Using the wavelet basis in Recurrent Dynamic Neural Network (RDNN) can improve the failure event estimation of software defect tracking in telecommunications. Non-linearity of the system is represented by proper selection of the wavelet function. This RDNN handles noisy data and enhances the speed of convergence as compared with alternate approaches. A new adaptive RDNN is presented where software deployment testing observations are used to synthesize intrinsic model parameters.

This paper presents a new reliability model for the time series of failures uncovered during testing of software products. The time series can be either an S-shaped curve or a slowing exponential. The model is statistically consistent; the variance of the estimated number of failures tends to zero as the test time increases. This model can be used for managing risks in the deployment of new products that support highly reliable telecommunication services.

This paper addresses risk management during the introduction of a new technology within a live network. Results from system test are combined with data collected from a field experiment with real traffic at selected locations for a limited amount of time. Compared with traditional sequential analysis, the proposed method allows an earlier decision with better estimates for the service reliability.

In this paper, we outline a procedure for quality assurance of network elements before their deployment. Software reliability is assessed using two models: a process-centric model (Musa's (1987) basic model) and a product-centric model (proposed by Hoeflin (2000)). Simultaneous use of both approaches is for sensitivity analysis of the results. In addition, we introduce the concept of deployability to measure the degree of confidence on the decision to deploy the equipment in the field.

Operation support systems (OSS) facilitate the development, deployment and maintenance or high-quality network-based services at a reasonable cost. This paper reviews the trends, particularly the customers' needs to manage their own data networks, and their effects on the architecture of OSS of public data networks.

Protocols such as the real-time transport protocol (RTP), are used to provide suitable voice/video communication channels over the Internet. RTP transports real-time media data along with synchronization information over a datagram protocol such as the user datagram protocol (UDP). The Casner-Jacobson (see Internet-Draft, 1997) RTP/UDP/IP header compression algorithm can significantly reduce the overhead and increase the throughput of a connection. We evaluate the behavior of header compression in noisy links. It is shown that error indication and error correction can improve the resilience of header compression over noisy links.

This paper presents a service architecture that combines SET and SSL to provide a lightweight secure electronic commerce system. Although both protocols secure data over an open network like the Internet, they differ in their mode of operation. SET defines secure transaction-oriented exchanges for payment systems whereas SSL secures point-to-point communications. SET and SSL are complementary and the combined architecture permits a gradual introduction of SET services by reducing the need for cardholder certification.

This paper presents the concept of multiple call establishment through the PSTN for multimedia traffic. The idea is to use new techniques for toll-quality voice compression to allow several calls between the end-user terminal equipment and the PSTN using the same telephone number and on the same local loop. The requirements for a control channel to achieve this functionality are reviewed and the implications of the proposed service are discussed.

This paper provides information on a laboratory study conducted by COMSAT Laboratories to evaluate the effect of echo and delay on facsimile call retention rates. The term "call retention" is defined as the ability of a pair of facsimile terminals to maintain a circuit connection until all pages have been successfully transmitted after negotiation of the high speed signaling rate. The results of this study indicate that the effect of network echo has the potential of degrading facsimile call retention. In addition, propagation delay reduces retention rates even further. The results of this work are currently under consideration by the International Telecommunication Union (formerly CCITT) in the formulation of test methodologies for assessing facsimile terminal implementation performance.

Integrated multiplexing schemes are needed to optimize the use of transmission bandwidth in integrated networks. In this paper, we introduce the idea of a dynamically controlled (T/sub 1/, T/sub 2/) multiplexing scheme. The advantages of this transmission policy is that it allocates the transmission capacity of the channel to two traffic types according to the instantaneous needs. We consider only two types of traffic, voice band traffic and digital data. The channel access times for voice and data are T/sub 1/ and T/sub 2/, respectively. In this paper, we present some preliminary results of our proposed scheme. In our simulations, we observe that a reduction in blocking probability of 15% or more is possible.

There have been significant advances in the design and development of "toll-quality" packetization systems. The paper presents various issues that affect the packetization of speech such as robustness to noise, reconstitution of the flow of speech and congestion control. The target audience consists of designers and implementers of speech packetization systems as well as service providers concerned with the quality of service and efficiency of transmission.

Embedded adaptive differential pulse coded modulation (ADPCM) algorithms quantize the differences between the input signal and the estimated signal into core bits and enhancement bits. CCITT Recommendation G.727, which describes embedded ADPCM encoding algorithms with 5, 4, 3, and 2 core bits, is virtually identical to the corresponding ANSI standard T1.310. The main features of G.727 and T1.310 and performance results are presented. A formal subjective evaluation of the speech performance of embedded ADPCM algorithms indicates that a midrise quantizer provides better voice transmission performance than its midtread counterpart when two core bits are used. The subjective data also show that the performance of the 40-kb/s midrise ADPCM algorithm with two feedback bits is indistinguishable from that of 64-kb/s pulse code modulation (PCM) for up to four tandem encodings. Embedded algorithms are therefore recommended for flexible congestion control of integrated traffic in multinode networks.>

Some methods of supporting voice in broadband ISDN, (B-ISDN) asynchronous transfer mode (ATM), including voice compression, are examined. Techniques for voice compression with variable-length packet format at DS1 transmission rate, e.g., wideband packet technology (WPT), have been successfully implemented utilizing embedded adaptive differential pulse code modulation (ADPCM) coding, digital speech interpolation (DSI), and block-dropping schemes. For supporting voice in B-ISDN, voice compression techniques are considered that are similar to those used in WPT but with different packetization and congestion control methods designed for the fixed-length ATM protocol at high speeds. Possible approaches for packetization and implementation of variable-bit-rate voice coding schemes are described. ADPCM and DSI for voice coding and compression and cell discarding (CD) for congestion control are considered. The advantages of voice compression and CD in broadband ATM networks are demonstrated in terms of transmission bandwidth savings and resiliency of the network during congestion.>