SD-RTN Millisecond Level Network Acceleration Technology Change

If the end-to-end latency of live video streaming must be tightly controlled within 500 milliseconds, so that it has the same low-latency experience as VideoLink, will the current CDN acceleration technology still work? First of all, the data cannot have cache, and the delay accumulation of TCP must be eliminated. Even RTMP protocol has to switch to the technical architecture of WebRTC. The real-time video application field of mobile Internet is rapidly evolving. Let us first call it the network delay revolution.

Operators are obviously sensitive to the above-mentioned network delay revolution and are making efforts to promote the construction of SDN/NFV: Layer 3 decoupling, network cloudization and development of ONAP have become the consensus of the industry. With the increase of the underlay network service capability, will it greatly weaken the dependence of underlay network acceleration technology? The answer is yes. However, the actual situation is that we currently cannot fully rely on the operator network to develop a completely free global real-time audio and video call applications. For example, if you want to carry out services like QQ micro-channel audio and video chat, you must also rely on the underlay network relay acceleration technology to solve NAT failures and "four-span" problems. The “four-span” mentioned here refers to the problem of fluctuations in the quality of network transmission across countries, operators, regions, and times.

The SD-RTN (Software Defined Real-time Network) discussed in this paper is a kind of above-mentioned underlay network transit acceleration technology. Compared to CDN, SD-RTN provides UDP-based real-time data transmission cloud services with millisecond end-to-end network delay. SD-RTN is a service architecture that can carry any point-to-point real-time data transmission needs: just call the open API, whether it is real-time video (conference, education, live, social, surveillance, VR), file transfer (short video, Office) or high-speed data synchronization (game, AI, IOT, Internet of Things) can easily access SD-RTN real-time data transmission cloud services.

The so-called converged SD-RTN refers to the SD-RTN, which is built on a shared network and server basis instead of renting POP nodes exclusively and purchasing hosting servers. Integration of SD-RTN is a perfectly elastic accelerated transit network, can solve the contradiction between resource investment and business development speed network node. After all, not every company's business is large enough to support the high costs of buying servers and renting BGP lines.

The converged SD-RTN utilizes shared node networking, so it is necessary to use network node resources flexibly and not to have a strong impact on other existing services. To this end we have introduced two technical ideas:

1. End-to-end dynamic routing algorithm that avoids automatic congestion avoidance;

2. Build a 99.9% stable network with 99% reliable nodes.

The end-to-end dynamic routing algorithm requires that the server's data relay module has real-time detection capabilities and instantaneous path switching capabilities for the available idle bandwidth of the network node, and this path switch must be service-aware. To this end, we have developed a comprehensive QoE mechanism that includes packet loss, real-time statistics of delay and jitter, and multipath hot backup. Built with 99% 99.9% reliable and stable network nodes, topology convergence means that SD-RTN must be updated in real time, highly self-healing. Therefore, we use the QoE mechanism and increase the speed measurement function of the entire network so that each network node has a real-time, dynamically updateable topology tree structure.

The integration of SD-RTN as a common point-to-point real-time data transmission cloud platform is embodied in the following technical characteristics: 1. Shared nodes; 2. Protocol optimization; 3. Close access; 4. Dynamic routing; 5. Cloud QoE; 6, common architecture. Share the nodes first, so you won't go into details. Secondly, protocol optimization refers to the simultaneous support of UDP/TCP protocol access. In the case of TCP access, SD-RTN can be internally converted to the QUIC protocol to reduce the delay accumulation. Near-access refers to using the global IP experience database and measured data to allocate the best transit access point. If the carrier's underlay network transmission quality is good enough, then avoid introducing multi-hop data transfer. Dynamic routing refers to the fact that the end-to-end transmission path can be switched in real time and is completely unaware of specific services. Cloud QoE not only provides a sound mechanism for the reliable transmission of data, but also forms the basis for networking of shared nodes. Drive QoE must be flexible and configurable, for packet loss, delay and jitter sensitivity of different traffic types, the right quality of the transmission path arrangement automatically re-calculated. The common architecture refers to the real-time data transmission cloud services that cover all the business scenarios using the APIs of server to server (S2S), client to server (C2S), and client to client (C2C).

The above-mentioned common architecture S2S access process mainly aims at real-time data synchronization between servers. Its characteristic is that both ends have a public network IP, and the API can be called to inform the SD-RTN access point to send data to the specified IP and port. The C2S access process is currently the most widely used application scenario. It is characterized in that one end is a server and the other end is a terminal that uses a private IP behind the NAT. After accessing SD-RTN, this terminal must use the public network IP after NAT mapping and send a data packet to the SD-RTN access point after creating the session. In this way, the complete TCP/UDP bidirectional data can be established. aisle. Finally, the C2C access process is directed to a direct connection scenario between terminals, such as a two-person audio and video chat in social networking. In the C2C access process, the API server in the SD-RTN actually assumes the role of the NAT puncture hole punch server, and the SD-RTN access point is the role of the TURN server.

Crimp Tool

Our Remove tool are used to insert and extract terminals from connector housings respectively without damaging the terminal or housing, a key criterion for quality wire processing, especially for OEM and aftermarket rework and repair important.
Whether standard or custom-made, our ejection tool kits are designed to perform precision tasks of the highest professional standards, yielding high-quality end products and bringing greater operational efficiency to your business.
The crimp toolare used to clamp the BNC connector jacket and base. It has a hexagonal notch for crimping. Generally, this kind of crimp tool also has the functions of stripping and cutting wires. Three-purpose network cable pliers, multi-functional, strong and durable, is a regular tool in the information age and modern families. Can make RJ45 network cable connector. RJ11 telephone line connector, 4P telephone line connector.
All the functions of the network cable pliers are integrated, and it is convenient to perform operations such as cutting, crimping, and stripping. It is made of iron, suitable for RJ45, RJ11 network cables, hardened and blackened, light and strong.

contact remove tool, contact crimp tool, remove tool

Kunshan SVL Electric Co.,Ltd , https://www.svlelectric.com

This entry was posted in on