SPIE Digital Library Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:
SPIE Defense + Commercial Sensing 2017 | Call for Papers

Journal of Medical Imaging | Learn more


Print PageEmail PageView PDF

Optoelectronics & Communications

Self-managing intelligent optical networks for smarter, tougher, and cheaper systems

Thanks to their self-monitoring and ability to dynamically reconfigure themselves, ‘self-managed networks’ hold great promise in reducing the enormous expenditures currently required to deploy new nodes or upgrade existing links.
1 May 2006, SPIE Newsroom. DOI: 10.1117/2.1200603.0128

When envisioning the 10-year horizon of optical networks, there are certain laudable goals that may be pursued, such as higher capacity, throughput, stability, reconfigurability, flexibility, and security. These network goals come with a rich set of technical challenges.

Today's optical networks function in a fairly static fashion and are built to operate within well-defined specifications.1 These networks require an enormous expenditure in order to deploy new nodes or upgrade existing links. This is quite primitive when compared to a typical wireless local-area-network (LAN) that can accommodate new users in a robust, autonomous manner. Specifically, it might be desirable for a highly-efficient future optical network to accommodate a self-managed, intelligent, well-monitored, and dynamically-reconfigurable network that can accept new nodes in a plug-and-play manner. Ideally, it should also be able to support the convergence of different types of traffic over the same network.

Automated self-managed optical networking

Deployment of current optical links and nodes is a labor-intensive and onerous task due to the numerous system variables that must be balanced. In general, this includes extensive initial measurements of the existing fiber plant and building equipment that has been ‘customized’ to meet the narrow range of specifications for the specific deployment. The resulting equipment then requires expensive, specialized technical experts to fine tune it upon its initial deployment, and as the network changes over time. This scenario is quite challenging for higher-capacity future systems since network paths are not static and channel-degrading effects can change with environmental temperature, component drift, aging, and fiber plant maintenance.2 Moreover, today's equipment does not allow us to simply plug an optical node into an existing network and let the network management and control allocate resources to ensure transport of a high signal quality of service (QoS). These resources include amplifier gain, signal wavelengths, dispersion compensation, path determination, and bandwidth.3 Moreover, reducing the need for a ‘person-in-the-loop’ for routine and even some non-routine tasks may provide enhanced operational efficiency and cost savings.

In order to enable robust ‘self-managed’ automated operation, the network should probably be able to intelligently monitor its own physical state as well as the quality of propagating data signals. It should also be capable of automatically diagnosing and repairing itself, as well as allocating resources where needed, and redirecting traffic (see Figure 1).4 Intelligent optical performance monitoring should isolate the specific cause and location of the problem rather than simply sounding an alarm. Such monitoring could also provide valuable information so that routing tables can reflect the state of the physical links and the addition/deletion of nodes, rather than simply keeping track of the fewest number of links/hops between source and destination.5


Figure 1. (a) An example of a WDM network that can accommodate ‘plug-and-play’ operation and adaptive resource allocation. (b) A graphic showing applications of ubiquitous monitoring.

An added benefit of monitoring is the security to prevent denial of service. As a simple example, an unwelcome high-power wavelength that is added and dropped somewhere in the middle of a long-distance link could cause severely-degrading nonlinear effects on all the existing channels.6 If this wavelength disappears, the network returns to normal without any lasting trace of the ‘culprit.’

Dispersion—an example of optical performance monitoring

Monitoring of signal quality should be simple and sensitive, and it could provide an ‘error’ signal for a compensator or equalizer. It is valuable to think of many degrading effects as reducing the signal integrity of the 1 and 0 bits, even though there may still be a high optical signal power. The monitoring of these non-catastrophic problems is therefore much more difficult that simply determining a fiber cut. One straightforward monitoring method is to tap off some optical power, use a high-speed photo-receiver to recover the bits, and then measure the quality or bit-error-rate. Unfortunately, this approach cannot distinguish between the various transmission-based channel-degrading effects such as chromatic dispersion, polarization-mode-dispersion (PMD),7 and nonlinearities.

Chromatic dispersion is a well-known effect that arises from the frequency-dependent light propagation speed in an optical fiber. A technique for chromatic dispersion monitoring uses a tunable narrowband optical filter (i.e., the passband is narrower than the channel's bandwidth) to select first the upper and then the lower halves of the transmitted data spectrum in order determine the relative group delay caused by dispersion (see Figure 2).8 The amount of chromatic dispersion can be determined by using a tunable filter to obtain upper and lower sidebands from the signal and then comparing the time delays of the two recovered data pulses. Using this measurement technique, the phase delay (in terms of nanoseconds) between pulses can be directly correlated to the frequency spread between the sidebands (in terms of nanometers). The time delay can be measured by measuring the phase of the clock component in the two sets of data seen by the photodetector. Note that many wavelength-division multiplexed (WDM) channels can be accommodated by sweeping the optical filter across all channels. The effectiveness of this technique was verified experimentally for a 40Gbit/s data signal. A key advantage of this technique is that it isolates chromatic dispersion from other deleterious effects. (Note that PMD and optical signal-to-noise ratio—OSNR—monitoring can also be obtained through a different use of a tunable narrowband filter.9,10)

Figure 2. An illustration of chromatic-dispersion monitoring using narrowband filtering. The amount of dispersion can be determined through a comparison of the time delay between the data pulses recovered from the upper and lower sidebands.

Data traffic convergence

The future optical network will probably be used by many users for a variety of applications, each with a different set of optimal requirements. In these circumstances, it is quite possible that there will be a wide variety of data formats that might need to be simultaneously transmitted over the network. It seems fairly inefficient to build a separate optical network to accommodate each application. Instead, there could be one network to accommodate a wide range of data rates services, and requisite and QoS levels. Likewise, a future optical network might also be required to support both circuit and packet switched traffic, multiple modulation formats, as well as concurrently carrying commercial and military traffic in a secure manner. Given that some information is undesirably lost when digitizing analog signals, one can even envision future scenarios for which one network might be required to carry both digital and analog channels. Of course, WDM of parallel channels in the fiber readily lends itself to mixing different data types within the optical network. A key issue is to dynamically and optimally allocate system bandwidth where it is most needed, whether to an individual channel or to a band of channels.11

Alan Willner
Electrical Engineering, University of Southern California
Los Angeles, CA
Alan Willner received his Ph.D. from Columbia University, has worked at AT&T Bell Labs and Bellcore, and is Professor of Electrical Engineering at University of Southern California. He has received the NSF Presidential Faculty Fellows Award from the White House. He has received numerous other awards including the Packard Foundation Fellowship, the NSF national Young Investigator Award, the Fulbright Foundation Senior Scholars Award, the IEEE LEOS Distinguished Traveling Lecturer Award, the University of Southern California-University-Wide Award for Excellence in Teaching, and the IEEE and OSA Fellow. Prof. Willner's activities have included: President of the IEEE LEOS, Editor-in-Chief of the IEEE/OSA Journal of Lightwave Technology, Editor-in-Chief of the IEEE Journal of Selected Topics in Quantum Electronics, Co-Chair of the OSA Science and Engineering Council, General Co-Chair of CLEO, and Steering and Program Committee Member of OFC. In addition to presenting many papers, Prof. Willner has served as a member of the International Steering Committee and Technical Program Committee for the Asia-Pacific Optical and Wireless Communications Conference (APOC), and as program chair for the Telecommunications Engineering sessions at Photonics West.

1. T. Chen, L. Paraschis,
Proc. SPIE,
Vol: 5626-45, 2004.
2. A. E. Willner, B. Hoanca,
Optical Fiber Telecommunications IVB,
Vol: 14, 2002.
3. D.C. Kilper,
IEEE/OSA Journal of Lightwave Technology,
Vol: 22, pp. 294-304, 2004.
4. A. E. Willner,
Proc. SPIE,
Vol: 5882, 2003.
5. X. Yufeng, G. N. Rouskas, H. G. Perros,
IEEE/OSA Journal of Lightwave Technology,
Vol: 21, pp. 904-915, 2003.
6. A. R. Chraplyvy,
IEEE/OSA Journal of Lightwave Technology,
Vol: 8, pp. 1548-1557, 1990.
7. H. Kogelnik, L. E. Nelson, J. P. Gordon,
IEEE/OSA Journal of Lightwave Technology,
Vol: 21, pp. 482-495, 2003.
8. Q. Yu,
IEEE/OSA Journal of Lightwave Technology,
Vol: 20, pp. 2267-2271, 2002.
9. T. Luo,
IEEE Photonics Technology Letters,
Vol: 18, pp. 430-432, 2006.
10. L.-S. Yan,
IEEE/OSA Journal of Lightwave Technology,
Vol: 23, pp. 3290-3294, 2005.
11. K. Son,
IEEE Int'l Conf. on Comm.,
pp. 1766-1770, 2004.