In this paper, a new family of wavelength-time codes with expanded code cardinality and the maximum cross-correlation function of 2 (i.e., lambda_c=2) is constructed and analyzed. One application of the large code cardinality of our lambda_c=2 codes is multicode-keying wavelength-time optical code division multiple access (O-CDMA), in which each user is allocated with multiple code matrices, instead of just one code matrix in the conventional ON-OFF keying (OOK) O-CDMA. System throughput is increased because a lower baud rate O-CDMA system can be used to support higher bit-rate transmission since each code matrix is used to represent a "symbol" of several data bits. User code confidentiality is improved because of symbol transmission. The performances of two multicode-keying O-CDMA schemes with the new lambda_c=2 wavelength-time codes are also analyzed. The results in this paper show that there is a tradeoff between the performance and the number of code matrices per user

To support multimedia services with different discrete bit-rate requirements, families of multilength optical codes, such as the carrier-hopping prime code (CHPC), extended CHPC, and multiwavelength optical orthogonal code, were recently constructed for wavelength-time optical code-division multiple-access (O-CDMA). In this paper, the performances of these multilength optical codes in a multimedia O-CDMA system with a variable-weight operation are analyzed. Our study shows that short-length codes generate stronger interference than long-length codes. This supports services prioritization in O-CDMA. Our study also shows that code weight is a more important factor than code length in determining code performance.

Quadratic-congruence carrier-hopping prime codes (QC-CHPCs) with zero autocorrelation sidelobes, cross-correlation values of at most two, and expanded cardinality were recently constructed for wavelength-time "multicode-keying" optical code division multiple access (O-CDMA) for improved data throughput and code obscurity. To support multimedia services with different discrete bit-rate requirements, "multiple-length" QC-CHPCs are constructed algebraically in this paper. In contrary to conventional single-length codes, our analysis shows that the performance of these multiple-length codes improves as the code length decreases, thus supporting services prioritization in O-CDMA. Moreover, the relationship of the normalized spectral efficiency (NSE) and code lengths of the multiple-length QC-CHPCs is studied. Our results show that the NSE improves as the number of simultaneous users with short code matrices increases, which, however, decreases the total number of simultaneous users in the system. The choice of which code-length distribution to use depends on whether system efficiency or total number of simultaneous users is more important.

Optics InfoBase is the Optical Society's online library for flagship journals, partnered and copublished journals, and recent proceedings from OSA conferences.

Two-dimensional (2D) wavelength-time coding schemes have recently been studied for optical code-division multiple access. In this paper, we propose a new family of 2D codes, which use "synchronized" quadratic congruence sequences for wavelength hopping and (n, w, 2, 2) optical orthogonal codes for time spreading. The new 2D codes provide larger (asymptotically optimal) code cardinality than other 2D codes by relaxing the maximum cross-correlation function from one to two. Our study shows that the heavier code weight supported by our 2D codes results in better code performance than some of previously proposed 2D codes (of the maximum cross-correlation functions of one and two). The new codes potentially support more subscribers and heavier bursty asynchronous traffic with gradual performance degradation as traffic load increases.

To provide differential quality-of-service in optical code-division multiple access, variable-weight codes were proposed and analyzed under a conventional assumption of identical power in every optical pulses (i.e., same chip power). To account for power limit of some laser sources, we assume the use of same optical power in each bit duration (i.e., same bit power) in this paper. For illustration, we focus our study on double-weight codes under both same-chip-power and same-bit-power assumptions. A new and better approximate performance-analytical model for double-weight codes is developed and validated with computer simulation. We show that heavy-weight codes do not always perform better than light-weight codes under the same-bit-power assumption, contrary to previous findings under the same-chip-power assumption.