Is Muti CDN really easy to use?

How much impact will a single outage of a large internet enterprise have?

A foreign netizen replied, “I thought the end of the world was coming!”

This is the roast of some netizens after the massive downtime of servers on Facebook, Instagram and other platforms on April 15, 2019, which shows the importance of stable network services. How to effectively ensure the service quality and user experience of their own websites and APPs has become a priority issue that every enterprise must consider.

How cautious are these companies? This can be seen from the practice of Internet companies purchasing CDN (content distribution network) services. Especially for video platforms such as live streaming and on-demand, even under the pressure of burning money, they are willing to sacrifice costs and expend energy to purchase multiple CDN services at one time, spreading risks and praying for peace.

It is against this background that Convergence CDN emerged, opening up a new circuit in the red sea of CDN competition with advantages such as “high stability, ease of use, and fast service response.”. After tasting the sweetness, Internet companies of all sizes have included integrated CDNs in their procurement lists.

Where there is a market, there is competition. In a few years, integrated CDNs have blossomed everywhere, starting with the introduction of integrated CDNs by manufacturers of all sizes. It seems that everyone can do integrated CDNs. But is this really the case?

Playing with concepts can’t solve real problems

Conceptually, anyone can indeed do it – by establishing agency cooperation with multiple CDN manufacturers at the same time, and being able to switch traffic between various manufacturers according to their own or customer wishes.

However, this approach only stays at the conceptual level and is essentially meaningless. The so-called integrated CDN platform only serves as the customer’s outsourcing procurement team or CDN operation and maintenance department. The pain points of the original CDN service have not been resolved: when customer traffic is sudden, it is still necessary to manually ask around to see who can pick up; If there is a network fault, it is still necessary to ask, troubleshoot, and resolve it one by one. If switching is necessary, it is also necessary to repeatedly communicate. The process has not changed, the efficiency has not changed, the stability has not changed, and the end of the world (customer business damage) results have not changed.

Through interviews with industry insiders, it is understood that integrating CDN and aggregating multiple resources is only the first step, even the simplest one. In the context of the red sea competition in the CDN industry and the business pressure faced by various manufacturers, it is not difficult to establish business cooperation with them. Currently, the vast majority of integrated CDNs on the market are still in the first step. The key factor to truly solve the pain point is actually the technical part under the water.

We know that the pain point of the original CDN service model is not that it cannot be switched, but that due to the isolation and fragmentation between multiple vendors, it is difficult to quickly identify and locate problems, and it is difficult to master the entire network status. After discovering problems, the switching process requires repeated communication and wastes time, thereby harming the business.

Correspondingly, the key to integrating CDN is not to be able to switch, but to be able to control the quality of the entire network at any time, detect network traffic bursts or failures at the first time, quickly locate problems, and then make scheduling arrangements. The entire process is smooth and efficient, even without manual communication.

Take a chestnut. For example, from City A to City B, there are many roads, and the integration of CDN with advertised concepts is just to give you a pass to ensure that you can walk on multiple roads. If this road doesn’t work, you can switch to another one – as old drivers know, it will be difficult to turn around when the road ahead doesn’t work. A truly integrated CDN is not just about having a pass, but rather having the navigation function of Baidu Map or Gaode Map, which constantly tells you which roads are easy to walk with fewer cars, which roads are slow in accidents, and which roads are under renovation. Never go

Therefore, the “predictive ability” of network monitoring, the “positioning ability” of data analysis, and the “decision-making ability” of intelligent scheduling are the core difficulties and core values of integrated CDN, and such integrated CDN has its professional name in the market – intelligent integrated CDN.

Putting aside intelligence and integration, we can basically conclude that we are playing with concepts.

Integrate the “smart” components of CDN

So, in the field of integrated CDN, when we talk about intelligence, what do we talk about? Let’s learn from Huosan Cloud that in order to achieve “intelligence” and achieve the goal of stability, ease of use, and efficiency, integrated CDN must have the following four technical elements:

1. Monitoring system

Intelligent Convergence CDN does not require CDN nodes, but requires a sufficient number of monitoring nodes, coupled with the network monitoring, channel, and customer monitoring systems of Fireumbrella Cloud, to be able to grasp the status of CDN nodes, operator backbone networks, customer origin sites, HTTPS certificates, and DNS resolution in real time, and to collect sufficient data information.

Real-time merging, denoising, and correlation analysis of the data collected by various monitoring methods can identify problems in the first time and form conclusions that can be used by the scheduling system or operators for decision-making.

2. Big data analysis

The network quality information collected by the monitoring system may be invalid, chaotic, and repetitive. A powerful big data analysis system is needed to filter and analyze, then identify normal business fluctuations and abnormal changes from it, and provide coping strategies based on historical data. Some intelligent fusion CDN manufacturers’ big data systems are not only capable of real-time analysis of monitoring information, but also capable of learning evolution. For example, their thresholds are automatically adjusted based on historical data, achieving automatic detection of normal business fluctuations and abnormal changes. This is also an important reason why Convergence CDN is named “smart”.

3. Intelligent scheduling

The intelligent scheduling system is linked with the monitoring and analysis system to make intelligent scheduling decisions based on the fault level. For availability failures, node failures, abnormal status codes, etc., it automatically switches based on the set threshold value; Conduct manual optimization for abnormal situations such as network jitter, download speed, and response time.

It is worth mentioning that due to its numerous channels, intelligent fusion CDN can implement a quality first scheduling strategy after finding faults that are difficult to solve, automatically scheduling first, and then solving the problem, with almost no user awareness.

Note that by this time, scheduling has occurred, and most of the monitoring and analysis work before that has already been completed. Therefore, the core issue that intelligent fusion CDN solves is actually why to schedule; The conceptual integration of CDN only achieves “schedulability”. This is the essential difference between the two.

4. Aggregation management

On the market, every sizeable Internet enterprise has many channels connected to it, and theoretically requires management. The intelligent fusion CDN is more professional and further developed, with a management system specially configured for its research and development. It can control the operational status of the entire integrated CDN platform, achieving comprehensive control from basic data to operation, maintenance, and operations, whether it is resources, services, status, task system, API system, or monitoring, scheduling, billing, and problem follow-up, to achieve global controllability.

For customers, this is reflected in ease of use. Whether it is to increase or decrease channels or change configurations, they only need to communicate with one platform. “One purchase, multiple stores available, one-on-one communication, is really a lot of convenience.”.