TACS  Delivers the Insight and Vision on Technology for Strategic Decisions

Optical Services

 

     
Home ] Up ] Services of TACS ] Competences of TACS ] Markets of TACS ] About TACS ] Views & News ] Links ] Site Map ] Search ]
The Vision-An All Optical Network ] Broadband Networks ] Optical Networking ] [ Optical Services ] Optical WANs ] Optical MANs ] Optical LANs ] Optical Technology & Architectures ] Competing Optical Technologies ]
 



 

Up

  Optical Services

Optical networks can provide circuit, packet, and cell services at the UNI. Each optical service can be categorized as analog or digital. We consider four major categories of optical services:

        Analog Circuit Service
        Digital Circuit Service
        Analog Packet/Cell Service
        Digital Packet/Cell Service

Analog Circuit Service

Analog circuit service provides a dedicated band-limited optical channel between ONAs. Point-to-point, multi-point, or broadcast connections may be supported. The network guarantees delivery with minimal distortion of any signal conforming to the channel bandwidth and input power specifications. Analog circuits may be full-bandwidth, occupying an entire waveband, or time-shared on a slotted basis.

To establish an analog circuit between ONAs, a circuit is assigned a waveband which is transparently routed through the network. This process is called wavelength routing. If wavelength changers are used within the network, the circuit may terminate on a different waveband than it originated on.

Digital Circuit Service

This service provides a dedicated logical digital channel between the ONAs. The network guarantees delivery with minimal errors of any digital signal conforming to the channel modulation specifications (e.g., solitons). Digital circuits may be full-rate or interleaved on a bit, byte, or slotted basis.

Analog Packet/Cell Service

This provides a packet/ cell service where the packet/cell body is a time slot which is transparently routed through the network. The time slot may contain any analog signal conforming to the channel bandwidth and input power specifications. The packet/cell header is a digital address which is sent out-of-band (e.g., in time or frequency) and travels along with the body on the same path and in synchronization. The header must be interpretable by network nodes and ONAs. There are a multitude of methods for encoding the header.

Digital Packet/Cell Service

Provides a digital packet/cell service analogous to conventional electrical packet/cell services.

If ON user traffic is bursty it might be beneficial to offer packet/cell services in the optical layer (at the UNI) in order to more efficiently utilize optical network resources. The ONAs could then utilize these optical packet/cell services to provide electronic packet/cell services to the ON users. Such an ONA would likely contain buffering (optical or electronic) and might be responsible for such functions as traffic shaping and flow control.

Current technology severely limits the range of optical packet/cell services that can be offered. Memory is perhaps one of the most difficult functions to implement optically.

There seems to be little motivation for digital packet/cell services at conventional electronic rates.

Today current technological limitations allow only analog circuit services (Fig. 3).

Figure 3 - Possible future trends in optical network technology 

Note that that there is an important distinction between the optical services (which are provided by the optical network to the ONAs), and the transport services (which are provided by the ONAs to the ON users). For example, to be specific, ONAs may provide SONET/SDH transport circuits (e.g., OC-48/STM-16) over optical analog circuits by conventional modulation and demodulation of an optical carrier; similarly, an analog transport service can be provided over a digital optical circuit via sampling and reconstruction.

For the most part, transport circuits are carried by optical circuits, and transport packet/cell services are supported by optical packet/cell services. A much researched and demonstrated exception to the latter occurs in LANs and MANs where the distances are short enough to implement a fast circuit-switched optical layer via a medium access control (MAC) protocol. This enables transport packet/cell services over optical broadcast circuits. In addition, it is theoretically possible, but technically difficult, to support transport circuits over optical packets/cells.

In the optical networking community, analog channels are often called fully transparent, or simply transparent, implying that any modulation format may be used on the channel. Very often, analog channels are only transparent to certain modulation types (e.g. amplitude modulation), while nontransparent to others (e.g., phase modulation). Therefore, to fully describe a channel its "level of transparency" must be determined, which is simply a specification of the modulation formats that can be transmitted over the channel with acceptable quality.

There are two fundamental reasons channels may be nontransparent to certain modulations:

        First, the fidelity of the optical channel usually degrades with distance and with the number of signals utilizing a fiber due to additive noise, multi-access interference, and nonlinearities of the optical amplifiers and the fiber itself; therefore, long links which are highly utilized may not support certain modulation types. Because of these difficulties, a significant amount of effort has gone into investigating long-haul WDM links for on-off keying (OOK) and the results have been very successful with 1 Tb/s links demonstrated over ~150 km and 100 Gb/s links demonstrated over 9100 km. Other modulation techniques, the mixing of different modulation techniques on different wavelengths, and the system implications of coding are not fully understood at this time.
        The second reason a channel might be nontransparent to a modulation type is that certain useful network elements may not support certain modulation types. An important example of this is a "gain saturating wavelength changer" which changes the wavelength of an optical signal. These devices use the amplitude of the input signal to amplitude-modulate light on a desired output wavelength. Therefore, all phase information is lost and only amplitude modulated signals are passed. In fact, only OOK signals will pass with sufficient fidelity. Note that the OOK signals can be of any bit rate, but this wavelength conversion process becomes increasingly difficult at rates faster than ~10 Gb/s. When a channel will pass any bit rate of a certain modulation type, it is called bit rate transparent.

Below we describe each service in more detail with an emphasis on the implementation and technical difficulties which make certain services impractical in certain situations.

Before entering that discussion, we first note that each optical service is offered on a waveband (at the UNI) and that there are trade-offs between

        network capability,
        upgradability,
        ONA complexity.

In particular, ONA hardware can be simplified by offering an optical service over a fixed, standardized waveband. The network would then be required to use wavelength converters to convert this waveband to another waveband in order to use WDM within the network. An alternative solution is for ONAs to use wavelength-agile transmitters and receivers, in which case the network would not be required to perform wavelength conversion. The latter is currently far more economical.

 

 

 
 
 
 
 
 
 
 
 
 
 
 
 

 

Home ] Up ]

 


Copyright 1997-2015 TACS
Last modified: July 13, 2016

 

The Best Networks Start with the Best Consultants, TACS