Azure CDN improves application performance by caching content to the location nearest the customer.
CDN outbound data transfers
|Service||Standard version2||Standard Plus version3|
|CDN data transfer (0-10 TB1/month)||¥0.19/GB||¥0.396/GB||¥0.228/GB|
|CDN data transfer (10-50 TB/month)||¥0.16/GB||¥0.347/GB||¥0.198/GB|
|CDN data transfer (50-100 TB/month)||¥0.13/GB||¥0.297/GB||¥0.169/GB|
|CDN data transfer (more than 100 TB/month)||¥0.1/GB||¥0.277/GB||¥0.139/GB|
Which regions correspond to Zone 1 and Zone 2?
CDN data transfer pricing is based on the node location from where the transfers are served, not the end user’s location. The following geographic areas correspond to the zones for Azure CDN as listed above ：
• Zone 1 — Mainland China (Excluding Hong Kong, Taiwan and Macao).
• Zone 2 — North America, Europe, Asia Pacific, Mid-East, Africa, South America, please refer to oversea POP for more information.
Can I designate a CDN data center to provide services to end users?
No. The service will select a CDN data center based on an end user's network configuration, developers cannot decide which CDN data center to use. Users will be provided services according to their ISP's preferred location or a logically determined "closer" (not necessarily physically closer) node.
What’s the difference between Standard CDN service, Standard Plus CDN service?
- Standard CDN service refers to static contents acceleration, including Web Acceleration, Download Acceleration and Video-On-Demand Acceleration.
- Standard Plus CDN service refers to live-streaming acceleration.
Peak Value Bandwidth Billing Explanation
Peak value bandwidth billing mainly comprises monthly 95th percentile billing and average peak billing. You need to contact 21Vianet to set this up. The details are as follows:
1.Monthly 95th percentile billing
The billing period is a calendar month. In a calendar month, the billing subscription number takes the bandwidth values for every 5 minutes of each valid day (>1Kbps), sorts them in descending order, and removes the first 5% of statistical points. The remaining largest statistical point is the billed bandwidth. Generated bandwidth >1 Kbps is recorded as a valid day.
For example, in one 30-day month that contains 20 valid days, the bandwidth value is taken every 5 minutes, and 12 statistical points are taken every hour. The number of valid points is 12 x 24 x 20 = 5,760 points. All statistical points are then sorted in descending order and the first 5% (5,760 x 5% = 288) are removed, meaning that the billing point (Max95) is the 289th point, and the billing amount is Max95 x unit price x 20 / 30.
2.Average peak billing
The billing period is a calendar month. In any one calendar month, the peak values for each valid day of a billing subscription ID (>1Kbps) are averaged together, and this is the bandwidth used for billing. Generated bandwidth >1 Kbps is recorded as a valid day.
For example: if there are 20 valid days in a 30-day period, the peak value of each active day is taken as that day’s active bandwidth. This is the method used to calculate the bandwidth for each valid day. The sum of the values taken for all valid days is divided by the number of valid days, so the bill is calculated as follows: (Peak1 + Peak2 + ...+ Peak 20)/20.
What is the relationship between Content Delivery Network flow and back-to-source flow?
- Content Delivery Network traffic indicates cache hits.
- Back-to-source traffic indicates the missed portion.
What is the difference between log traffic and paid traffic?
The amount of network traffic data generated by CDN-accelerated domain names is 7–15% higher than the amount of log traffic data. This difference occurs because log traffic data is counted as application-level traffic. The Internet transmission process not only includes app log traffic, but also additional network usage caused by TCP/IP headers and TCP retransmission.
- TCP/IP header usage: Internet data packets based on the HTTP protocol are 1,500 bytes long, including the 40-character headers for the TCP and IP protocols. Traffic generated by headers cannot be counted at the application level, so only 1,460 bytes of traffic are recorded in application-level logs. For this reason, packet headers account for 2.74% (40/1,460) of application-level log traffic, so there is a data error of around 3%.
- TCP retransmission: Fluctuations in Internet networks cause around 3–10% of data packets to be discarded by the Internet. However, servers retransmit these lost data packets using the TCP protocol’s retransmission mechanism. Retransmitted data traffic is also not counted by the application layer.
What factors impact the availability of content in the CDN's local caches, and how can I reduce the need to make frequent origin requests?
The availability of content in the CDN's local caches (often called "cache efficacy" or "offload") will be influenced by many factors, including:
- Expiration (“max-age”) header values
- Overall total size of the developer's content library (how much can be cached)
- Active working set (how much is currently cached)
- Traffic (how much is being provided)
- Cache churn (how often are objects being added to cache, or aging out)
For example, a developer with higher churn and traffic has less cache efficacy than other users because objects are swapped in and out more frequently. This will lead to higher storage and data transfer charges, as more origin requests are required.
To reduce the need to make origin requests, you can set longer max-age headers, thereby allowing the CDN to hold objects longer.
Are all data transfers charged at a higher rate once the amount of data transferred each month exceeds 10 TB?
No. Usage at each tier will be charged according to the rate for that tier. For example, if you generate 50 TB of Standard version CDN data transfers in Zone 1, the first 10 TB will be billed at a rate of ￥0.19/GB, and the remaining 40 TB will be billed at a rate of ￥0.16 per GB.
Do CDN charges include requests to Storage to retrieve data, and data transfers from Storage to the CDN location?
No. When the CDN receives an object request for a non-edge location, it will make a request to Azure Storage to obtain the data. Operations for reading data from Storage and transferring it from Storage to the CDN will be charged at standard data transfer rates.
Support & SLA
If you have any questions or need help, please visit Azure Support to select self-service or any other method to contact us for support.
We guarantee that CDN will respond to client requests and deliver the requested content without error at least 99.9% of the time. We will examine and accept data from any commercially reasonable independent measurement system that you choose to monitor your content. You must select a set of agents from the measurement system's list of standard agents that are generally available and represent at least five different geographical locations in major metropolitan areas of the People's Republic of China. If you want to learn more about the details of our server level agreement, please visit the Service Level Agreement page.