Thursday, December 6, 2012

Blueprint: The Transport Network Challenge

by Scott Wakelin, Product Line Manager in PMC-Sierra’s Communication Products Division

Optical network operators worldwide are faced with a tremendous challenge – expanding their networks to keep up with massive traffic growth and doing so profitably.

In 2012, Cisco’s Visual Networking Index (VNI) projected network traffic would quadruple between 2011 and 2016 to 1.3 zettabytes or 1.3 trillion Gigabytes annually. Video will continue to grow and eventually consume a 55% share of network traffic. Likewise, mobile traffic will grow 18x, driven by the transition to HSPA+, LTE, and LTE-Advanced.

Market research firms project that by 2015, optical spending will increase 25% over the $12B spent in 2010 as carriers prepare to build out their metro and access networks to deal with the massive increase in Ethernet and packet traffic.  

What will the new metro network look like and what capabilities will be required?

Before exploring these questions, let’s review the architecture of today’s typical carrier network.

Today’s Carrier Network

In the access network, TDM services (T1/E1 private line, ISDN, voice, 2G wireless) dominated until only recently. The last few years have seen dramatic changes in the access service landscape with Ethernet replacing T1/E1 for both enterprise and mobile access. At the same time, demand for native Video and Storage Area Network (SAN) transport has accelerated, adding to the service mix that carriers must support.

Meanwhile, outside of China, Layer 1 transport in the metro continues to be largely SONET/SDH based. Today, carriers aggregate client traffic into SONET/SDH (generally at 10G). The resulting OC-192/STM-64 signal is then fed into a transponder which converts the 10G client signal into a 10G wavelength using first generation OTN (ITU-T G.709 Optical Transport Network) equipment. At this point, the signal is ready for transport over the ROADM based DWDM infrastructure.

The access transition to Ethernet coupled with exploding bandwidth demands has exposed three fundamental weaknesses of SONET/SDH based Layer 1 aggregation, which fundamentally limits the ability of carriers to scale their metro networks:

  1. Fixed switching granularities which are only a fraction of the 10G line rate
  2. Inefficient support for Ethernet without the use of VCAT
  3. Little deployment beyond 10G and no roadmap beyond 40G
As a result of these challenges, carriers are preparing to deploy a new metro network. The next section explores the coming Metro Transport Network evolution.

 The New Metro Network

In order to scale their metro networks to handle the growth in access traffic, carriers seek a network technology that:

  • Supports the full range of protocols that exist in the metro, including Ethernet, SONET/SDH, SAN, and Video, without the use of Circuit Emulation or Pseudo-wire emulation techniques,
  • Supports efficient transport of packet services such as Ethernet
  • Is able to scale to 100G and beyond,
  • Offers a simple to manage Layer 1 network that extends end-to-end.
Today, carriers have broadly deployed OTN as the basis for their DWDM core networks and it has proven an effective technology in providing both the management, protection, and reach extension required in the core network.  The desire for continuity at layer 1 between the core and metro networks made OTN a primary candidate for the Metro transport network as well.  However, OTN technology, as originally deployed in the core, fell short in terms of efficiency of Ethernet transport, and switchability.  Nevertheless, the G.709 standard has evolved to become a highly efficient transport technology for Metro applications, with the result that OTN is the nearly unanimous choice of carriers globally to base their Metro networks.

PMC refers to this evolved OTN technology as Metro OTN.

Metro OTN

Let’s look more closely at how well OTN meets the needs of the new Metro network. 

Multi-Service Transport

Metro OTN provides standards-based methods to enable full bit and timing transparent transport of Ethernet (1GE, 10GE, 40GE, or 100GE) – which is critical for the growing Ethernet private line services market. In addition, OTN also supports GFP-F mapping of packet based services such as:
  • MAC terminated Ethernet
By virtue of this capability, and when coupled with Carrier Ethernet features such as IEEE 1588v2 (Precision Time Protocol) and Synchronous Ethernet, OTN is ideally suited for the quickly growing mobile backhaul market. 

Now, Ethernet is not the only client in the metro. SAN services such as Fiber channel and Infiniband are commonly used for datacenter to datacenter interconnect. Uncompressed HD and SD video streams are increasingly used in video contribution networks due to their superior quality and low latency. Prior to OTN, these bit and timing transparent services would generally be transported directly over DWDM but did so at the expense of reduced or no manageability. OTN provides the bit transparent transport these services require coupled with enhanced end-to-end OAM that includes 6 layers of Tandem Connection Monitoring (vs. the single layer offered by SONET/SDH).

Furthermore, there remains a tremendous installed base of SONET/SDH with new deployments still expected for at least the next 5 years. OTN was designed to accommodate both asynchronous and synchronous mapping of OC48/STM-16 and OC192/STM-64 clients. In this manner, OTN can provide the means for the bit and timing transparent transport of SONET/SDH, whether point to point or ring based – and importantly, without the need for PWE3 or CES.

 Efficient Resource Utilization

The efficiency issues associated with transporting Ethernet over SONET/SDH are well known. But even 1st generation OTN suffered from efficiency issues. Take for instance a GE to be transported over an OTU2 operating at 10 Gbps. First generation OTN equipment either:
  1. did not support this capability,
  2. did not support it efficiently, or
  3. did not support it in an interoperable manner
In contrast, Metro OTN naturally supports Ethernet, and unlike SONET/SDH does so with a single ODU container to provision, switch and manage. This greatly simplifies provisioning and management, ultimately leading to reduced OPEX. Furthermore, as Ethernet scales in the future, so will OTN.

With the development of Metro OTN, carriers can now efficiently map GE into the new ODU0 container operating at 1.25G – right sized for GE. The GE may be mapped in a bit and timing transparent manner for private line service, or may be MAC terminated for managed service delivery. Figure 6 illustrates that in comparison to 1st generation OTN, Metro OTN will double the efficiency of GE transport.

Figure 6 also illustrates how the new variable rate ODUflex container drives efficiency gains for other common metro access clients. Take for instance 3G-SDI. In 1st Generation OTN equipment, this video client was at best 30% efficient when transported using a 10G ODU2 signal. ODUflex enables a container to be assigned that closely matches the client rate. ODUflex can also be used to transport subrate 10GE signals, which has the power to open up new private line service options for enterprises and revenue streams for carriers, while at the same time allowing the carrier to efficiently use its fiber resources. Furthermore, each ODU container contains all of the OAM flexibility that OTN is known for.

The new ODU0 and ODUflex containers are also switchable. Let’s explore the final aspect of Metro OTN: the support for flexible, granular and distributed OTN switching.

Flexible, Granular and Distributed OTN Switching

The vast majority of access services are sub-10G, with GE the access currency of choice for broadband and enterprise access. At the same time, the metro network is generally built around 10G wavelengths, with carriers preparing for broad deployment of 40 and 100G wavelengths in the metro. As a result, the gap between client rate and wavelength bandwidth is increasing.

In recognition of this trend, early OTN deployments were based on muxponders which multiplex client signals into a single outgoing OTU2, OTU3, or OTU4 as shown in figure 5.

Muxponder based compact metro access solutions are ideal for aggregation of mobile, broadband, and enterprise services, and are a growing trend among equipment vendors and carriers alike. In a fiber-rich access network, muxponders can cost-effectively provide bit and timing transparent mapping of SONET/SDH, Ethernet, SAN, and Video into grey or colored OTN signals.

However, when used in multi-slot / multi-wavelength systems deeper in the metro and the core, muxponders and transponders can lead to inefficient wavelength utilization as a full wavelength must be assigned regardless of the total client bandwidth. Client Add / Drop and Continue is also hindered by the inflexible nature of Muxponder/Transponder architectures. Only clients that are physically connected to a particular board can be mapped into that boards specific outgoing wavelength. This leads to a more complicated service provisioning and management model. For example, if a client needs be moved from one muxponder to another (in order to be transmitted on a different wavelength), human intervention is required. This inflexibility leads to increased OPEX for the carrier.

Metro OTN addresses these challenges through the deployment of OTN switching systems.

In comparison to muxponders, the benefits of OTN switching include:
  • Efficient grooming of any sub-wavelength client onto any outgoing lambda,
  • Maximum wavelength utilization
  • The ability to switch an ODU from any outgoing line interface to any outgoing line interface
  • The ability deploy remote management, eliminating the need for manual patching,
  • Separation of client and line optic interfaces, which enables a carrier to deploy 100G wavelengths as traffic dictates

Unlike SONET/SDH, OTN imposes no limitations on switching granularity. All ODUs may be switched between any ingress and egress line card through a cell, TDM, or off-the-shelf packet fabric using the new OIF OTN over Packet Fabric format.

The deployment of an OTN switching system in the metro is a critical requirement if carriers are to achieve the most efficient use of their network resources at the lowest possible OPEX.  

Silicon Impact of Metro OTN

Just as the metro transport evolution is driving new requirements for OTN equipment vendors, Metro OTN also drives new requirements for silicon vendors. No longer is a simple implementation of G.709 sufficient. The following fundamental features are also required:

  • Any-Service, Any-Port, Any-Rate SERDES and mappers  in order to deliver true multiservice capabilities,
  • High density deeply channelized OTN framing, mapping, and ODU0/ODUflex granular switching,
  • High Density SONET/SDH framing, mapping and switching to enable carriers to transition from SONET/SDH to OTN without stranding their legacy network,
  • Onboard Carrier Ethernet PCS and MACs with integrated packet timing capabilities in order to address the requirements of mobile backhaul in the age of LTE,
  • Packet and OTN fabric interfaces to enable both packet and OTN switching applications,
  • Ability to address OTN, packet and lambda switched deployments with the same device
These features enable the equipment vendor to address all present and future requirements imposed by Metro OTN while minimizing total cost of ownership.


PMC-Sierra has introduced a new family of OTN products that uniquely delivers on the requirements of Metro OTN enabling OEMs to deliver a new class of transport equipment upon which carriers can build their next generation Metro transport networks which are:

  • Multi-service, with seamless transport of Ethernet, Storage, Video, SONET/SDH, and Private Line
  • Scalable with the rapid growth in packet traffic
  • Switchable, providing fine-grain sub-lambda grooming
  • Efficient, especially for the transport of packet centric services
  • Compatible with the core network, providing end-to-end Access-Metro-Core continuity for flexibility, protection and management.
With this new class of equipment, carriers can achieve reduction in  OPEX and CAPEX necessary to enable profitable scalability to support the upcoming 4x growth in network traffic. 

About the Author

As a Product Line Manager in PMC-Sierra’s Communication Products Division, Scott Wakelin has helped define some of the industry’s most successful communication semiconductor solutions including PMC’s HyPHY, TEMUX, and FREEDM product families. Currently focused on packet-optical transport solutions, Mr. Wakelin has over 12 years of experience delivering OTN, SONET/SDH, and Ethernet products to market. Mr. Wakelin holds a Master of Applied Science degree in network infrastructure and security.

About the Company

PMC (Nasdaq: PMCS) is the semiconductor innovator transforming networks that connect, move and store big data. Building on a track record of technology leadership, the company is driving innovation across storage, optical and mobile networks. PMC's highly integrated solutions increase performance and enable next-generation services to accelerate the network transformation. For more information visit


Post a Comment

See also