Alibaba Cloud Zhang Yiping: Build an edge cloud full-site acceleration network system

Introduction to On June 9, 2021, the Asia-Pacific Content Distribution Conference and CDN Summit will be held in Beijing. Zhang Yiping, senior technical expert of Alibaba Cloud Edge Cloud Network, was invited to the conference and shared Alibaba Cloud’s efforts in building an edge cloud full-site acceleration network system In practice, the edge cloud node supports the hierarchical transmission acceleration of various application protocols, and provides wide-coverage network acceleration basic services for differentiated enterprise applications.

On June 9, 2021, the Asia-Pacific Content Distribution Conference and CDN Summit was held in Beijing. Zhang Yiping, senior technical expert of Alibaba Cloud Edge Cloud Network, was invited to participate in the conference and shared Alibaba Cloud’s practice of building an edge cloud full-site acceleration network system based on the edge The cloud node supports the layered transmission acceleration of various application protocols, and provides basic network acceleration services with wide coverage for differentiated enterprise applications.

Zhang Yiping believes that speed is a business indicator that has received widespread attention in Internet business, and speed generates value. In the scene of video and interactive interaction, the speed corresponding to the video second opening rate, playback freeze rate, video clarity, and interaction delay directly affect the user experience and retention; in the financial and e-commerce transaction scenarios, the response speed of each transaction request And the success rate is related to the operation of high-value business; in the scenario of full site business docking, one-stop accelerated docking can greatly improve the efficiency of business docking is the key.

From content acceleration to full site acceleration

Site-wide acceleration is an upgrade from content acceleration. Content acceleration is mainly the acceleration of static content, such as video on demand and file download. CDN improves user request efficiency and bandwidth carrying capacity through a large number of nodes, and realizes acceleration based on cache. In addition to services that can be cached for a long time, such as video-on-demand, file downloads, and static web pages, a large number of real-time services appear, and need to rely more on real network acceleration rather than experience acceleration brought by nearby static caches, such as online games, financial transactions , Real-time audio and video communications, etc.

Unified routing and scheduling based on edge cloud nodes is the key

The essence of routing and scheduling is to achieve the best overall performance/cost in differentiated customer services/scenarios/resources. Resources are nodes and networks. A single node includes CPU computing power, memory capacity, storage IO, network IO, etc.; services include long videos, short videos, downloads, live broadcasts, financial transactions, edge rendering, etc.; scenarios include cached content (consumption of storage Capacity/IO, network IO), dynamic content (consuming network IO, CPU computing power), traffic calculation (consuming CPU computing power, memory capacity);
Route selection and scheduling can be understood as a real-time dynamic solution process. Under the premise of ensuring service quality, each type of resource can be consumed as evenly as possible, especially the effective use of edge cloud node CPU computing resources.

Intelligent protocol stack is the basis for improving service quality

Combining business, scenario, and environmental differences to match a more suitable transmission protocol stack can reduce the service differences of different resources. In the scenario of protocol stack optimization, it is divided into end links, intermediate links, and back-to-source links.

1) End link

For the terminal, make full use of the data difference of the terminal environment to drive and match the most suitable transmission algorithm. The differences in the network environment, service type, time and other aspects of terminal access reflect different service qualities. Transmission protocol matching is the most effective means of optimizing quality, while also optimizing network costs.

2) Intermediate link and back-to-source link

Large-scale link convergence, the pursuit of optimal transmission quality and stability, allows to sacrifice part of the cost, combat fluctuations and interference of public network transmission, and achieve link disaster tolerance.

Full link diagnosis is the cornerstone for fast fault location and fast recovery

Quickly detect and locate abnormal points from the business perspective, improve the speed of abnormal recovery, and improve the efficiency of operation and maintenance. It can be divided into two types of detection:

1) Passive detection

Including TCP marking data statistics, cutting in from the seven-layer business, marking tracking based on the business layer Request/Response dimension, performing packet coloring according to the seven-layer business marking when the seventh layer is transferred to the fourth layer, and the coloring of various gateways in the intermediate link Packet statistics report, and finally diagnose abnormal points based on statistical data.

2) Active detection

With the help of a large number of terminals to feed back the service quality from the perspective of the end user, through the refined mutual detection of its own thousand-level nodes, the automatic abnormality treatment is linked, and the third-party broadcasting and testing platform can also be used to obtain the service quality evaluation from the third-party perspective.

In the end, the original independent quality inspection of the third/fourth/seventh layer was upgraded to the third/fourth/seventh layer through correlation, forming a global perspective of accelerating the quality of the whole link of the whole station.

Create systematic security protection capabilities

When talking about network security, the first thing that comes to mind is traffic attacks such as DDoS. As a basic capability, it must be possessed. It is a resource-based capability. A single node does not have a bandwidth reserve of hundreds of G. It is difficult to form an effective The anti-D ability. Alibaba Cloud has sufficient resource pools to prevent traffic attacks.

Zhang Yiping pointed out: In addition to resource-side protection, it is also necessary to build a security system, especially for dynamic content. Dynamic acceleration content has the characteristics of small request bandwidth and high business value. Compared with the most basic violent traffic attacks, it is aimed at dynamic content. More is the crawler management, business anti-scratch and other functions. For example, various prices and orders placed on e-commerce websites are a large amount of dynamic content data, and there are more cases of crawling, ordering, etc., and the protection that needs to be done is more aimed at these. Through the establishment of a complete set of security data analysis system and in-depth cooperation with customers at the business level, we jointly define what dynamic requests are normal, what are abnormal, and what are the characteristics of abnormal requests for different businesses , How to deal with it in order to improve the security capability as much as possible while reducing the probability of accidental injury;

The establishment of a security protection system needs to rely on technologies and strategies accumulated in offense and defense, and this is the most valuable.

The core capabilities of network acceleration are also the foundation for building an edge cloud system. In various edge application scenarios, the acceleration of the underlying network and the interconnection between multiple clouds can help build an edge cloud ecosystem. Standardized interfaces, cross-network collaboration, and cross-cloud service deployment provide a broader and universal foundation for upper-layer services Facilities capacity.

At the end of the speech, Zhang Yiping expressed his views on future development: In the development of the edge cloud ecosystem, the delivery of scattered basic resources will become more and more difficult to implement. The delivery of scenario application solutions will be the mainstream direction, and open cooperation and win-win will be the main theme. Look forward to strengthening industrial cooperation and co-creating an edge cloud ecosystem.

Copyright statement: content of this article is contributed spontaneously by Alibaba Cloud real-name registered users, and the copyright belongs to the original author. The Alibaba Cloud Developer Community does not own its copyright and does not assume corresponding legal responsibilities. For specific rules, please refer to the "Alibaba Cloud Developer Community User Service Agreement" and the "Alibaba Cloud Developer Community Intellectual Property Protection Guidelines". If you find suspected plagiarism in this community, fill in the infringement complaint form to report it. Once verified, the community will immediately delete the suspected infringing content.