Scaling your Web Print environment

When using Sandbox mode, we recommend a minimum of two Web PrintWeb Print enables printing from user-owned devices without the need to install printer drivers and manage server authentication. servers. Using only one Web Print server causes a single point of failure and potential bottlenecks. Best practices in server management recommend building redundancy into your network and monitoring the health of your environment.

However, two Web Print Servers might not be enough for sites that see very high Web Print throughput. While you can simply continue to add Web Print Servers until the Web Print performance during peak printing hours is acceptable, sites with particularly demanding throughput will benefit from using statistics for their capacity planning.

Analyzing these statistics will help you plan your Web Print deployment. While we recommend a minimum of two Web Print servers as a starting point, the unique demands and performance characteristics of your environment are the key drivers in determining what you need. You will need to balance the acceptable average wait time that is appropriate for your users, against the resources needed to deploy additional Web Print servers.

Determine the number of pending Web Print jobs

To statistically determine the number of Web Print servers you need to adequately handle your throughput, you should evaluate the number of pending jobs your system has awaiting Web Print processing.

The Print System Health interface provides a monitoring endpointAn endpoint is a connection point where HTML files or active server pages are exposed and can be accessed by an external application. Endpoints provide information that can be consumed by an external monitoring tool. that tracks the number of pending print jobs in a Web Print queue at a specific point in time. You can use this data to determine the average number of pending jobs as well as the maximum number during peak periods. For more information about this endpoint, see Base system health URL.

A direct link to this endpoint is available in the Admin web interface:

1. Select Options > Advanced.

   The Advanced page is displayed.

2. In the System Health Monitoring section, click the Detailed system information link.
   

The attribute is webPrint > pendingJobs.

You can use a monitoring tool to accurately track how this value changes over time. A large number of pending Web Print jobs that rarely or never varies during printing hours, might indicate that users are experiencing consistently high wait times. If, however, the number of pending jobs peaks at random points in time, the average wait time might not be an issue. You can test if this is the case by submitting a number of test Web Print jobs during a peak print period, and timing how long it takes for the job to process. Good test documents would be a one page PDF and a one page Microsoft Word document.

1. Log in to the User web interfaceThe User web interface provide a range of services for users, including a summary of usage and balance history, a list of the shared accounts that the user can use for printing, the current costs for printing usage, ability to add balance by using a TopUp/Pre-Paid Card or an external payment system (when using the payment gateway module), transfer funds to other users, view a history of balance transactions, view a list of the user's recent printing, and view print jobs pending release (when using a Release Station)..
2. Submit a test Web Print job.

3. On the Web Print page, time how long it takes for each job to progress in the queue from having a Status of Queued in position to Finished: Queued for printing.

   Tip:

   You might need to manually refresh the page to keep across the changes in job status.

Averaging out the wait times incurred for test jobs during these periods of sustained high load provides an idea of the average time it takes to process a Web Print job in your current environment.

View historical Web Print trends

Historical data is another good tool for tuning your Web Print scaling setup. The number of jobs processed during key periods (for example, during peak times), will provide you with an estimated mapping of job numbers to average queue length. You can use this value to estimate future average queue lengths based on expected changes in the Web Print job load.

To generate a Web Print report:

1. Select Reports > Printer.

   The Printer Reports page is displayed.

2. In the Print Log Reports area, for the Print logs report, select Adhoc from the drop-down list.

   The Ad-hoc Report: Print logs page is displayed.

3. In From date, enter the start date for the period you want to run the report.
4. In To date, enter the end date for the period you want to run the report.
5. In Print method, select Web Print.
6. In Report Format, select CSV.

7. Click Run Report.

   The report will be a spreadsheet containing information for all Web Print jobs performed over a specified interval of time.

It is a good idea to also compare the average wait times for jobs that can be processed by any of your Web Print servers (for example, PDF files) versus those that can be processed only by servers equipped with additional licensed software (for example, Microsoft Word documents). If jobs that require licensed software are taking significantly longer than those that can be processed by any Web Print server, you might want to consider purchasing additional licenses for the software and deploying it to a larger proportion of your Web Print servers.

You can use this report to determine the proportion of Web Print jobs of each file type, which can indicate if you need to license additional software for your Web Print servers. To do this, generate a separate report for each file type by entering values, such as, .doc, .pdf, .xls, .jpg, in the Document name field. Sum the data provided in the Total Printed Pages or Size (KB) columns to give you an idea of the typical complexity of submitted jobs; a higher file size or page count indicates that a printed document was more complex, and thus would take longer for a Web Print server to render than a smaller job.

For a Web Print scaling example, see Web Print Scaling Example.