To address the limitations in description，positioning，and real-time performance as well as the dynamic scheduling requirements of traditional identifier systems in the internet of computing，this paper proposes a computing resource identifier system for the internet of computing. Firstly，it designs the overall architecture of the computing resource identifier system oriented to the internet of computing. Secondly，it clarifies the mechanisms of computing resource identifier collection，verification，resolution，and identity authentication. Finally，it introduces the practical applications based on the computing resource identifier system. Compared with existing identifier systems，the proposed computing resource identifier system demonstrates significant advantages in comprehensive resource description，dynamic resource perception，and real-time positioning and addressing，providing theoretical support and a technical foundation for advancing the development of the internet of computing.
As a new type of digital infrastructure，the internet of computing （IoC） serves as the foundational support for the high-quality development of the digital economy. Its characteristics of logical centralisation，physical decentralisation and heterogeneous resources have disrupted traditional security boundaries，giving rise to complex risks. This paper systematically analyses security challenges across IoC’s three-tier architecture，synthesising a comprehensive risk framework. By integrating core concepts such as trusted computing，reliable operation，and AI-enabled security，it constructs a security and trustworthy framework for IoC while proposing governance pathways. This aims to provide theoretical guidance for security practices within IoC and lay a robust security foundation for building a integrated computing power network.
With the explosive growth in demand for large models inference，traditional centralized or static multi-data center deployment models face severe challenges in latency，data compliance，and resource elasticity. This paper proposes a cloud-edge collaborative wide-area distributed inference network architecture，focusing on building a new intelligent-computing service system for the emerging computing-power internet. The architecture introduces a prefill-decode separation mechanism： the latency-sensitive prefill stage is offloaded to edge nodes closer to data sources，while the high-throughput decode stage is deployed in the central cloud，enabling secure collaboration over a wide-area network.
To address the need for analyzing the computing resource identifier system aimed at achieving resource-aware description and positioning addressing，this paper proposes a DNS-like hierarchical resolution mechanism and implements a three-level resolution processing architecture for computing resource identifiers. Verified in an optical fiber interconnected experimental environment and large-scale identifier scenarios，the mechanism’s resolution performance and resource scheduling efficiency are validated. Results show that regional computing resource identifier resolution nodes possess low-latency resolution capability，while root computing resource identifier resolution nodes maintain a high resolution success rate. Compared with existing identifier resolution mechanisms，this computing resource identifier resolution mechanism can effectively support the resolution needs in scenarios of large-scale distributed resource convergence and interconnection for the internet of computing.
The rapid development of artificial intelligence （AI） large models has posed new challenges to high-performance network architectures. This paper systematically analyzes three core issues： the conflict between high-bandwidth demands and costs，bottlenecks in low-latency stability，and complexities in topology adaptation. It reveals the critical roles of enhancements in remote direct memory access （RDMA） technology，improvements in edge-side capabilities，and end-to-end network collaborative optimization. The study finds that there is an urgent need for industry-academia-research collaboration to address the challenges faced by high-performance network，and provides suggestions for development directions. Future trends point to collaborative innovation in protocol openness，hardware optoelectronic integration，and paradigm autonomy，which breaks the limitations of single-dimensional optimization through cross-layer design，thereby providing high-performance，low-energy-consumption interconnection support for ultra-large-scale AI clusters.
To address the problem that existing wide-area distributed computing power scheduling，a technology based on dynamic latency-aware scheduling （DLAS） is proposed. This technology detects backbone network latency through a 30 second cycle，combined with intelligent fault detection and adaptive migration strategies，to achieve better routing decisions and high availability guarantees. Establish a delay optimization mathematical model，which has been theoretically proven to reduce the average delay by 35% to 50% and shorten the fault recovery time to the minute level. Experiments have shown that the dynamic latency aware scheduling of DLAS reduces the response latency of computing power services by 42.3% compared to traditional polling scheduling，28.7% compared to static geographic scheduling，and improves service reliability to over 99.95%. This can provide theoretical and practical guidance for constructing efficient and dependable distributed computing power networks.
Amidst the rapid development of the Internet of computing，escalating computational costs during the post-training phase of large language models （LLMs） have become a critical bottleneck hindering widespread technology adoption. First，by systematically organizing and training cost optimization technology system，a comprehensive framework is constructed to reduce computational，storage，and data overheads，leveraging the cross-domain collaboration characteristics of the computing power internet. Second，the limitations of existing mainstream techniques are analyzed，and the evolution trends in this field are summarized to explore new directions for post-training cost optimization techniques of large models in distributed computing power interconnection environments.
The large-scale development of nternet of vehicles （IoV） is in urgent need of the support of computing-network convergence featuring low latency and high reliability. To address the problems including unclear mapping between hierarchical IoV services and hierarchical computing-network capabilities，lack of cross-domain collaboration mechanisms，and fragmented cloud-edge-end resource scheduling，this paper proposes a three-level collaborative architecture of “access-convergence-core” edge and central cloud，designs a peer-to-peer cloud collaboration scheme for cross-operator，cross-region and heterogeneous computing power，constructs a full-cycle closed-loop scheduling mechanism of “perception-decision-execution-audit-recovery”，and puts forward suggestions focusing on computing-network convergence standardization，cross-domain collaboration enhancement and intelligent scheduling optimization，so as to provide systematic support for the large-scale commercial application of IoV.
With the rapid iteration of Artificial Intelligence （AI） large models and the evolution of wide-area intelligent computing networks，the traditional single computing-optical synergy architecture can no longer meet the stringent requirements of large model training and inference for full-link resource coordination across the “optical-network-computing-application” continuum. Focusing on this core challenge，this paper conducts research on the key technologies and implementation approaches of optical-augmented computing，and designs an integrated optical-network-computing-application scheduling platform scheme based on Optical Circuit Switching （OCS）. A five-layer integrated architecture consisting of the adaptation layer，perception layer，control layer，service layer，and full-chain simulation layer is constructed，forming a closed-loop scheduling mechanism of “perception-decision-execution”. This scheme can effectively support the differentiated demands of diverse AI services and facilitate the construction of an efficient and reliable wide-area intelligent computing network ecosystem.
Generative artificial intelligence （AI） video technology is driving innovation in content creation，but also brings some challenges. Focusing on three typical tasks—Text-to-Video （T2V），Image-to-Video （I2V），and AI-powered video editing—this study analyzes their distinct risks and systematically sorts out general security technologies covering the entire lifecycle of pre-generation，in-generation，and post-generation. It discusses precise adaptation strategies for different tasks and the role of the internet of computing. By examining industry practices，it identifies key governance shortcomings，including the gap in security capabilities，fragmented identification standards，and ambiguous responsibility definition. This paper provides support for building a targeted AIGC video security system.
Computing power measurement and transaction management technologies serve as the core support for the market-oriented operation of computing power networks，directly influencing the allocation efficiency and value realization of computing resources. Based on the practical experience of China Telecom Cloud Xirang Platform，this paper analyzes the implementation paths of key technologies such as unified measurement of heterogeneous computing power，dynamic duration estimation，flexible pricing adaptation，and intelligent scheduling optimization. Measured data show that the unified measurement accuracy error of heterogeneous computing power on the platform is small，the utilization rate of computing power resources is significantly improved，and the user’s computing power usage cost is obviously reduced after cost scheduling optimization. The research results provide technical references for the standardization of computing power measurement and marketization of transactions.
In the era of large models，the demand for industrial computing power has witnessed a significant upsurge. However，its supply still faces three practical challenges： discovery，allocation，and utilization. This study examines the industrial paradigm shifts driven by the internet of computing power，integrating the methodologies of enterprise digital-intelligent transformation with the multidimensional maturity model of the German Academy of Science and Engineering. It constructs and validates a maturity assessment framework that encompasses four dimensions： network，resource，operation，and application. The results indicate that the internet of computing power accelerates the industry’s transition from resource procurement to task-based delivery，proving the feasibility of logical-layer resource integration. In the future，the internet of computing power will evolve toward intelligent perception and global interconnection，thereby providing a strategic reference for the intelligent upgrading of industries.