In total station acceleration, intelligent scheduling is one of the important technologies for achieving CDN network acceleration. The purpose of this technology is to gather users in the same network environment into the same service group, and allocate the optimal service node based on the network and capacity of the CDN server. Simply put, the whole station accelerated intelligent scheduling is to provide users with the most suitable CDN node to access and obtain network services nearby. There are generally two methods of intelligent scheduling: DNS scheduling and central scheduling.

DNS Scheduling

Mainly based on Local DNS, it is used for scheduling across regions and operators, with high scalability and good performance. It is generally used in full-site acceleration scenarios such as web pages; This scheduling method relies on DNS recursive resolution, passing CNAME to the GLSB of the CDN, and then allocating adaptive nodes based on the geographical location of Local DNS. This method has some shortcomings:

1. Scheduling accuracy may be insufficient, such as: Guangdong Telecom users resolve through local DNS, and CDN will allocate Guangdong Telecom’s service nodes; If the user customized and configured a local DNS address (114.114.114.114), this method will assign a service node in Nanjing, Jiangsu Province,

2. DNS service provider servers are vulnerable to DDOS attacks, resulting in DNS resolution failures and affecting user access.

How to implement intelligent scheduling in DCDN?

Central dispatching (302 dispatching)

Based on the IP address of the end user, precise HTTP redirection requires protocol support and considerable latency, and is generally used in streaming media acceleration scenarios; The scheduling method is to obtain the IP address of the GLSB cluster of the CDN through DNS resolution. The user sends an HTTP request, and the GLSB server returns “302 Found” to redirect the access to the appropriate service node. This method also has some shortcomings:

1. Limited to HTTP applications, with insufficient scalability;

2. “302 jumps” are added to the scheduling process, resulting in a longer delay compared to DNS scheduling. “.

For Web pages, users have high requirements for the loading and display time of the web page, so the scheduling method used for site wide accelerated applications is generally intelligent scheduling based on DNS. In response to the problems and shortcomings of DNS scheduling, functional optimization has been carried out, mainly including: nearby access, CloudDNS intelligent cloud resolution, EDNS support, and self recovery from failures.

Accelerated access to nearby sites throughout the station

1. With a wide distribution of users, it is necessary to ensure the access effectiveness of users in various regions. The global load system of CDN will allocate the best node for services based on the location of DNS exits, such as: users from Guangdong Shenzhen Telecom are allocated to nodes from Guangdong Shenzhen Telecom for services;

2. If a bandwidth burst occurs in a single node computer room, which temporarily affects the service performance of the node, the nearby access function of the intelligent scheduling system can ensure the service performance of the node. For example, if a Guangdong Shenzhen Telecom node fails, it can allocate a nearby Guangdong Dongguan Telecom node for service, ensuring customer service quality, and ensuring that the end user experience is not affected.

The global load balancing system for accessing CDN nearby provides users with access to the nearest CDN service node based on the location of DNS exits and real-time monitoring of node performance.

CloudDNS intelligent cloud resolution

To address the vulnerability of DNS resolution services to attacks and the pressure of domain name management, we provide self-service, reliable, stable, and secure DNS advanced defense solutions, providing customers with domain name management, intelligent resolution, domain name monitoring, traffic management, backup scheduling, log services, advanced defense attacks, and other services.

1. Domain name resolution server attack scenarios: Domain name resolution servers are vulnerable to attacks that exceed the attack threshold, the domain name resolution service is affected, and customers are blackmailed. CloudDNS can provide domain name protection services.

2. Numerous management and frequent modification scenarios for domain names: Numerous domain names, complex management, and slow effective time for modification and configuration; CloudDNS provides API self-service management, improving deployment efficiency, simple configuration, and resolution and modification taking effect in seconds.

3. A wide range of users can be accessed to accurately analyze scenarios: distributed source stations, where users from different regions access appropriate source stations; CloudDNS has rich line resources and accurate IP positioning, which can distribute traffic to appropriate source stations.

4. Customers attach importance to resolution audit scenarios: it is necessary to regularly query the resolution of domain names, as well as the traceability query of operational changes to domain names; CloudDNS provides resolution log viewing capabilities, log queries, and real-time understanding of all domain name operations.

EDNS support

Scheduling to a non optimal node due to incorrect DNS settings on the user side; Support EDNS service, support DNS to carry IP packets, and accurately schedule based on IP, effectively avoiding cross zone access caused by DNS settings errors on the user side. For example, for users of Guangdong Telecom, the DNS configuration is Jiangsu Nanjing Telecom (114.114.114.114), and the CDN supports EDNS, which means that the DNS server carries the user IP. The global load balancing system of the CDN will allocate nodes of Guangdong Telecom for service based on the user IP.

Fault self recovery

CDN’s intelligent scheduling system can monitor the operation of node computer rooms in real time. If a service failure occurs in a single node computer room, the intelligent scheduling system judges the failed node and automatically switches to a normal node, ensuring customer service quality and ensuring that the end user experience is not affected.

Huosan Cloud Integrated CDN Whole Station Acceleration Platform can monitor the service status of CDN nodes in real time, dynamically select the nodes with the best service to provide users with network access services, ensure customer service quality, and ensure that the end user experience is not affected.

相关新闻

微信

15323421841