دانلود کتاب انقلابی در مدیریت شبکه – سفر به سوی استقلال بومی هوشمند، چاپ اول

5G and 6G networks are expected to provide virtually-unlimited-gigabit and ultrareliable connections to people and objects, when and where it matters, supporting diverse use cases with an extremely demanding range of requirements (in terms of latency, throughput, reliability, coverage, and security, cost targets, etc.). However, building a network that supports a diverse set of new services, all of which can be set up, dynamically reconfigured, scaled, and torn down at a moment’s notice, introduces several challenges. The network’s performance, coverage, and capacity should be constantly assured to satisfy the requirements of the active services. Any fault or degradation that might adversely impact the services should be resolved. Furthermore, unprecedented operational agility is required to allow services to be rapidly deployed, dynamically adapted, and continuously and seamlessly assured — similar to the way in which cloud providers offer ondemand, managed cloud services. Each and every service spans multiple technological domains and is composed of cloud resources, connectivity, virtual and physical network functions, augmenting services, and application logic. The network includes hundreds of thousands of network functions and each function has multiple versions. In addition, 5G and 6G run on a virtualized infrastructure and are designed with a fully programmable approach to software and microservices, capable of supporting diverse use cases. This makes the network exponentially more complex to operate and manage. When we add distributed clouds, heterogenous networks, the 10X densification of RAN sites, meshing, and integrated multisupplier technologies to the mix, we introduce even more complexity, driving up the timescale and cost. The accelerated worldwide deployment of 5G networks drives a radical change in the way networks and services are created, orchestrated, and managed. Evolving toward zero-touch and full end-to-end network and service automation have become an urgent necessity to manage this complexity and deliver services with agility and speed while adapting, assuring, and ensuring the economic sustainability of the highly diverse service portfolio. The ultimate target is to achieve the highest degree of automation while enabling fully autonomous operations driven by highlevel business goals and policies. The autonomous networks will be able to self-manage and self-organize (configuration, healing, assurance, optimization, etc.) without human intervention beyond the initial transition of the business intents. Realizing this vision requires a novel end-to-end architecture framework and enablers designed for self-management, near-real-time closed-loop automation, and optimized for data-driven analytics. Advanced machine learning (ML) algorithms and artificial intelligence (AI) will be essential tools to deal with the complexity and the diverse services. End-to-end automation is a challenging objective and represents the industry’s collective target for the technology evolution in years to come. The key components for increasing automation include the expression, as well as the translation, of high-level abstract business goals (e.g., intents) into efficient and impactful actions, deployment and interaction with closed loops, real-time data collection, extraction of meaningful insights from data, analytics and intelligence, and orchestration and resource control. The practical implementation of an intent-based network requires substantial automation technology embedded in the network. The translation of the intent into efficient and impactful actions in each context requires self-evaluation and self-measurement capabilities. Many of these aspects still require further research work, learning, and sharing of best practices and guidelines in applying them in operational service providers’ environments. Autonomous networks rely on analytics and intelligence to learn, reason, produce causality analysis, predict, etc. The networks should be able to execute even in conditions of uncertainty and lack of knowledge or in cases of aleatory (randomness/variability). Analytics can learn from collected information, identify patterns, and generate models for various optimization and classification tasks. AI technologies may be used to empower the automation processes. While AI technologies are developing and expanding fast, some gaps must be overcome to fully leverage their potential. AI and ML in service providers’ networking are still in the early days. We only start to see how AI/ML can be usefully applied. Currently, AIbased solutions are driven by use cases and work in isolation, with limited systematic reuse. The view and use of the AI potential are limited, mainly with ML and deep learning at the descriptive and explanatory levels. AI is still resource-hungry and works on big dump data. The integration of AI in network and service automation still requires significant human involvement in all steps of the AI development and operation processes. Automation is not only about technology. It requires changes in the processes and the mindset of people. Trust is a key barrier for adoption and building it will be a continuous learning process; as more automation and AI-empowered closed control loops are deployed and start to operate safely/efficiently, human trust will increase and the requirement for a level of supervision/visibility will diminish. The purpose is to apply transparent, reliable, and robust automation. To increase the comfort level of the human in automation, a level of supervision and transparency is needed. The ability to explain and trace the machine’s decision-making process and the reasoning behind the decisions are key in supporting it. The use of AI in network and service automation will evolve incrementally to the ultimate goals of realizing zero-touch intuitive AI and autonomous operations, leveraging machine reasoning and symbiotic human—AI interaction. This book explores the megatrends and wireless evolution jointly influencing the development of network and management technologies. Chapters 1 and 2 will provide an overview of the technology, industrial, and standardization landscape to understand the need for automation, their technical means, and the main standard definition organizations producing relevant studies, reports, and normative specifications. Next, each of Chapters 3-13 provides a deeper view into one of the key technology enablers of network management automation and network autonomy, starting from fundamental and established ones, and building toward increasingly specialized and advanced topics. Chapter 14 provides a summary of the book and an outlook on the technology trends shaping the future of telecommunication systems.