All Products
Search
Document Center

CDN:Performance indicators

Last Updated:Sep 21, 2023

This topic describes the major performance indicators that are used to evaluate the performance of content delivery before and after a website is accelerated by Alibaba Cloud CDN.

You can make informed business decisions based on the performance indicators. Performance indicators include:

Common performance indicators

You can monitor the following performance indicators to evaluate the performance of content delivery before and after your website is accelerated by Alibaba Cloud CDN. Common performance indicators include but are not limited to the ones described in this topic.

Note
  • After a website is accelerated by Alibaba Cloud CDN, the network latency, packet loss rate, and back-to-origin rate are typically reduced, and the cache hit ratio is increased. However, the actual performance may vary based on the business scenario and workload type. The performance indicators that are described in this topic are based on pure statistics.

  • The back-to-origin rate and cache hit ratio that are described in the following table are used to evaluate the performance of content delivery for a website after it is accelerated by Alibaba Cloud CDN. If your website is not accelerated by Alibaba Cloud CDN, the back-to-origin rate is 100%, and the cache hit ratio is 0.

  • How content is delivered to a client上传

  • How content is downloaded by a client下载

Performance indicator

Description

DNS time

The amount of time required for a client to initiate a request and receive the IP address of the destination host.

TCP time

The amount of time required for a client to establish a TCP connection to the destination server.

SSL time

The amount of time required for a client to establish an SSL connection to a web server.

Delivery time

The amount of time required for a client to complete sending a request after SSL handshakes are completed.

Connection time

If a POP uses HTTP to accelerate content delivery, the connection time consists of the DNS time and TCP time. If a POP uses HTTPS to accelerate content delivery, the connection time consists of the DNS time, TCP time, and SSL time. The connection time shows the coverage of POPs and the capabilities of the POPs to deliver content.

Response Time

The amount of time required for a web server to process an HTTP request and return a response to a client.

Download time

The amount of time required for a client to receive and download the first packet returned from a web server.

Time to first packet

The amount of time required for a client to send a request and receive the first HTTP packet from a server. The time to first packet shows the overall performance of POPs.

For content uploading and downloading, the time to first packet consists of the DNS time, TCP time, SSL time, request time, and response time.

Note

A new domain name may require a longer period of time for DNS resolution than other domain names. However, this does not affect the cache retrieval time.

Overall performance

The amount of time required to upload or download an entire file.

Packet loss rate

The rate of lost packets to total packets during transmission.

Back-to-origin rate

The back-to-origin rate is classified into two: back-to-origin request rate and back-to-origin data transfer rate.

  • The back-to-origin request rate refers to the rate of requests for resources that are not cached, have expired, or cannot be cached on POPs to the total number of requests. Back-to-origin request rate = Number of back-to-origin requests from POPs/Total number of requests sent to POPs. A lower back-to-origin request rate indicates better performance. However, if the original user requests are fragmented after POPs redirect the requests to the origin servers, the number of back-to-origin requests becomes greater than the total number of requests that are sent to POPs.

  • The back-to-origin data transfer rate refers to the rate of data transfer that is returned by the origin servers to data transfer that is returned by POPs to clients. Back-to-origin data transfer rate = Number of bytes returned from the origin servers to POPs/Number of bytes returned from POPs to clients. A lower data transfer rate indicates better performance.

Cache hit ratio

The cache hit ratio of Alibaba Cloud CDN is classified into the byte cache hit ratio and request cache hit ratio. The cache hit ratio of Alibaba Cloud CDN is represented by the byte cache hit ratio. A higher cache hit ratio indicates a higher performance.

  • Byte hit ratio = (Total number of bytes returned from POPs to clients - Total number of bytes returned from the origin servers to POPs)/Total number of bytes returned from POPs to clients.

    Note

    A lower byte hit ratio indicates a higher volume of origin traffic. A higher volume of outbound traffic from the origin server indicates a larger bandwidth value and heavier workloads of the origin server. Origin traffic represents the amount of workloads on the origin server, and the byte hit ratio is a major concern in actual business scenarios.

  • Request hit ratio = (Total number of requests to POPs - Total number of back-to-origin requests)/Total number of requests to POPs.

Performance indicators for small file distribution

Small files refer to HTML, JavaScript, JPG, and CSS files. The delivery of these types of files is latency-sensitive. Lower latency indicates higher performance. The latency is determined by the following factors:

  • Time to first packet (most critical factor)

  • Connection time

  • Download time

Alibaba Cloud CDN provides common scenarios that small file distribution is suitable for. For more information, see Image and small file distribution.

Performance indicators for large file distribution

Large files refer to files that are larger than 20 MB. The key indicators for large file distribution include:

  • Download rate

  • Download time

Alibaba Cloud CDN provides common scenarios that large file distribution is suitable for. For more information, see Large file distribution.

Performance indicators for on-demand video and audio streaming

On-demand audio and video streaming uses FLV, MP4, WMV, and MKV files. The key indicators for on-demand video and audio streaming include:

  • Initial load time

    The amount of time required to complete loading the first frame of a stream. The initial load time is determined by the DNS time, connection time, and time to first packet. A shorter initial load time indicates higher performance.

  • Stalling rate

    Stalling events may occur when a video or audio stream is played or a resource is loaded. The stalling rate is calculated by using the following formula: Number of viewers that have stalling events/100. A lower stalling rate indicates better performance.

Alibaba Cloud CDN provides common scenarios that on-demand video and audio streaming is suitable for. For more information, see On-demand audio and video streaming.