Transparent LAN service (TLS) refers to adjoining multiple sites over a metro solution using a single subnet or VLAN. TLS is easy to implement but presents scalability issues and a single failure domain for all customers using the metro solution. TLS supports point-to-point and point-to-multipoint operation.

In a transparent LAN service, data transmitted across the metro solution acts in a manner similar to an Ethernet hub using a single broadcast domain. This type of behavior presents a single failure domain to all customers of the metro solution. One traffic flow from a single customer within the core network of the metro solution may affect traffic from another customer. Figure 16-5 illustrates a TLS metro solution where the metro solution acts as a single broadcast domain.

Figure 16-5. Sample TLS Topology

For the enterprise using a TLS metro solution, all of its routers and multilayer switches reside on the same subnet on the edge. The service provider easily provides a TLS solution for the enterprise by maintaining the same VLAN across its entire infrastructure. However, a TLS metro solution has two potentially serious performance implications for the enterprise:

The key benefit of TLS is that its shared backbone provides simplicity of implementation. All remote locations appear on a single subnet and in a single VLAN. Service providers and enterprise customers use VLANs to segment and differentiate traffic across a TLS metro solution. However, from a service provider standpoint, only 4096 individual VLANs are available for customer use. Generally, 4096 VLANs are insufficient for service providers who provide multiple VLANs on metro Ethernet solutions to multiple customers. In addition, supporting multicast and implementing QoS on TLS metro solutions is difficult because no discrete method of traffic separation exists.

In summary, only enterprises that purchase their own dark fiber and implement their own metro solution platforms use the TLS design, with or without the use of VLANs. Generally, service providers use directed VLAN services, as discussed in the next section, to take advantage of the cable plant while maintaining differentiated and autonomous metro Ethernet solutions to enterprise customers.

In a directed VLAN service (DVS), the metro solution uses VLAN IDs to forward traffic to select destinations. DVS supports both point-to-point and point-to-multipoint operations. Typically, DVS isolates VLANs of the enterprise from the VLANs of the network by using additional header information. As such, enterprise VLANs are separate from the metro VLANs. One such example is to add a second 802.1Q header to a frame to achieve a Q-in-Q transport.

Figure 16-6. Metro Solution Deploying DVS

The switch in the core sees the VLANs defined at the edge of the network and provides VLAN switching. Note that these VLANs are not the VLANs configured in the multilayer switched network, but rather VLANs locally significant only to the metro solution. In DVS, the core switch must know the enterprise VLAN assignments. This implementation requires Per VLAN Spanning Tree Plus (PVST+).

DVS allows enterprises to support VLANs that appear in multiple locations across the core. In Figure 16-7, VLAN 10 appears in Remote Offices 1 and 3, whereas VLAN 20 appears in remote offices 2 and 3. The core switch in the cloud, generally physically located in Enterprise Edge of the Campus network, provides VLAN switching capability. As a result, the core switch knows about VLAN assignments on the remote location switches.

Figure 16-7. Sample DVS Metro Solution

At a high-level overview, DWDM provides a transparent medium for Ethernet connectivity similar to SONET. In Figure 16-10, the DWDM metro Ethernet network appears to the end user as a point-to-point Gigabit Ethernet link. As with SONET, the metro Ethernet system essentially creates a long extension cord for the Gigabit Ethernet transmission between enterprise campus A and enterprise campus B.

Figure 16-10. Metro Solution Using DWDM

At Layer 1, DWDM converts Ethernet packets on a bit-by-bit basis into an optical stream at 1, 2, or 10 Gbps. DWDM does not perform framing conversion and simply performs optical conversion. As a result, Ethernet over DWDM does not require specific encapsulation. The core network moves data bits transparently from campus to campus. Because DWDM simply converts data bits to an optical stream, DWDM suffices as a transport media for not only Ethernet, but also storage protocols such as Fibre Channel, FICON, and ESCON.

Figure 16-11 illustrates a sample deployment of DWDM for multiple protocols and autonomous networks.

Figure 16-11. Sample Metro Solution Using DWDM

The two most commonly used wavelength regions for optical transmission over long distances are 1310 nanometers (nm) and 1550 nm. Short distance requirements commonly use the 850-nm window. These 1310-nm (S band) and 1550-nm (C band) windows, along with a fourth window at 1625 nm (L band) for upcoming technologies, have significantly lower optical loss compared to other windows near the same wavelength range. Figure 16-12 illustrates the optical loss characteristics at different wavelengths.

Figure 16-12. Optical Loss at Different Wavelengths

Transmissions using the 1310-nm wavelength (band) are popular for metropolitan or short-range applications because of their relatively low cost. The 1310-nm single-mode GBICs are readily available for Catalyst switches. Because transmissions in the 1550-nm band can travel greater distances because of that band’s intrinsic properties, the 1550-nm range is preferable for very long-distance transmissions (between 100 km and 600 km). The increased cost in CPE equipment to process wavelengths in the 1550-nm range is still cheaper than the aggregate cost associated with continuous retransmission of a 1310-nm range optical signal. Figure 16-13 illustrates these windows relative to the electromagnetic spectrum.

Figure 16-13. Wavelength Bands in the Electromagnetic Spectrum

DWDM technology assigns incoming optical signals to specific frequencies of light (wavelengths, or lambdas) within the 1550-nm frequency band. Figure 16-14 illustrates the principle of multiplexing different wavelengths across a single medium.

Figure 16-14. Illustrative DWDM Functional Schematic

This multiplexing closely resembles the way radio stations broadcast on different wavelengths without interfering with each other. Because each channel transmits at a different frequency, a tuner is necessary to select the desired channel.

Current models of the Cisco optical platforms support up to 32 usable lambdas plus a 33^rd lambda for management. DWDM optical platforms use different frequencies and wavelengths of the ITU draft standard G.692 for systems based on 100-GHz wavelength spacing with a center wavelength at 1553.52 nm. Table 16-2 illustrates the ITU grid for these types of DWDM systems.

As mentioned previously, most LAN-based Cisco IOS routers and Catalyst switches use either the 850-nm band or the 1310-nm band for optical transmission. For a DWDM system to transmit a signal in the 1550-nm range, it converts the ingress 850- or 1310-nm signal into a 1550-nm signal using a device called a transponder. Transponders accomplish the wavelength altering by doing an optical-to-electric-to-optical (OEO) conversion. In addition, transponders support regeneration, reshaping, and retiming functionality.

DWDM systems use optical multiplexers and de-multiplexers to combine and separate the multiple wavelengths for transmission and upon reception of optical signals, respectively. Optical amplifiers increase optical signals as they transmit across the optical fiber. Optical amplifiers differ from traditional amplifiers in that the optical signals are not converted to electrical signals for amplification. Erbium doped fiber amplifiers (EDFA) are the most commonly used optical amplifier in today’s DWDM system.

Furthermore, DWDM systems use optical add/drop multiplexers (OADM) to drop or add a single lambda at a specific location. OADMs are useful in aggregating remote sites where each remote site uses only a subset of lambdas. For instance, Figure 16-15 illustrates the use of an OADM. In this DWDM system, the New York metro site only uses one lambda. As a result, the DWDM topology uses an OADM to add and drop a single lambda at this location.

Figure 16-15. Sample OADM Use in DWDM Topology

This section is meant to be only a brief overview of DWDM; for a complete understanding of DWDM, consult Cisco.com.

CWDM technology harnesses the same WDM technologies as DWDM systems. DWDM and CDWM primarily differ in distance ranges, spacing of lambdas, number of channels, and the ability to amplify optical signals. In brief, CDWM uses 8 lambdas using a 20-nm wavelength grid, whereas DWDM technology has 32 or more lambdas using a 1-nm wavelength grid. CWDM technology does allow for optical signal amplification using any external amplifier, which limits its capability to transmit under 100 Km. CWDM basically provides for the same topology as DWDM except the Cisco IOS routers or Catalyst switches originate the signal and OADMs add or drop lambdas at a specific location. As such, CWDM is far less complex than DWDM, requires less hardware, and is easier to install. Figure 16-16 illustrates a sample metro Ethernet over CDWM topology.

Figure 16-16. Sample Metro Ethernet Deployment over CDWM Topology

The CDWM solution uses two main components:

CWDM does support EtherChannel using multiple CWDM GBICs and OADM modules between sites. An alternative to EtherChannel is using multiple Layer 3 equal-cost paths over multiple CWDM connections. In summary, CWDM has the following benefits as a metro Ethernet solution:

The only main downside to deploying CWDM as a metro solution is distance; CWDM does not support amplification, which limits distances to 100 km for 30-dB optical budgets.