Cloud video conferencing: In the era of mobile Internet, with cloud computing technology, virtual machines are used to replace traditional expensive hardware resources. Relying on a new generation of audio and video compression technology, a new generation of algorithms, and with the help of cloud computing, high quality, no lag, and no loss of sound can be ensured even when network packets are lost on the Internet.
Nowadays, the cloud video conferencing market is complex, and only a few truly achieve cloud video conferencing. The video system should have the following characteristics at the same time to be called a true cloud video platform:
Super High-efficiency video processing
High-quality audio and video processing requires a large amount of computing resources. To meet the ultra-large-scale access requirements of the cloud video platform, the efficiency problem of a single media node of the platform must first be solved. An MCU with a traditional video editing and decoding mechanism requires more than one CPU core to process each channel of 1080P video. The extremely high hardware resource requirements make it difficult to support large-scale video access. A true cloud video platform must adopt more advanced SVC flexible coding and multi-stream intelligent routing mechanisms. The 1-core CPU with the same main frequency can support more than 20 channels of high-definition video, improving the efficiency of a single MCU media node without reducing audio and video quality. It is improved by more than 10 times, and multiple media nodes are integrated into a unified management resource pool through distributed computing architecture to support ultra-large-scale video access.
Virtualization deployment
Virtualization technology decouples computing resources from physical hardware, which is the basis for providing segmented and quantitative services to the outside world and a necessary condition for cloud computing. Therefore, virtualization deployment is also one of the necessary conditions for cloud video platforms. Even traditional video giants such as Cisco (launched in 2015), Polycom (launched in 2016), and Huawei (launched at the end of 2017) have also launched virtualization platforms, proving Traditional hardware MCUs have completely fallen behind.
High reliability cluster
Ultra-large scale inevitably requires higher reliability. Traditional video emphasizes the stability of a single machine and is suitable for self-built small systems. However, there will be many single points of risk through the MCU cascade architecture during large-scale expansion. What the cloud video platform needs is the reliability of the overall service, separation of signaling media, each module can be horizontally expanded, creating high-reliability cluster resources, and supporting multi-active and hot backup in remote locations.
Based on general network
Traditional video coding H.264AVC is based on dedicated line design (low latency, stable bandwidth, extremely low packet loss), which is difficult to adapt to the application requirements of the general Internet and greatly limits range of application scenarios. The cloud video platform uses SVC flexible coding and multi-stream intelligent routing mechanisms to provide high-quality audio and video services based on universal networks, adapting to access to various network conditions such as dedicated lines, the Internet, 3G/4G, and satellites.
High scalability
The term "capacity expansion and transformation" does not exist in cloud computing. The overall architecture can be elastically expanded, and there is no need to change the deployment as the capacity changes, just like electricity, Same as tap water. When the overall basic hardware resources are sufficient, they can be called on demand, dynamically migrated, automatically expanded, and delivered quickly.
SAAS-level on-demand service
Cloud video service belongs to the SAAS layer (software application as a service) in cloud services. To meet this standard, in addition to the underlying architecture, an operator is also required. Level management backend meets three major requirements:
1) Users can use services independently and provide exclusive enterprise backend.
2) Highly automated, resources can be quickly delivered according to user needs (such as capacity expansion requirements).
3) Data operation provides users with various data statistics to help users better complete summary reports, capacity planning and other tasks.
To sum up, in order to meet the six characteristics of cloud video, four technical standards must be met at the same time: first, the SVC flexible coding algorithm must be used, second, virtualization deployment, and third It must be a top-level flat design, a distributed cluster architecture, and fourth, it must have a SAAS-level operation platform. Only when all the above requirements are met at the same time can it be called "cloud video".
According to the above standards, giants such as Cisco and Polycom in the video conferencing industry do not have flexible coding algorithms and SAAS-level operating platforms, so they mostly enter the Chinese market in the form of hardware equipment, so they can only be regarded as traditional video conferencing manufacturers. .
At present, there are only three representative manufacturers in the world that can provide real "cloud video" services: ZOOM, Vidyo and Xiaoyu Yilian. ZOOM and Vidyo are both American manufacturers. They mainly rely on agents and agent brands to operate in China. With the Sino-US trade Further expansion of the war, rising costs and availability are obvious issues. Xiaoyu Yilian is the only cloud video service provider in China with completely independent intellectual property rights, and has more than 30 core audio and video patents.
Compared with ZOOM and Vidyo, Xiaoyu Yilian’s biggest advantage is that it not only has a global deployment cloud platform that rivals them in performance, but also has a complete intelligent hardware R&D team with full-scenario intelligent hardware terminals. The product line can provide users with a complete "cloud + terminal" overall solution. Solutions from manufacturers such as ZOOM, Vidyo and all agents are flooded with a large number of third-party products, which are very difficult to match and debug. In the case of large-scale deployment, it is difficult to guarantee users' video conferencing needs.