Because of the Internet and broadband networks were introduced, the emerging applications – such as teleconferencing, video on demand, and high quality audio transmission has strict high flux optical access network quality of service (QoS) capabilities. However, the infrastructure of current access networks suffers from limited bandwidth, high network-management costs, poor flexibility, and low security, which prevent networks from delivering integrated services to their users. Due to sophisticated optical components and electronic circuits, optical fiber links has become used in access networks. Passive optical network (PON) and different multiplexing technology is put forward under this background, including the wavelength division multiplexing (WDM), 1 time division multiplex (TDM), 2 and optical code division multiplexing (OCDM).

PONs have been standardized for FTTH solutions and are deployed by network-service providers worldwide. Even though PONs based on TDM (TDM-PON) effectively use fiber bandwidths, they have limitations regarding transmission speed, burst synchronization, security, dynamic bandwidth allocation, and ranging accuracy. Wavelength division multiplexing (WDM) technology has also been proposed for PONs. When used in conjunction with PONs (WDM-PON), this emerging technology becomes more favorable as the required bandwidth increases, but it failed to attract attention from industry due to the high cost of optical component. Other schemes for optical access networks are currently under study worldwide.

Optical code division multiplexing access (OCDMA) systems have attracted attention in recent years, because the number of advantages, including its asynchronous access ability and flexibility of user distribution, support for variable bit rate, traffic and security “bursts” unauthorized users. OCDMA is an very attractive multi-access technique for access systems like local-area networks and the first mile, but no detailed network -design schemes have been developed to date. A PON based on code-division multiplexing access (CDMA) has been proposed using pseudo-random and Walsh codes for user identification. However, signature processing for multiple access is done in the electrical domain using an application – specific integrated circuit (ASIC), and not in the optical field as we are pursuing.

We have developed the OCDMA-PON, a network structure of PON in conjunction with OCDMA, i.e., a different multiple-access technology from TDM and WDM. OCDMA technology achieves signature processing light rather than the electrical domain using an optical encoder/decoder.

Optical-line terminator (OLT) of the optical-code-division multiplexing access/passive optical network (OCDMA-PON) system. Rx, Tx: Receiver, transmitter. OOC: Optical orthogonal code. ODN: Optical distribution network. ONU: Optical network unit. ONT: Optical network termination. n: Counter.

For the forward channels, the source is encoded at the OLT and the downstream signal is transmitted at a wavelength of 1550nm. Every user is a assigned a unique optical orthogonal code and identified by a correlation operation at the optical decoder based on fiber Bragg gratings (FBGs). To reduce multi-user interference (MUI), the nonflattened source spectrum can be compensated by a flatness +compensator before entering the FBG encoder. Another benefit is a flattened source is that it relaxes the accuracy requirements regarding the achievable precision with which the phase can be controlled and maintained stably when the number of users increases. For the backward channel from the optical network unit (ONU) optical network terminal (ONT) to the OLT, the upstream is transmitted at a wavelength of 1310nm. After upstream traffic passes the optical transceiver module, which it is sent to multiple decoders, each of which recovers the information for each user.

Upstream of a stable wavelength is usually required with a stabilized laser source at the ONU’s transmitter. The downstream signal from the OLT to ONU/ONT passes through a circulator to the detector, where the user’s information is separated through optical correlation with his or her unique OCC using a balanced receiver. The downstream control signal is also obtained and the control unit is passed to the network. For the upstream, the signal from the ONU to the OLT is encoded by the OCC for user identification by the optical encoder. This is then transmitted through optical fiber link from tifert. Our scheme has several advantages. For example, any user may add or drop into the network at random and the network is running asynchronously.

Our OCDMA-PON system is composed of an OLT and an ONU. Every ONU is identified by its own cde address. The signal is modulated with both frame information and an address-code sequence. The former is used to complete the data load switch, which help to identify different users.

Figure 1 compares the bit-error rate (BER) with the ONU-transmitter input power for two cases, one affected by interference (for example, MUI or from different noise contributions) and a second based on a back to back configuration. In this case, we assume that the amount ONUs/ ONTs is 30 and the bit rate of downstream flow is 1.25gb/s. We found that the sytem that includes interference exhibits much worse performance than that using the back to the back configuration, with approximately 6dB penalty at a BER of 10−9. For the OCDMA-PON, we confirm that MUI is the dominant degradation source (see Figure 3) and must be included in the network design.

Figure 1. Performance of the OCDMA-PON for back-to-back configuration and with interference. N-active: Number of active users. R-downstream: Downstream-traffic bit rate.

Figure 2 compares the BER performance as the number of active users increases for the CDMA-PON and OCDMA-PON systems. We assumed that the bit rate of downstream traffic is 1.25Gb/s for a fiber-link length of 10km. The OCMDA-PON scheme exhibits a similar performance to the CDMA-PON when the number of users is large. For example, for 30 users and a BER of 10−9, OCDMA-PONs can only increase the number of users compared to that supported by CDMA-PONs by 10%. However, the advantage of OCDMA-PONs over CDMA-PONs becomes obvious when the number of users supported by the OCDMA-PON is small (see Figure 4). Note that the signature is processed in the electrical domain by ASIC with Walsh code for user identification in CDMA-PONs, while the optical domain is used by the FBG encoder/decoder based on source-spectrum flattening and a balanced detector with the prime codes for user identification in OCDMA-PONs. Taking into account the higher bit rate and higher bandwidth offered by OCDMA-PONs based on optical processing of user signatures, we still prefer an OCDMA-PON scheme with intensity modulation. To further improve the performance of the OCDMA-PON, we need to solve the MUI-imposed degradation problems and improve the optical encoder/decoder.

Figure 2. CDMA-PON versus OCDMA-PON comparison as the number of ONUs/ONTs increases. L: Length.

Our OCDMA-PON system is different from classical PONs in that OCDMA is used for multiple-user access instead of TDM or WDM, and the OCDMA-PON scheme combines the advantages of both the PON and OCDMA technologies, including flexible network assembly, fair bandwidth division, differentiated services or QoS in teh physical layer, Asynchronous access, support for variable bit rate and burst, and the security of an unauthorized user traffic. Based on comprehensive comparisons between the OCDMA-PON and the previously studied CDMA-PON, we find that the former scheme exhibits a better BER performance than the latter for small numbers of users.

In summary, OCDMA-PON offers several advantages over CDMA-PON, including higher bit rate, higher bandwidth, and better security against unauthorized users. We continue to work towards realizing an experimental demonstration of the OCDMA-PON system.