Xiaoyuan Gu

Network-centric music performance: practice and experiments

Advances in information technology and the great proliferation of the Internet have changed nearly every aspect of the work and life of human beings. Despite progress in networked entertainment, many music professionals and enthusiasts are still sticking to the traditional way of carrying out rehearsals and concerts. Music performance in this way requires physical presence of the participants and has a number of inherent limitations. We introduce a novel system called network-centric music performance (NMP) that enables multiparty music performance through cyberspace. Our target is to support real-time multichannel natural audio streaming over the network, using audio compression schemes that can provide acceptable audio quality. A system like this is bandwidth-demanding and highly delay-sensitive, and requires synchronization of the audio streams. Hence, support from the underlying end systems and networks is critical. However, the current source coding mechanisms and the best effort nature of the internet pose many challenges to achieve the desired quality of service. We have implemented a prototype of NMP, and exploited end system and network influences on NMP. The work was done in a LAN environment using Linux PCs. The system enables two different application scenarios: real-time rehearsal and rehearsal on demand. Real-time multichannel audio transport and different audio compression schemes are supported. Our evaluation results based on both subjective and objective measurements show that the system provides sufficient audio quality level for the target application in such an environment. The scalability test also revealed that the system scales well with increase of clientele. In the future, we extend our system for networks spanning larger distances and experiment with more realistic network conditions in the Internet.

Real-time service provisioning for mobile and wireless networks

As mobile devices and wireless networks are becoming ubiquitous, the interest of users to deploy real-time applications, e.g. online gaming or Voice-over-IP in such environments is also increasing. Due to the difference between traditional and wireless networks, in particular in terms of available bandwidth and network structure, the concepts used for supporting real-time applications in both networks are different. This paper gives an overview of the key technical challenges that are fundamental and need to be solved in order to easily support real-time applications in wireless and mobile environments. In a first step, issues related to service provisioning in mobile networks are discussed. This is followed by a look at the Quality of Service supported by wireless networks and possible techniques for improving it. Finally, concepts for securing the communication between the users of real-time applications in wireless and mobile networks are presented. For each of these issues, we provide a detailed analysis and an overview of the state-of-the-art. Moreover, we illustrate the main points using distributed online games as an example.

Towards self-optimizing protocol stack for autonomic communication: initial experience

The Internet is facing ever-increasing complexity in the construction, configuration and management of heterogeneous networks. New communication paradigms are undermining its original design principles. The mobile Internet demands a level of optimum that is hard to achieve with a strictly-layered protocol stack. Questioning if layering is still an adequate foundation for autonomic protocol stack design, we study the state-of-the-art from both the layered camp and its counterpart. We then outline our vision on protocol stack design for autonomic communication with the POEM model and its internals. A novel cross-layer design approach that combines the advantages of layering and the benefits of holistic and systematic cross-layer optimization is at the core of this work. With inspirations from the natural ecosystem, we are working on the role-based Composable Functional System for self-optimization that features proactive monitoring and control. By doing so step-by-step, we envisage reaching the goal of self-tuning autonomic network with high level of autonomy and efficiency, with minimum human management complexity and user intervention.

Enabling network-centric music performance in wide-area networks

NMP on the Internet is not only possible, its delay bounds can satisfy the tight requirements involved in human perception while generating high-end audio quality for musicians and listeners alike.

NMP - a new networked music performance system

Although IT has penetrated into nearly every aspect of the work and life of human beings, music professionals still stick to the traditional way of carrying out rehearsals and concerts. Music performance in this way requires physical presence of the participants and has a number of inherent limitations. We introduce in this paper a prototype of a novel networked music performance system that enables the music professionals and enthusiasts to play with each other through cyberspace. An application like this is bandwidth demanding, highly delay-sensitive and requires the synchronization of the audio streams. Hence, the support from underlying end-systems and networks is critical. However, the current source coding mechanisms and the best-effort nature of the Internet poses many challenges to achieve the desired quality of service. We have implemented the prototype in a local area network environment on Linux PCs. The system contains four major interactive components and enables two different application scenarios of real-time rehearsal and rehearsal on-demand multi-channel audio transport and different audio compression schemes are supported. Our evaluation results based on both subjective and objective measurements show that the system suffices the audio quality level for the target application. In the future, we extend our system to larger scale and consider more realistic network conditions in the Internet. Moreover, the support for MPEG-4 AAC codec for even better quality of multi-channel audio streaming and enhanced audio quality under the same bandwidth constrains are added.

Autonomic Multimedia Communications: Where Are We Now?

Internet protocol-based multimedia applications are gaining momentum thanks to the great proliferation of the Internet and mass deployment of broadband. The abundance of rich content, heterogeneity of networks, user terminals, and services, along with the surges of multimedia traffic, have convoluted ever-increasing complexity and costs in the operation and management of multimedia systems. Autonomic communications (AutoComm) promotes the autonomy of communication networks with minimum human administration. In a multimedia system that features AutoComm, the networks and their elements strive to realize efficiency, immunity, resilience, and evolvability through intrinsic self-governance. By encapsulating complexity within the networks, exploiting their intelligence, and using technology to manage technology, human administrative efforts can be greatly reduced while management efficiency can be maximized. In this paper, we want to illustrate how traditional way of management and control of multimedia networks can be gradually replaced by network autonomy and self-management. Related work is extensively reviewed and compared to our approach. Initial experiment results are presented as proof of concept.

Performance evaluation of DCCP: A focus on smoothness and TCP-friendliness

Recent years have seen dramatic increases of the use of multimedia applications on the Internet, which typically either lack congestion control or use proprietary congestion control mechanisms. This can easily cause congestion collapse or compatibility problems. Datagram Congestion Control Protocol (Dccp) fills the gap betweenUdp andTcp, featuring congestion control rather than reliability for packet-switched rich content delivery with high degree of flexibility. We present aDccp model designed and implemented withOpnet Modeler, and the experiments and evaluation focused on largely the smoothness of the data rates, and the fairness between concurrentDccp flows andTcp flows. We foundDccp-ccid3 demonstrates stable data rates under different scenarios, and the fairness betweenDccp andTcp is only achieved under certain conditions. We also validated that the throughput ofDccp-Ccid3 is proportional to the average packet size, and relatively fixed packet size is critical for the optimal operation ofDccp. Problems in the slow start phase and insufficient receiver buffer size were identified and we hereby proposed solutions on this.RésuméLe nombre d’applications multimédias sur l’internet a considérablement augmenté ces dernières années. Ces applications généralement ne gèrent pas les congestions ou bien utilisent des mécanismes de contrôle propriétaires. Il peut en résulter facilement des congestions graves ou des problèmes d’incompatibilité. Le protocoleDccp (Datagram Congestion Control Protocol) se positionne entreUdp etTcp puisqu ’il fournit, pour des paquets multimédias, un contrôle de congestion hautement flexible mais sans la fiabilité de délivrance. Cet article présente un modèleDccp conçu et implémenté avec le simulateurOpnet. Les expérimentations et les analyses portent principalement sur le lissage des débits et l’équité entre des fluxDccp etTcp concurrents. On montre queDccp-Ccid3 affiche des débits stables sous divers scénarios et que l’équité entreDccp etTcp n’est atteinte que sous certaines conditions. On valide également que le débit deDccp-Ccid3 est proportionnel à la taille moyenne des paquets et que le fait d’avoir des paquets de taille relativement constante est un facteur essentiel pour un fonctionnement optimal deDccp. Les problèmes pendant la phase de démarrage lent et ceux dus à une mémoire tampon insuffisante côté récepteur sont également analysés et résolus.

A Proactive Policy-Based Management Approach Towards Autonomic Communications

The growth and evolution of the Internet and the rapid adoption of new wireless, mobile, and self-organizing networks and new services are constantly adding complexity and costs in network management. The traditional manual and static network management paradigm fits poorly into the changing environment where human efforts should be kept at a minimum, and gradually replaced by technologies themselves. We introduce a novel network self-management scheme. Our goal is to seek the synergy between autonomic management, policy-based manage- ment, and autonomic control. They all form the Architecture for Network Autonomy (ANA) that strives to realize proactive policy- based network management for autonomic communications. In this paper, we study the related work, describe the design considerations of ANA as well as its current implementation status, and present the initial results gathered. The Internet continues to grow and evolute, with rapid adoption of new wireless, mobile, and broadband networks. The trend of everything over IP and IP connects everything is being made self-evident by a myriad of sensors, devices, terminals, hosts that are penetrating into each single domain of modern work and life. Numerous services and applications, span WANs, LANs, vehicles, offices, homes, down to human bodies. All of this has convoluted the crisis of Internet manage- ment, in particular the spatio-temporal complexity of network dynamics, and the management and control difficulties. The traditional way of manual planning, configuration, trouble- shooting, and performance tuning will be exorbitantly ex- pensive or even dominate operational cost, as opposite to hardware/software improvements that continuously help to reduce capital expenses. The shortage of skilled labor and the limits of human management of complex systems are complicating the problem further. Besides the looming complexity, the relatively static and time-consuming procedure of conventional network manage- ment calls for new efforts to react to the dynamics and scale of the networks (considering their sheer sizes and traffic volume) in a timely manner. Even more essential is the mapping and bridging between the high-level management constraints and low-level configuration rules, with the latter being the implementation of the former. In doing so the users and administrators can remain control at a high and abstract level by specifying only objectives, and the system will self-govern itself towards realizing the high-level policies, and behaves as expected. Such paradigm requires new system architecture, network model, protocol stack, as well as algorithms for network self-management. In the rest of this paper, we first review the related work in Section 2 that covers three main areas, namely autonomic communications, policy-based management, and the semantic web technologies. Based on this, we present the architecture for network autonomy in Section 3. This is followed by Section 4 on the details of the design and implementation of our prototype. We also show some initial evaluation results in Section 5. Finally, Section 6 highlights the concluding remarks and outlines the directions of future work.

To layer or not to layer: architectural considerations on autonomic communications

We introduce a new communication paradigm dubbed Autonomic Communications, for which context-awareness and distributed policy-based management plays a key role. An extensive study of network architectures is performed, followed by our vision on network architecture for Autonomic Communications, with the POEM model and its protocol design.

Seamless Mobility and Autonomic Computing

The aim of Seamless Mobility is to provide a world of simple, intuitive, and uninterrupted access to information based on the users context. This paper describes a unique architecture that uses autonomic computing principles, anchored by information models, ontological engineering, machine-based learning and reasoning, and policy based management.