This topic describes the benefits of Dynamic Route for CDN (DCDN). Compared with other whole site acceleration service providers, Alibaba Cloud DCDN has the following benefits: stability, high access speed, scalability, and cost-effectiveness.
- Wide service coverage: Alibaba Cloud has deployed over 2800 nodes across the world. A bandwidth of 80 Gbit/s per node and a total network-wide bandwidth of 120 Tbit/s are supported. The node network also supports tens of millions of QPS to ensure the stability of the acceleration service.
- Advanced distributed system architecture: Supports load balancing to ensure node availability.
- High stability and performance: A cache hit ratio of more than 95% can be reached for static caches. Access speeds in mainland China and outside mainland China can be improved by 100% and 200%, respectively. In addition, the availability in mainland China and outside mainland China can reach 99.99% and 99.95%, respectively.
- Optimized transmission protocols: Supports HTTP/2 and SPDY transmission protocols for fast and stable data transmission.
High access speed
- Accurate caching: Uses the smart object algorithm to cache content based on object popularity in a multi-level and hierarchical manner to achieve accurate content delivery acceleration.
- High-speed caching: A high-performance caching system is designed to balance the use of multi-core processing capabilities, facilitate the efficient memory usage and control, and maximize Input/output operations per second (IOPS) and throughput of SSDs.
- High read and write speeds: All nodes are configured with SSDs to support high-speed data reads and writes. With the acceleration capabilities, these SSDs can greatly reduce the response time and improve the availability.
- Efficient back-to-origin process: Implements the session persistence feature.
- Divides the back-to-origin path based on the client IP address. This ensures that logon session information is not accessed across origins and resolves the issue that information on multiple origins cannot be shared.
- Provides failover and retransmission mechanisms to ensure efficient back-to-origin process and information synchronization.
- Intelligent scheduling
- Multi-level scheduling: Even if some nodes fail, service unavailability will not occur.
- Multi-system coordination: Coordinates with the security defense system, refresh system, and content management system to achieve the best performance of each module.
- Data-based real-time scheduling: Supports node-level traffic prediction to improve the scheduling quality and accuracy.
- Ordered processing of back-to-origin requests:
- Automatically responds to and adjusts for traffic spikes. DCDN implements QPS-based throttling for back-to-origin requests to protect the availability of the origin server.
- The waiting room solution allows you to customize the waiting page, waiting duration, and release rules to maximize user experience.
- Resources can be elastically expanded and paid based on actual usage. You can achieve network-wide coverage across providers and regions.
- DCDN provides a wide range of simple custom configuration options to allow you to add, delete, modify, and query domain names. DCDN also allows you to customize caching policies, HTTP response headers, and other features.
- DCDN provides a wide array of API operations. You can call these API operations to use multiple DCDN features including service activation, content refresh, monitoring, and log download and distribution.
- Performance optimization
- Intelligent compression: compresses content transmitted over the network. This can reduce the number of bytes transmitted over the network, shorten the data transmission time, and improve acceleration.
- Refresh and prefetch: provides the cache refresh and object prefetch features.
- Optimal link: Makes routing decisions and selects the optimal back-to-origin path for dynamic content delivery by using origin monitoring, real-time network monitoring, and smart routing technologies.
Link multiplexing: Edge nodes accept connection establishment requests from users and establish persistent connections based on actual back-to-origin requests. This can reduce the I/O pressure on the origin server.
- Analyzes inbound and outbound traffic conditions to ensure that the client can be reached within a limited number of hops, avoiding network failures or congestions.
- The last hop to connect to edge nodes adopts BGP multi-line. This can ensure efficient back-to-origin process under the same region and network provider.