Page tree
Skip to end of metadata
Go to start of metadata

To measure the performance of web applications that rely on external content providers, the system uses page-render time (PRT), which is the time for a page to load all content (whether it is delivered from your origin server or from a caching server).

Understanding PRT

PRT is an alternate measurement of the page-loading time and is not related to the end-to-end (E2E) time. It can be smaller or larger than the E2E time; furthermore, there is no breakdown of the PRT into network, host, SSL, redirect components as in case with E2E time.

PRT is calculated from the browser's perspective, so it is affected by the browser's host operating system and hardware. For example, if you are accessing a web application from a slow computer or from a smartphone, the PRT might be a lot larger than traditional E2E time, and this is the true user experience that the system is expected to report. Page Render Time (PRT) is calculated using the following metrics:

  • Parameters passed by the web beacon (page-render marker) in the JavaScript instrumentation
  • TCP-level time metrics calculated on the HTML page served

PRT calculation explanation

 

Page PRT = A + 1/2 RTT + B

 

  • A starts when the system identifies the beginning of an HTML page request. If there is a redirect request leading to the page, the time taken by the redirect is added to A. The A timer stops when EUEM detects the start of the HTML page response content (or end of response headers if no content) in the normal mode, or the end of the response in the instrumented-by-proxy mode.
  • RTT is the TCP level average round trip time from a packet sent to its ACK received.
  • B is the value of the crd_olt URI Query parameter passed by the web beacon (page-render market, or PRM).

If you are using a proxy server to inject the Edge instrumentation JavaScript routine, you must configure the system to specify which traffic is subject to that injection method.

Process for setting up the system to report page-render time

To perform this procedure, you must have Administrator-level access, or higher.

StepTask
1Connect the Real User Collector and the Real User Analyzer to the same time synchronization server or make sure that the difference of the system times on these components does not exceed 5 seconds.
2In the Real User Analyzer, ensure that you select Manage configuration for each Collector feed. For details, see Establishing traffic data feeds between an Analyzer and a Collector.
3Set up the Akamai instrumentation.
4Deploy a page-render beacon on the origin server.
5Enable detection of the web beacon.

After you configure the system to report page-render time, you can use various dashlets and reports to monitor the performance of applications whose content is delivered to end users by external providers. For example, the Global Application Delivery dashlet (Real User Analyzer) offers a visual representation of how the end-user-experience performance of your web application varies by geographic location.

Related topic

Defining page SLTs to measure application performance compliance on the Analyzer