Since the introduction of mobile phones in the 1980s, wireless technology has undergone major changes over a period of about 10 years. The main feature of the first generation wireless technology was analog voice, and the second generation was 2G, the digitization of voice. The third generation includes data access and the fourth generation features ultra-high-speed data and end-to-end IP networking technology (now available in highly industrialized countries).
These network upgrades were needed to meet bandwidth demands from growing consumer users and to enable new services to be provided over networks. As all service providers know, bandwidth demand continues to grow at an average rate of 45% per year, and this demand will be further fueled by the billions of object Internet devices that connect networks over the next few years.
Service providers are also looking for ways to earn revenue and compete with key competitors in this area. These are some of the key factors affecting the next generation of mobile standards known as IMT-2020, or more commonly known as 5G networks.
Unlike previous evolutions, 5G does not completely replace 4G, but it works like 4G in the early stages of implementation and complements 4G. The first 5G standards, known as non-standalone mode (NSA), are expected to be completed by the end of May 2018, and 5G is expected to be 100 times faster per cell area, 10 times slower, Density.
All three of these features will enable countless new services, applications and business models when combined. When combined with all newly connected Internet devices, the model will have more than 1,000 times more data flowing through the 5G network, both user area data and control area data.
A 5G network is possible when there is ongoing network evolution as follows.
Implementation of Software Defined Networking (SDN) in which the data area of the application area, control area, telemetry area, and packet transmission scheme is completely decomposed into different entities.
Evolution of Network Architecture from Existing Network Architecture to Virtualized Networking Architecture with Network Functional Virtualization (NFV)
The desire for open networking application programming interfaces (APIs) and the use of data models and the desire to discard existing protocols and managed objects
The need for more intelligent and more automated network operation and network control
Introducing multi-access edge computing (MEC) and deterministic transport technology to solve the low latency problem when connecting contents and users
In all network evolution, especially in the case of redevelopment network scenarios where existing sets of networking equipment are operating, it is almost impossible to realize the next generation network through a strategy of full replacement. This applies when operators begin to introduce 5G mobile networks into existing LTE or advanced LTE networks.
The three flows of 5G network evolution are expected to be as follows.
Flow 1 (currently underway): IT Convergence and Telecommunications Company (Telco) Focus on Telcos Making Investments in Telecom for Cloud
Flow 2 (after 2018): Focus on the evolution of 5G approach to investments in networking infrastructures for large bandwidth, high-capacity, and large volumes
Flow 3 (since 2020): Focus on distributed intelligence in networks where artificial intelligence, machine learning, and advanced autonomous networking developments enable what is impossible in networks and applications
Especially with the adoption of MEC platform technology, I think that both wireless and wireline will be mainly virtualized and distributed in different telco cloud network sectors in the target architecture view (see Figure 1 below). From an operator's point of view, I think there will be three networking cloud segments, including an access cloud, a local cloud, and a central cloud. Moving from the central cloud to the access cloud means that handling lower latency application traffic and different MEC platform conventions will be realized in both the access cloud and the local cloud, respectively.
Some key features of this target architecture are:
Infrastructure & Networking Fabric: Packet forwarding protocol agreement will be realized here. As the path intelligence moves into the software data domain, the networking fabric will be simplified and utilize a common set of protocols for WAN, inter-data (data and data), center communications, and intra (data) data center communications. Apart from protocol conformance and simplicity, this network area, also known as 'network based', is an advanced packet networking technology that can maintain a very low packet error loss rate and provide a deterministic movement guaranteeing low latency delivery in a consistent manner .
Distributed Data Plane: In this network plane, we focus on network programability and also focus on the common data plane stack realization across multiple different cloud segments. Network protocols and path intelligence can be dynamically programmed here and can be resiliently changed throughout the life cycle operation of network functions. Another important aspect of this lower layer is the open networking data model that allows for maximum code reuse despite variations in base program network I / O technology in the infrastructure / networking fabric plane and the joint programability API implementation.
Integrated Control Plane: The control plane and user domain decomposition began with the delivery system and are introduced separately into 4G Evolved Packet Core (EPC) in 3GPP Rel-14. With 3GPP Rel-15, both the 5G radio access network (RAN) and the 5G next generation core (NGC) are also defined with this decomposition structure. Control software innovation to realize an integrated control plane framework in which higher-level autonomous end-to-end networking control can be executed by having real-time processing and non-real-time processing elements in it when all network function control plane elements are disassembled Let it start. Such autonomous network control includes the following. 2) dynamic movement of network functions between cloud segments based on network load; and 3) cloud-based mobile packet core resources based on user patterns. Total autonomic optimization of cloud-based RAN resources with
Intelligence, Automation, Planning, Management & Revenue Creation: A wide variety of software modules will be realized in the network plane, where it becomes possible for operators to help their operators differentiate themselves in competition. These software modules will use technologies such as artificial intelligence, machine learning, enhanced autonomic networking technologies, and flexible policy engine implementation. By combining some of these software modules together, operators can enable what was impossible in their own network era.
Deterministic Networking in 5G Packet Transmission Networks
As discussed briefly above, an enhanced packet transmission technique will be introduced into the 5G network transmission scheme to enable the following at the same time.
- Provides reliable packet transmission and low latency with a low rate of packet error loss in a consistent manner.
- Provides statistical multiplexing and grooming, and oversubscription of low priority traffic.
- It enables to support both hard slicing and soft slicing in transport network slicing.
Advanced Ethernet technology known as IEEE 802.1 Time Sensitive Networking (TSN) can be utilized in the network to realize the three networking requirements outlined above. A summary of the component specifications of the 802.1 TSN is given below.
Each TSN component is capable of delivering deterministic backhaul for 5G (and LTE-A Pro) packetized front-end (Fronthaul) networks as well as SLA-guaranteed network slicing in mobile wholesale networks and for highly reliable low- backhaul. & lt; / RTI & gt;
The TSN component criterion is primarily a major enhancement to the IEEE 802.1Q functionality, and some components of the TSN feature have been incorporated into the IEEE 802.1Q-2014 specification and additional fixes were included in the year 2015/2016.
Also. The 802.1 specification has been improved (eg, an IEEE project on ongoing frame replication and removal) and there are specific improvements to the 802.3 specification (eg, Frame Preemption 802.3br with 802.3 improvements).
TSN component criteria fall into four main categories:
- Provides timing synchronization within the deterministic Ethernet network.
- Guarantee bounded low-latency Ethernet payload delivery.
- Provides a robust reliability transport mechanism within the Ethernet network.
- Management and control through open networking APIs and data models.