The Making of an All-Optical Buffer

There has been tremendous progress in research and commercialization of dense wavelength division multiplexing (DWDM) optical fiber communications. Transmission capacity as high as 10 Terabits/second through a single fiber has been demonstrated in laboratories. This huge capacity can create enormous data traffic congestions at major interconnections. An all-optical packet switched network can potentially eliminate this major bottleneck, by allowing the data packets to remain in the optical domain and to route through the
network towards a final destination without optoelectronic conversion. Electronic routers at sub- Terabits/second rate exist today. The scalability to higher throughput is very difficult. And even when realizable, the power and space demanded by such a router make the electronic routers highly undesirable.
One of the most important components in a router is a buffer. A buffer must be able to store the data packets for a substantial period of time and must be able to release the data within an acceptable delay when the switch is clear for routing. There have been many research efforts on all-optical packet switching. However, there have not been any optical buffers with the necessary properties. Fiber delay lines have previously been referred to as an "optical buffer". However, since the delay is for a fixed amount of time, there is no way to guarantee contention-free connections in the optical switch or through the network. It clearly does not meet the necessary requirements for an optical buffer.
In this program, we propose to work on a novel all-optical buffer with variable memory. The basic idea centers on slowing down the group velocity of the optical data packet in the buffer with a controlled reduction, such that it is effectively an optical memory. By varying the group velocity reduction factor, the memory length and the delay time can be adjusted. It is essential that we engineer the buffer such that a large velocity reduction can be obtained without much pulse dispersion or optical loss.
We propose to develop quantum-dot III-V devices to realize a room-temperature optical buffer. The research will include material research to fabricate quantum dots on III-V compounds, theoretical modeling of such material and its coherence property, and experiments to verify these properties. By providing much sharper quantum confined
energy levels than conventional QWs, we expect to improve the EIT temperature to room temperature and reduce the required optical control beam intensity to a much smaller level. We expect to achieve group velocity reduction and thus switchable optical memory in such samples. Experiments will be designed to explore key controlling parameters of memory size or velocity reduction factor.