What impacts the availability of content in the Content Delivery Network’s local caches, and how can I reduce the need to make frequent origin requests?

The availability of content in the Content Delivery Network’s local caches (often called "cache efficacy" or "offload") is influenced by multiple factors including:

  • Expiration (“max-age”) header values
  • Overall total size of the developer’s content library (how much could be cached)
  • Active working set (how much is currently cached)
  • Traffic (how much is being served)
  • Cache churn (how often are objects being added to cache, or ageing out) For example, a developer with high churn and high traffic has less cache efficacy than other users because objects are swapped in and out more frequently. This incurs higher Storage and Data Transfer charges, as more origin requests are required. To reduce the need to make origin requests, you can make longer max-age headers, allowing the Content Delivery Network to hold objects for longer.

Content Delivery Network

Related questions and answers

  • No. When the Content Delivery Network receives a request for an object that is not at an edge location, it makes a request to Azure Storage to obtain the data. The cost of reading data from Storage and transferring data from Storage to Content Delivery Network is based on regular Storage and Data Transfer charges.

  • Content Delivery Network data transfer pricing is based on the node location from where the transfers are served, not the end user’s location. The following geographical areas correspond to the zones for the Azure Content Delivery Network as listed above. For a detailed listing of node locations, please refer to the documentation for Point of Presence (PoP) locations.

    • Zone 1: North America, Europe, Middle East and Africa
    • Zone 2: Asia Pacific (including Japan)
    • Zone 3: South America
    • Zone 4: Australia
    • Zone 5: India
  • One price covers geographies served by existing Content Delivery Network zones.

  • No. The data centre is selected based on the end users’ network configurations and cannot be controlled by the developer. Users may be served by locations that are preferred by their ISP or nodes that are “closer” in a logical sense, not necessarily in physical proximity.