Viewpoint
Trends and Challenges
With the emerging 5G, IoT, and VR coupled with the increasing requirements for service optimization and low latency, computing is required to be carried out at the edge of networks. Typically, the computing functions are deployed on the cloud infrastructure in a Data Center (DC) which is at a higher level and far from the end users. For some latency-sensitive or experience-sensitive services like Big Video, the delay in such a long path leads to a bad user experience. Since the Access Office(AO) is the closest to end users, it is reasonable to place the cloud infrastructure to avoid the complex transmission or data processing between AO and DC, hence helping enhance user experience.
However, deploying a cloud infrastructure in the AO encounters a lot of challenges.
- Some edge access offices have very limited spaces. Some of them are rented and extra space will cause an increase in rent and OPEX.
- The conditions of some equipment rooms are simple and unsatisfactory, which makes equipment deployment difficult.
- The cloud infrastructure at the edge of a network can be deployed at geographically scattered sites, which makes management difficult.
ZTE “Light Cloud” Solution
ZTE ”Light Cloud” solution is targeted at experience-sensitive services, decoupling software and hardware, and unified management. It can be used by the operators themselves or leased to third parties as a high-quality and low-latency network infrastructure, meeting the requirements of uRLLC and Big Video services in a wide range of application scenarios, and reducing costs.
ZTE "Light Cloud” solution has the following three types of applications.
- Large-traffic and low-latency accelerated video services including Access CDN, video cache, and TCP acceleration.
- Be targeted at experience-sensitive services such as autonomous driving, Internet- of-vehicles, industrial control, remote health care, and VR, or provides eMBB and uRLLC services as a MEC or vSTB (virtual STB) .
- The “Light Cloud” devices and pipelines act as high-quality and low-latency IAAS/ PAAS or are bundled with telecommunications pipelines to be leased to third parties.
Considering the AOs’ conditions, service requirements, and the development of IT technology, especially the soaring development of low-power SoC CPU and SSD. ZTE "Light Cloud” solution provides built-in blade that can be inserted in the slots of the existing OLT, enables on-demand introduction of computing and storage resources without the need of extra room spaces, and supports the access network NFV.
The blade is 300mm in depth and is plug-gable in the OLT subrack, effectively utilizes the existing resources of the OLT without extra footprint in equipment rooms comparing to commonly-used universal servers. The power consumption is less than 200W, 50% lower than that of the standalone blade servers. The built-in blade connects the PON OLT backplane via a high-speed bus and makes use of existing racks, shelves, as well as uplink, downlink, and fiber resources, which reduces external connections and interfaces, and provides higher reliability. Meanwhile, it leverages state-of-the-art technologies such as lightweight SoC CPU and SSD to provide better forwarding performance, stronger I/O storage capability, and broader temperature adaptability.
The blade is based on ZTE’s self-developed open virtualized operating system to provide open services and capabilities. In the downstream direction, it shields the hardware-based differentiation and services, and reduces costs. In the upstream direction, it supports self-developed or third-party APPs. It provides multi-dimensional performance acceleration, enables rapid provisioning of edge services, supports more third-party APPs, and enriches user experience. The blade is managed by the PIM (Physical Infrastructure Manager) system, which is included in the overall management system in compliance with ETSI NFVI via the PIM to achieve automatic management and configuration.
Application Scenarios
ZTE "Light Cloud” solution provides overall openness, broad adaptability, excellent performance, and rapid market response capabilities. It supports localized deployment of cloud applications, rapidly processes edge data services, provides new value chain and new segmented markets for ICT networks, and spurs flexible deployment of network applications.
Access CDN
With the booming development of video services, the proportion of video traffic on broadband networks is increasing. The growth of video traffic consumes large amount of backbone network resources, and brings a heavy burden to the operators’ networks. At the same time, industry analysis and statistics show that hot spot contents, in particular, may form a huge amount of repetitive traffic from the video source to the users in a short time. These repetitive traffic carries the same content and their paths from the content source to the network edge are identical. By making use of this characteristics, the CDN edge nodes are getting closer to users.
Access CDN is a kind of CDN form in which the CDN node is deployed downward to the access network. Figure 2 shows the deployment of Access CDN in OLT’s built-in blade. The Access CDN mainly stores the TSTV programs, TVoD programs, and hot spot movies on-demand.
On the one hand, moving the CDN downward to the fixed network will make use of the high-bandwidth and excellent QoS features of the PON networks, which avoid congestion and packet loss. On the other hand, Access CDN will provide services near users which dramatically reduces the network transmission latency, saves the bandwidth of the uplink network and relevant network devices, sharply eliminates the load of the CDN central nodes, and effectively drives down the bearer cost of the video networks.
A single built-in blade in an OLT with a depth of 300mm can store up to 15T contents and process 30Gbps contents, and consume less than 200W of power, all of which provide a strong technical foundation for the CDN moving downward, and to be embedded in the OLT. Besides, it is predictable that with the IT technology development, the storage and computing capabilities will be stronger, and the cost will be lower, which will make the performance-price ratio of Access CDN increasingly higher.
Cloud Based vSTB
With the users’ growing demand for video experience, the provisioning of value-added services attached to the STB put forward higher requirements on the software and hardware of the STB. The replacement of STBs is not only expensive but also has many engineering problems. Meanwhile, due to the diverse STB models, the hardware capabilities vary greatly, resulting in a series of problems such as difficult adaptation and slow provisioning of new services.
ZTE’s vSTB solution emerges to address this situation. The vSTB operates on the built-in blades, processes services and UI, converts the interface into a video stream and distributes to the pSTB. The pSTB only needs to reserve the video coding/ decoding, and I/O interface capabilities, and does not need to provide strong CPU, memory, and Flash capabilities. The vSTB solution can effectively shield the differences between the STBs from different vendors, facilitate unified distribution and deployment of new services, and dramatically reduce TCO.
Figure 3. Light Cloud for vSTB Deployment
From the perspective of latency reduction and traffic offloading, deploying vSTB in OLT’s built-in blade has tremendous advantages. Through downward deployment of the computing to offload the bandwidth, the vSTB solution can be deployed on-demand and conveniently without the need of considerably upgrading or transforming the aggregation core network.
MEC
With the requirements for network service optimization, downward deployment of contents, and low latency become more and more urgent. To ensure service experience quality and to reduce the transmission cost, these services need to be provided at the network edge, i.e. MEC application. At the same time, integrated AO has become the deployment mode adopted by a majority of mainstream operators. Deploying the BBUs, DUs and OLTs in the AO requires that the AO provides the NFVI needed by edge computing. ZTE’s “Light Cloud” solution provides an economic-friendly and efficient mode for the NFVI construction in the AO. The MEC VNF and APP can be deployed in AO NFVI as per demand to form a dynamic and highly-efficient edge computing system.
vRG(virtual Residential Gateway)
Some control, management and service functions of HGW deployed in existing networks are moved upwards to the “Light Cloud” to simply the functions and reduce procurement and maintenance costs.
vFirewall
Encrypting the upstream and downstream data in the “Light Cloud” including management data and service data makes the networks more secure and reliable, avoids the risk of data theft and loss, and ensures smooth service provisioning.
AI processing
Supports the operating of the APPs for network quality detection and traffic detection, implements smart network O&M, and helps targeted marketing and advertising.
Vertical Industries
Helps the development of vertical industries including IoT-based applications (energy saving management, smart fence, etc.), QCell indoor positioning and navigation, information service management, etc. The deployment sites can cover a variety of indoor and outdoor scenarios including campus, enterprise zones, large shopping malls, airports, conference and exhibition centers, indoor parking lots, sports centers, etc.
According to the actual service development requirements, “Light Cloud” solution based OLT is not only limited the previous applications, but also supports more flexible scenarios for service acceleration, cache and fixed network VNFs at a low cost and low power according to the characteristics of different services.
Summary
Aiming at the development trend of FMC in the future, ZTE launched the innovative "Light Cloud” solution with built-in blade by making use of the large amount of idle OLT slots to open up a broad imagination for multi-access computing in the future. ZTE will continue to do research in the lightweight edge computing arena, leverage cutting-edge technologies to implement low-cost, low-power edge computing infrastructures in support of flexible combination as per service, and lay a solid foundation for the operators’ network build out and service development.
