leaving the remaining 30% of traffic within the router layer. Comparing the cost of investing in router capacity only with investing in a combination of OTN switches and routers reveals that the cost of introducing the MPLS-TP layer would be recouped within one year. The cumulative CAPEX saving over a five-year period amounts to about 60% In addition, significant OPEX savings would amount to a 60% saving, while reduced power consumption would equal CO2 savings of up to 590 tonnes over five years. OTN switches that support MPLS-TP also help to future-proof the architecture by enabling the migration of legacy circuit-switched transport networks to next-generation packet-optimized optical transport networks. Last but not least, little investment is required to implement MPLS-TP capabilities, since it is merely a software feature running on the OTN switch.


11

Capex

Opex -80*%

IP CORE ROUTER

Weight -60%

IP CORE ROUTER

OTN switch + MPLS-TP


IP CORE ROUTER

-60%


* Depending on final configuration. Example 50% OTU-2 and 50% Client 10G MPLS interfaces

Figure 12: Benefits deriving from OTN switching for the discussed network scenario

OTN switches equipped with MPLS-TP can also carry out many of the traffic management functions that would otherwise be handled exclusively by IP-layer routers. Furthermore, by incorporating them directly onto the transport layer, these advanced features become part of the transport network, regardless of the IP router actually used. This approach also helps CSPs to generate significant savings by reducing the router capacity needed in the network.

4.3 Robust protection promotes carrier-grade availability As we have seen, combining OTN switches in a mesh topology provides an extremely robust network. The OTN switches typically have redundant hardware installed for maximum reliability, but in the event of a major failure or an incident on the DWDM transport layer, such as a severed fiber,

C

traffic may need to be rerouted. The right topology can even protect against multiple failures.

the functionality of the nodes where the switching occurs in the event of a failure.

It’s possible to provide protection in the IP layer, although it can be relatively slow to recover. There are also different techniques for protection that can be applied directly on the DWDM layer, based on GMPLS or simple 1+1 optical switching. The best solution for a particular network depends to its specific topology, but OTN switches offer a new degree of sophistication for traffic protection. Protection occurs on the service level, making it possible to apply different measures to traffic with different QoS.

OTN switches enable protection to be configured at the level of the individual ODUs, which provides much finer granularity than protection provided by the line interface alone. Features include sub-network connection protection (SNCP), which works on different ports of the interface cards and can protect particular ODUs. There are also the same protection architectures familiar in other layers of the network, such as multilayer GMPLS (Och, ODUk and VC4). OTN switching also supports techniques such as hold-off timers to prevent the triggering of multiple protection measures and wait-torestore timers combined with revertive protection, which automatically restores the original traffic configuration when possible.

OTN technology supports the full range of protection techniques to prevent failures and speed up recovery times. The first step to increase resilience is to build alternative paths into the network using ring topology, dual nodes and high connectivity, for example. The next consideration is

F 2

Initial route 1st alternative route

1

2nd alternative route

A

D

3rd alternative route

G N 3

B

Failure N

E

Figure 13: OTN switches within a mesh network topology supports protection switching and creates a robust network

12

The deployment of MPLS-TP with OTN provides additional protection at all levels from the end-to-end transport path to individual links. The above options are complemented by packet-oriented protection schemes specifically for LSP and pseudo wire (PW) connections.


4.4 OAM capabilities support a great customer experience Effective OAM comprises a set of network-oriented mechanisms for monitoring and managing the network

MPLS-TP PE

LH-OTP

MPLS-TP NNI

P

P

LH-OTP

LH-OTP

MEP

PE

MPLS-TP NNI

LH-OTP

MEP

Throughput test

Figure 14: OAM capabilities shown for a performance management throughput test

infrastructure. Since the OTN switch offers a lot of functionality within the network environment, OAM support optimizes the integration of OTN switches into the overall network management. In particular, the deployment of MPLS-TP within the OTN switch provides carrier-grade OAM capabilities.

Common functions are performance management (PM) and fault management (FM). Together with MPLS-TP, PM comprises packet loss, packet delay and throughput, end-toend across the entire network. For FM, features such as trace routing and LSP pings are available, as well as a set of proactive functions.

Figure 14 shows a throughput test for PM. This is the typical approach for verifying the bandwidth of an MPLS-TP transport path before it is brought into service. This is the configuration typically used in the traffic offloading scenario. In the architecture shown here, we consider a transport network domain with OTN switches, comprising PE switches at the edge and P switches in between. The throughput tests are performed on the transport layer between the two end points. The OAM systems control the end points and a test-traffic sequence is sent between them, complete with a byte count. The benefits of effective OAM include fast end-to-end service provisioning, maintenance and the fast restoration of services. Improved performance and the ability to trace network faults in case of failures translate into less downtime, which in turn helps to drive down OPEX for CSPs.

5.0 Packet Optical Transport from Nokia Siemens Networks

Nokia Siemens Networks offers a P-OTS solution with advanced transport and optical and electrical switching capabilities. The P-OTS solution also supports the end-to-end provisioning of services across a network and provides restoration functions for greater network resilience. Nokia Siemens Networks P-OTS can be tightly integrated with existing

DWDM platforms and features advanced network architecture that places no restrictions on services or applications. DWDM transport capacity is upgradable up to 96 channels with 40 Gbit/s or 100 Gbit/s, while switching is scalable from a capacity of 0.2 Terabytes per second (Tbps) up to more than 24 Tbps, complete with 40 Gbit/s and 100 Gbit/s optimized line interfaces from a single platform.

Seamlessly integrating with the P-OTS solution is the hiT 7100 OTN switch that offers multilayer switching including ODU -0/1/2/3/4/flex, MPLS-TP, L2 Ethernet and VC-4/STS-1. Using intelligent multi-service interfaces, the switch is future-proofed, because any service can be processed transparently by adding the relevant interface card. This protocol-independent switch features 3 Tbps switching capacity per


13

chassis that can be flexibly scaled up to 48 Tbps in a multi-chassis configuration. Carrier-class network resiliency is provided by the hiT 7100’s built-in switch fabric module and controller card redundancy offering N+1 protection, combined with several protection options offered on ODU layer. The hiT 7100 can be provided in different configurations to suit any network need, whether for a standalone network element, or on top of existing DWDM infrastructure. OTN capacity planning is supported by the Nokia Siemens Networks NPS10 planning tool and optical network planning is supported by TransNet. Combined, this offers a complete

process of planning, configuration and installation of packet optical networks including configuration, upgrade, optimization, ordering, commissioning, and network maintenance. P-OTS network management is performed via the carrier grade Nokia Siemens Networks Transport Network Management System (TNMS) solution that provides automated or manual provisioning of ODU connections and interface connections. The system also provides monitoring and performance management at the element and transport service levels.

database creation and adaptation, and replacement preparations for upgrading networks. Highly trained experts throughout the world can save CSP time and effort by offloading lengthy design and database management activities, helping to lower the total cost of ownership by allowing CSPs to focus on their primary business, providing customers the new, high quality services they demand. The use of Nokia Siemens Networks services not only ensures the highest network efficiency and integration of upgrades, but also paves the way for future upgrades.

Nokia Siemens Networks also provides a full set of professional services, including assessment, design,

6.0 C onclusion: OTN switches at the heart of efficient and flexible packet transport

Deploying OTN switching technology in metro and core networks is one of the most cost effective opportunities for CSPs to meet the booming capacity demand and manage the increasingly complex and dynamic mix of traffic generated by advanced IP-based services. While the concept of introducing an optical switching network layer may seem contrary to the trend over the last decade of simplifying networks by removing layers, the technology does promise substantial cost savings and performance gains. OTN switching can be combined more efficiently with an IP core network, saving significant CAPEX at the network level. OTN switching enables

14

traffic in transit to bypass many of the network’s IP routers and thus reduces the amount of expensive router capacity required. It also provides efficient grooming of the optical signal on a sub-wavelength level. This uses existing bandwidth more efficiently, thereby minimizing the capacity needed to carry a given volume of traffic. These benefits are compelling reasons for CSPs to adopt OTN switching. Using the Nokia Siemens Networks converged optical packet portfolio, CSPs can support a multitude of applications in metro and backbone optical transport networks. As these networks are typically multiservice in nature and contain complex mesh


topologies carrying many different bit rates, an electrical grooming layer is required, achieved when the P-OTS is combined with the Nokia Siemens Networks hiT 7300 DWDM platform. These factors shape the Nokia Siemens Networks ‘converged’ transport network vision, with IP optimized optics enabling the exponential growth of the Internet at the lowest possible cost per bit. With the P-OTS combination of the hiT 7100 OTN switch and hiT 7300 DWDM platform, CSPs can plan and deploy their entire optical transport network in the most cost effective way, seamlessly migrating towards nextgeneration packet optical networks.

Summary of benefits of OTN switching Capital and operational cost reduction

Maximum utilization of wavelengths by sub-lambda grooming IP traffic offloading reduces investment in IP network layer routers and reduces running costs Multi-vendor interoperability simplifies procurement and operation Common management system for OTN and DWDM layers reduces system complexity and simplifies network operations Carrier-class protection and migration to mesh network topology improves network availability and reduces maintenance

Rapid return on investment

Rapid service provisioning enables fast time to market and early revenue

Future-proofed network

Extreme scalability and flexibility to adopt new service types protects investments Migrate to IP/Ethernet services while supporting traditional SDH/PDH services

Abbreviations

ATM Asynchronous Transfer Mode BNG Border Network Gateway CAPEX Capital Expenditure CET Carrier Ethernet Transport CSP Communications service provider DSL Digital Subscriber Line DWDM Dense Wavelength Division Multiplexing FEC Forward Error Correction FM Fault Management GMPLS Generalized Multi-Protocol Label Switching GW Gateway IP Internet Protocol ITU-T International Telecommunication Union’s Telecommunication Standardization Sector LAN Local Area Network LSP Label-Switched Path MPLS MultiProtocol Label Switching MPLS-TP MPLS Transport Profile MSPP Multi-Service Provisioning Platform NMS Network Management System OAM Operations, Administration and Maintenance ODU Optical Data Units

OPEX OTN POTN PDH PM P-OTS PW QoS ROADM SAN SDH SFM SNCP SONET TDM TNMS VPN WAN WDM

Operational Expenditure Optical Transport Network Packet Optical Transport Network Plesiochronous Digital Hierarchy Performance Management Packet Optical Transport System Pseudo Wire Quality of Service Reconfigurable Optical Add-Drop Multiplexer Storage Area Network Synchronous Digital Hierarchy Switch Fabric Module Sub-Network Connection Protection Synchronous Optical Network Time Division Multiplexing Transport Network Management System Virtual Private Network Wide Area Network Wavelength Division Multiplexing


15

Nokia Siemens Networks P.O. Box 1 FI-02022 NOKIA SIEMENS NETWORKS Finland Visiting address: Karaportti 3, ESPOO, Finland Switchboard +358 71 400 4000 (Finland) Switchboard +49 89 5159 01 (Germany) Order-No. C401-00725-WP-201107-1-EN Copyright © 2011 Nokia Siemens Networks. All rights reserved. Nokia is a registered trademark of Nokia Corporation, Siemens is a registered trademark of Siemens AG. The wave logo is a trademark of Nokia Siemens Networks Oy. Other company and product names mentioned in this document may be trademarks of their respective owners, and they are mentioned for identification purposes only. This publication is issued to provide information only and is not to form part of any order or contract. The products and services described herein are subject to availability and change without notice.

www.nokiasiemensnetworks.com

Optical Transport Network Switching

Recommend Documents