Shulou(Shulou.com)05/31 Report--

What this article shares with you is about how to build a NFVI that can support the network transformation. The editor thinks it is very practical, so I share it with you to learn. I hope you can get something after reading this article.

NFV becomes the technological basis of operators' Network Transformation

In the past few years, the network transformation of operators has become an unavoidable core topic in the industry. On the one hand, the growth pressure of existing business and the change of market competition, as well as the acceleration of 5G process put forward new requirements for network infrastructure construction, which makes the demand for transformation increasingly urgent; on the other hand, the development of emerging technologies such as virtualization, cloud computing and SDN also provides a realistic possibility for the realization of transformation.

Since it was first proposed, in just a few years, with the active participation of many operators and manufacturers, NFV has rapidly entered the stage of implementation from the concept on paper, and is widely regarded as the capital construction mode of the future 5G network and other emerging services. Through the hierarchical decoupling of network infrastructure (NFVI + VIM, hereinafter referred to as NFVI), upper network service (VNF) and management allocation system (MANO), NFV draws an open, flexible and automated evolution blueprint for future operators' networks, and theoretically provides a basic basis for operators' network transformation and service innovation.

Some key demands on NFVI put forward by the Commercial of NFV

It should be noted that while NFV shows a bright future for network transformation, it does introduce some technical challenges that can not be ignored. Specific to the field of NFVI, through the analysis of the practice cases of operators at home and abroad in recent years, we can sum up the following issues that need to be paid attention to.

First of all, whether the NFVI implementation scheme based on IT technology system can carry telecom-level network services.

For a long time, telecom operators have been providing all kinds of network services with excellent experience, security and reliability with their advanced telecommunications technology system, thus supporting the smooth operation of the whole social economy and shouldering the vital responsibility of information infrastructure.

After entering the NFV era, the network business migrates to NFVI, which is composed of COTS hardware and open source software. Can its performance, reliability and security still meet the existing business requirements? Will the risks introduced by technological innovation affect the user experience? The answers to these questions have undoubtedly become a prerequisite for the commercial use of NFV.

Secondly, whether the NFVI architecture scheme can support the continuous evolution and innovation of network services.

It is not difficult to imagine that under the dual drive of market competition and technological change, the transformation and innovation of operators' networks will become a long-term and continuous process, and may even become a new normal faced by operators. In this context, all kinds of technical requirements for NFVI will continue to appear. In other words, the evolution of business will also drive the continuous evolution of NFVI itself.

Therefore, the planners and designers of NFVI need to face the following question: what kind of NFVI architecture solution can achieve a flexible evolution that continuously matches business requirements? For example, can the NFVI built for 4G services evolve smoothly to meet the requirements of 5G service deployment? These problems are directly related to whether NFVI can really support the landing of the strategic goal of network transformation, but also related to whether to avoid technical risks and investment waste caused by planning factors.

Third, how to correctly deal with the contradiction between the openness of NFVI and the difficulty of integration delivery.

The openness of technical architecture and system architecture is very important for NFVI. Only an open NFVI can effectively promote the diversification of network business and manufacturer ecology, so as to make the benefits of network transformation.

At the same time, it should also be noted that because the technical specifications related to NFV are not perfect, whether it is the docking of NFVI northbound with VNF and MANO, or the integration of NFVI internal hardware or virtualization software with COTS, there are still some areas that lack standards. This situation can easily lead to various risks in the integrated delivery of multi-vendor product solutions, thus hindering the commercial landing of NFV.

The focus of operators when building NFVI

In view of the above problems, operators need to pay special attention to the following technical aspects when planning, designing and building NFVI.

First, reasonable reinforcement and enhancement of NFVI to make it truly meet the quality requirements of carrier-grade services.

NFVI, which can carry carrier-grade services, needs to be comprehensively and deeply strengthened and enhanced in many directions, such as high performance, high reliability, high security and so on, so that it can meet the technical requirements of telecom-level.

In terms of high performance, the virtual network forwarding technology with high throughput and low latency is very important. On the one hand, the high-performance virtual switch is related to whether the forwarding capability of NFVI meets the commercial requirements of vEPC, vBRAS and other existing services, on the other hand, it is also the key to affect the commercial experience of 5G, video and other new types of services on the NFVI platform. At the same time, the improvement of performance also means a reduction in cost under the same business scale, which has a clear business value.

As Moore's Law is gradually coming to an end, the idea of relying on the performance improvement of general-purpose processors to continuously and significantly improve forwarding performance is no longer sustainable. At the same time, SR-IoV-based hardware virtualization technology provides another alternative to solve the problem. In the long run, different network forwarding technologies based on software virtualization and hardware virtualization will coexist in the field of NFV for a long time and play an important role in different business scenarios.

As far as high reliability is concerned, the key is to carry out panoramic, multi-level and multi-domain detailed analysis and reinforcement of the reliability of the system to ensure that there is no single point of failure in the system, and to ensure that the upper-level business can achieve 5-9 telecom-level high availability in various failure scenarios.

In order to achieve this purpose, it is necessary to reasonably strengthen the internal components of the data plane, control plane, management plane and operation and maintenance plane of the NFVI scheme, and comprehensively use a variety of technical means, such as second-level fault alarm, automatic fault recovery of control components, automatic fault recovery of virtual machines, automatic backup of business data, data center-level local / remote disaster recovery, and so on, to fully ensure the reliability of the system.

Security is a new topic that has attracted more and more attention in the field of NFV. From the closed and integrated traditional telecommunications equipment to the open and hierarchical decoupling NFV system, the security of business really needs more attention. In addition to the tenant-level security isolation capabilities supported by computing, network, storage and other virtualization technologies, it is also necessary to examine and strengthen the security of NFVI from the perspectives of physical infrastructure security, system software security, business data security, system security management and so on.

In addition, ease of operation and maintenance is also a key problem faced by NFVI in the commercial process. For NFVI constructed by IT technology system, its operation and maintenance work naturally includes all aspects of IT system operation and maintenance. Installation and deployment, upgrade patches, change and expansion, monitoring and alarm, and fault management are all practical tasks that can not be ignored by the operation and maintenance team. This also puts forward high technical requirements for the operation and maintenance system of NFVI. For example, when the Linux operating system on the physical host of the NFVI computing node needs to be upgraded, how to achieve full cluster automatic upgrade without business interruption is a difficult challenge. If these problems are not solved well, it will directly affect the negotiability of NFVI. To this end, it is necessary to introduce a powerful automated upgrade process management system into the life cycle management system of the NFVI solution, so that the sequence relationship between business migration and node upgrade operations can be coordinated in a programmed and automated way, so as to accelerate the upgrade process and reduce the upgrade risk, while reducing the upgrade pressure on the operation and maintenance team.

To sum up, the four aspects of performance, reliability, security and maintainability all need to be paid great attention to in the process of designing NFVI.

Second, in order to effectively support the continuous evolution of operators' services, NFV is bound to move towards NFC (Network Function Cloudification), so the design of NFVI needs full flexibility.

It can be expected that the future network of operators will change from the networking construction mode based on telecom room to the networking construction mode based on distributed multi-level and multi-data centers. Data centers of different sizes and levels are distributed in different geographical locations, and provide different types of cloud resource pools and different levels of network access latency to support distributed deployment and wide regional coverage of different types of business. On this basis, due to the demand of supporting millisecond ultra-low latency services in the future 5G network, NFVI needs to consider supporting a large number of small-scale edge sites at the edge of the network. Therefore, supporting the networking and centralized management of the above-mentioned multi-level, multi-center and edge sites has become the key capability of NFVI.

At the same time, diversified business will also put forward diversified resource types for NFVI. Therefore, in the architecture of NFVI scheme, we need to fully consider the possibility of heterogeneous access to different types of computing, storage, network and other resources. For example, in the computing field, we not only need to provide virtual machines as basic computing resources, but also need to consider providing bare metal resources directly for some special high-performance services; we not only need to be able to support general-purpose x86 processors, but also need to consider providing high-performance heterogeneous computing resources such as FPGA for some computing-intensive businesses.

On this basis, NFVI should also have the ability to integrate with the upper container cluster management platform to support containerized services, such as agile 5G core network services. However, NFVI itself should remain decoupled from the container management platform and even the PaaS layer, so as to avoid overly bloated NFVI architecture, affecting its own evolution and deployment flexibility.

Third, we need to consider the openness of NFVI vendors, jointly promote the ecological formation of the NFV field, and consider the ability of manufacturers to deliver end-to-end solutions.

In the face of the contradiction between system openness and integration delivery risk, operators can consider working closely with NFVI industry vendors to jointly promote commercial docking verification between VNF, MANO and other components of vendors and mainstream NFVI solutions in mainstream commercial scenarios, and form relevant technical specifications. At the same time, consider the delivery capability of the vendor's end-to-end solution to ensure fast delivery and follow-up services.

The above is how to build a NFVI that can support the network transformation. The editor believes that there are some knowledge points that we may see or use in our daily work. I hope you can learn more from this article. For more details, please follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.