Study On Ecological Network-Based Resource Dynamic Services Of Next Generation Internet

High automatic requirements of next generation Internet systems, especially service-oriented emergence, diversity, scalability, adaptability, self-organization, autonomy, self-recovery, and survivability, make IT professionals further evolve Internet architecture and technologies for supporting component-oriented applications, scalable resource sharing and large-scale peer-to-peer communications. On the other hand, over millions of years of evolution, large-scale natural ecosystems (including human society) have developed many mechanisms that allow them to scale, adapt, survive, and so on. These mechanisms fulfill exactly the key requirements of the next generation Internet services and applications. Based on our previous work, this paper further applies the mechanisms and principles from natural biological systems and ecosystems into designing a prototype system of the next generation Internet middleware for solving some key issues, such as dynamic resource discovery and composition, and load balance adaptation.First, we overviewed the state of arts of nature-inspired next generation Internet computing. After the introduction of current five types of computing paradigms, we explored the research approaches, relative research projects, design philosophy, system architecture, applications, and development of nature-inspired next generation Internet computing. Finally, we made some prospects on the further work of nature-inspired next generation Internet computing.Then, from the evolution history of grid systems, we discovered the characteristics of the next generation Internet systems and demonstrated the methodology of building such systems. According to real requirements of the next generation Internet systems, we designed a next generation Internet-oriented ecological network platform. The design philosophy, laws, and solutions are also given out. Considering indispensable importance of wireless computing environment in next generation Internet, we provided three key technologies, that is: (1) asynchronous migration and its strategies, (2)wireless QoS adaptation, and (3) wireless security certification. Finally, the software characteristics of ecological network-based platform were discussed.Next, the ecological network-based platform needs a flexible communication mechanism to search, assemble, organize, and coordinate autonomous agents for forming a community and emerging desirable services. From a low-level to high-level of implementation strategies of communication mechanism, we proposed a transport mechanism, Ecological Network Communication Language (ENCL), and Ecological Network Interaction Protocol (ENIP). Also, we gave out the XML-based implementation details of the ENIP and the ENCL messages. Several preliminary performance evaluation of the communication mechanism indicated that the low-level communication implementation worked effectively to reduce overheads in message dispatching and network connection, .and scales well in terms of the number of agents.Moreover, an implementation mechanism that ensures autonomous agents to emerge as a controlled and desired service was explored. A credit card management service in economic interactions was considered as a decentralized control approach. The control scheme, design, and implementation of the credit card management service were presented in detail. Finally, the simulation results showed that collaboration-independent application could emerge under the control strategies.Also, we applied the principles and concepts in social networks to designing a decentralized, survivable, and adaptive resource discovery approach, in which autonomous agents can assist each other in discovering desired resources. The wide-area discovery framework, models, algorithms, and message formats of the approach were presented in detail. A simulator was developed and coded with some representative scenarios being created and evaluated. The simulation results showed that our approach can: (1) form relationship clusters and significantly improve the discovery performance;(2) adapt well to different resource distributions and user request patterns;(3) survive from the changes of dynamic environments, including variable-biased user . requests and agent amounts as well as partial failure of the agents. The simulation still proved that the presented communication mechanism and decentralized controlmechanism benefited to the next generation Internet service and application emergence.Adopting autonomous mobile agent approach is one promising solution to the design of highly available, ubiquitous, self-managing, and adaptive network services. While a fundamental theory on migration model of vast numbers of agents that are geographically dispersed across the Internet is still required. We established a mathematical model for the macro migration behavior of autonomous agents. By analyzing the model, we brought to light that some interesting properties of agent stable distribution. We also found out the relation between the migration rates and agent amount in the stable distribution. These conclusions provided us with an effective measure to improve the macro quality of emergent services.Finally, based on the study on current grid architecture and the characteristics of the next generation grid, we proposed a novel grid computing-oritented next generation Internet system architecture and its prototype system named ecological network-based grid middleware. The design philosophy and key technologies of the middleware were discussed. Then, we introduced the application of our prototype system in supply chain management and bioinformatics fields.The emergence models and algorithms from natural ecological networks and human social networks presented in this paper enrich the research framework of ecological network architecture, and give out some novel concepts and methods for building a scalable, adaptive, self-evolutionary next generation Internet environment. The presented ecological network platform not only provides a good simulation environment for the next generation Internet, but also grows into an ecological network intelligent middleware embedding to current Internet network architecture.