40 Gbps and 100 Gbps services on metro networks

Long-haul 40 Gbps and 100 Gbps ITU / OTN transponders, based on active chassis-based and vendor proprietary systems, have been on the market for some time now.  As those are targeted for very challenging transport over hundreds if not thousands of kilometres, solutions, all solutions rely on complex technologies involving differing modulation formats such as DPSK (Differential Phase Shift Keying), DQPSK (Differential Quadrature PSK), DP-QPSK (Dual Polarisation-QPSK) etc. Since this complexity does not come for free, the IEEE bodies have standardised pluggable 40 Gbps and 100 Gbps transceivers for "shorter reaches", and while limiting the requirements have reduced the associated cost tremendously. Although the majority of those transceivers cover very short distances (e.g. 10 m, 100 m) the 40GBase-LR4 and 100GBase-LR4 versions are capable of being used for metro network applications and associated distances of up to 60 km.

Overlaying 10 Gbps networks with additional 40 Gbps or 100 Gbps services

In a typical scenario a metro network already exists, often relying on multiple 1 Gbps or 10 Gbps services which are multiplexed over the dark fiber network. The goal therefore, is not to take down the network and replace it by a new infrastructure but to add the new services without affecting the legacy installation, at lowest possible CAPEX and OPEX, while still being simple to install and maintain. This can easily be achieved by adding the 40 Gbps / 100 Gbps service via a passive transport approach. In this scenario the relevant 40GBase-LR4 / 100GBase-LR4 (or future -ER4) transceiver (as e.g. CFP, QSFP, CFP2 etc. MSA format) will be  directly plugged into the terminal equipment, into the Ethernet switch, router etc. The new service can then be overlaid on the existing DWDM (or CWDM) network by adding it through a 1310 nm band pass port on the (maybe existing) DWDM (or CWDM) passive multiplexer:

If this 1310 nm band port is integrated into a 40 channel 100 GHz DWDM passive multiplexer, then this set-up allows the transport of up to 40 services at 10 Gbps plus 100 Gbps over 1 fiber pair, in total 500 Gbps, implemented into 1HU/19" rack space without consuming any electric power.

Overcoming reach limitations

Although the IEEE bodies have standardised LR4 and ER4 versions for 10 km and 40 km reach, which are both operating in the 1310 nm optical spectrum, the availability of the later ER4 (40 km) versions is "limited". In practice this means ER4 transceivers are commercially not available and it is questionable if or when that might change. This effectively limits the use of metro applicable 40GBase and 100GBase to the LR4 versions. Those are nominally addressing distances of "up to 10km" - which are achieved for a perfect single mode fiber, without losses of splices, patch cords, ODFs or other reasons. Moreover the additional multiplexer will insert extra losses, hence the effective reach of such an LR4 transceiver will decrease to 6 km or maybe even only 4 km. This would then limit the use very drastically to some rare cases.

However, that obstacle may be solved by introducing a stand-alone but managed (SNMP etc) SOA (Semiconductor Optical Amplifier): since the LR4 optics operate in the 1310 nm range the usual metro transport EDFA (Erbium Doped Fiber Amplifiers) supporting the 1550 nm spectrum would not only not amplify but even block the LR4 signal. In contrast to EDFAs the SOAs are amplifying the 1310 nm spectrum but vice versa are blocking all 1550 nm signals. Accordingly the additional SOAs must be implemented between the 1310 nm multiplexer port and the LR4 transceiver.

As with any amplified transmission here as well the signal quality and signal to noise ratio is of high importance. The SOAs therefore have to be used as pre-amps (and not boosters) to amplify the incoming (Rx) signal on the receiving side of the link. Preferably and to extend reaches to the absolute maximum LR4 transceivers should be selected that exceed their specification in terms of signal quality. Doing so enables distances of up to 60 km reach.

Passive WDM transport means can be used to set up multiple, up to 24, 100 Gbps services in the same way that passive DWDM 10 Gbps networks have been used to augment capacity in metro environments. And similar to passive 10 G WDM networks, passive 100 G DWDM networks are installed at a fraction of the cost of the active transport counterparts.

Similar to the LR4 versions featuring 4 lasers and 4 detectors (each at 25 Gbp, each at specific lambdas), the heart of passive 100 Gbps DWDM solutions are 100 Gbps DWDM CFP MSA compliant transceivers - they only come without the optical WDM multiplexer integrated into the transceiver. This requires an external (e.g. 19" based, passive DWDM) multiplexer.

In contrast to CFP LR4 types, the CFP DWDM versions are based on tunable lasers (50 GHz DWDM grid) in the 1550 nm spectrum. A set of up to 24  differently coloured 100Gbps DWDM can be transported via a 96 channel passive DWDM mux (each CFP uses 4 lambdas) in parallel over a single-mode fiber pair.

As with 10 Gbps DWDM networks, those signals can be reach-extended by the use of standard EDFA (Erbium Doped fiber Amplifier) amplifiers. This enables the transport of up to 2.4 Tbps in a passive WDM manner without additional and costly amplifiers over distances of a hundred and more kilometres.