Documentation Center

Performance variables for Acceptance and Production

The performance of the main areas of your Tridion Sites implementation depends on a large number of performance variables, many of which are determined by your implementation choices. This topic enumerates for each performance area which variables you might measure or examine.

Typically, performance bottlenecks in a Production environment occur in one of the following areas:

  • Content Manager: users and/or processes interacting with the Content Manager server have to wait too long for a response.
  • Publishing throughput: publishing content takes too long.
  • Presentation Server serving of webpages: visitors of the website wait too long for a webpage to be loaded in their browser
Content Manager performance variables

As a rule of thumb, automated processes, most notably publishing, often impact scaling decisions more than, say, the number of concurrent users.

Automated processes include:

  • Publishing volume, that is, the total combined size of items being submitted for publishing per time unit
  • The complexity and speed of templating code
  • The complexity and speed of Event System code, if any
  • The complexity and speed of GUI extension code, if any
  • The complexity of Workflow Process Definitions, if any
  • The complexity and speed of other custom code (e.g integration with back-office systems, or on-the-fly data conversion)
  • The number of Core Service requests per time unit

Another set of factors is the number of users and the intensity of their interaction:

  • The number of concurrent users, defined by the formula (NAMED USERS * SINGLE USER USAGE PER PERIOD) / OVERALL USAGE PER PERIOD. For example, 200 different users interacting an average of 30 minutes per week with the system, while the system is in use for 40 hours per week, means that there are (200 * .5)/40 = 2.5 concurrent users.
  • The number of search requests per time unit
  • The number of item modifications by users per time unit

The third set of factors concerns the size and complexity of your implementation:

  • The expected number of Components and Pages
  • The number of Organizational Items, and the depth of the structure in which they are organized
  • The complexity of your Taxonomies, both the number of Keywords and the depth
  • The complexity of your BluePrint structure (including the number of Publications and the BluePrint depth)
  • The complexity of your security framework

And fourth, of course, the actual hardware you use affects performance. Note that because web and application server tiers are typically easier to scale out than database server tiers, the database tier will more often prove to be the performance bottleneck.

  • The number of Content Manager server machines
  • The CPU speed, number of CPUs and number of cores of your Content Manager server machine(s)
  • The RAM of your Content Manager server machine(s)
  • The number of your Content Manager database machines
  • The CPU speed, number of CPUs and number of cores of your Content Manager database machine(s)
  • The RAM of your Content Manager database machine(s)
  • The maximum throughput rate of the hard disk of your Content Manager database machine(s)
  • The client/server network latency
Translation Manager performance variables
Performance bottlenecks in a production environment typically occur in one of the following areas:
  • Translation throughput:
    • Sending/retrieving translated content takes too long.
  • Translation Manager performance variables (automated processes):
    • Translation volume—the total combined size of items being submitted for translation per time unit (or even translatable content)
  • The size and complexity of your implementation:
    • The expected number of translated items (Components, Pages, and so on)
    • The number of Organizational Items, and the depth of the structure in which they are organized (consider configurations)
    • The complexity of your Taxonomies, both the number of Keywords and the depth
    • The complexity of your BluePrint structure (including the number of Publications and the BluePrint depth)
    • The complexity of your security framework
  • The number of Translation Manager services configured in scale-out scenario.
  • Client/server network latency:
    • The bandwidth of the connection between the Content Manager and the translation system
    • The type and speed of the firewall you use, and how it is configured
Publishing performance variables
  • Publishing volume, that is, the total combined size of items being submitted for publishing per time unit
  • The complexity and speed of templating code
  • The complexity and speed of custom Content Deployer Modules, if any
  • The bandwidth of the connection between the Content Manager and Content Delivery systems
  • The type and speed of the firewall you use, and how it is configured
  • How static or dynamic your website is (dynamic websites publish faster)
  • The number of machines running a Content Deployer
  • The CPU speed and RAM of such machines
  • The number of database servers
  • The CPU speed and RAM of such servers
Web content serving performance variables
  • The visitor's connection speed
  • The number of page views per time unit
  • The number of concurrent visitors, that is, the number of active visitor sessions at any one time
  • The level of interactivity on webpages
  • The type of web and application server used
  • The amount and level of integration with other applications, such as back-office systems
  • The type and speed of the firewall you use, and how it is configured
  • The use of Tridion Sites's object caching
  • The use of file caching (e.g. proxy caching) - this slows down performance
  • The number of Content Delivery web service requests per time unit
  • The number of Content Delivery server machines
  • The CPU speed, number of CPUs and number of cores of your Content Delivery server machine(s)
  • The RAM of your Content Delivery server machine(s)
  • The number of your Content Delivery database machines
  • The CPU speed, number of CPUs and number of cores of your Content Delivery database machine(s)
  • The RAM of your Content Delivery database machine(s)
  • The maximum throughput rate of the hard disk of your Content Delivery database machine(s)
  • The connection speed between your various Presentation Servers
  • How static or dynamic your website is (static websites serve content faster)