Before installing PaperCut, Sys Admins often want to know what type of hardware they need to run a quick and responsive system in their specific environments. That’s quite a hard question to answer, but hopefully, these guidelines provide an answer that makes sense for you.
There are a large number of factors that affect the size of systems needed. For example:
Number of average and peak users simultaneously using the system. Type of print jobs (file size, printer language in use, printer options selected, print scripts in use etc.). Use and configuration of Find-Me printing queues. Features such as Print Deploy, Integrated Scanning, Mobility Print, Web Print, Email to Print, Print Archiving etc. Use of our PaperCut embedded solution on Multi-Function Devices if you’re using PaperCut MF. Use of Payment Gateways and other third party external interfaces can also impact performance and are not under the control of the PaperCut application. The number of print queues on a particular print server and volume of jobs that will be on hold at any one time. Scalability & Disaster Recovery We recommend virtual instead of dedicated physical servers. Virtual machines benefit from using any reserved amount of the host’s complement of processing, storage, and memory resources as well as the vendor’s built-in or customised back-up and restore capabilities. The basic recommendation for CPU, RAM, and Storage is for them to be to set statically rather than dynamically.
We also have additional information regarding your environment growing and scaling over time here.
Server CPU Don’t skimp on CPU resources! The typically recommended CPU requirement is to start with two and then to extrapolate from there depending on the size of your organisation and this KB’s suggestions. See the Server Sizing Table at the bottom of the page for additional help.
Server RAM When building a Virtual Server, specifically for PaperCut NG or MF, start at 4GB of RAM and make sure the memory is reserved, not dynamically allocated.
The PaperCut application will use 1/4 of the server’s available memory by default, but it’s possible to configure PaperCut to use up to 1/2 of the memory by making the change discussed here.
The PaperCut application itself is a “process virtual machine” operating in a Java runtime environment. The JRE’s garbage collection process tries to keep the memory as neat as possible which means having more RAM than required for your organisation can negatively impact the application’s performance. Sometimes, less is more!
Storage Database Consider that the PaperCut database grows at an approximate rate of 4.5Mb per 10,000 jobs. For example: allowing for 4 print jobs per user and a 30% increase that comes out at just under 1/2 Gb over three years. As our storage requirement is so low we do not suggest archiving. You can calculate your own value using the guidelines here.
Print Server Print servers require disks capable of fast data transfer and enough space for print job spooling. The minimum requirement overall is 100 GB of free space as discussed here. (10+ GB depending on the number and size of the print jobs). PaperCut’s job analysis is an intensive Read/Write operation, and modern enterprise-class server disks should support its basic requirements.
Server Sizing Table The table below shows guidelines for the server specifications that are required to support PaperCut NG or MF in order to give consistent performance for a-typical printing use at a site. Please note that Users refers to the number of active users per day. These guidelines apply both to physical hardware and to virtual servers / virtual machines hosted locally or in a Private Cloud.
PaperCut Authorized Partners are able to offer architecture and infrastructure advice for large deployments if you’re a current PaperCut MF customer. Also please see our System Requirements page for more guidelines on minimum and recommended server requirements.
Note: The specifications in the table below are meant as a guide only! Nothing substitutes volume testing before going live!
Printers vs Users 5
Printers 20
Printers 100
Printers 1,000
Printers 5,000
Printers 10,000
Printers 1,000
Users 2 CPUs
4GB RAM 2 CPUs
4GB RAM 5,000
Users 2 CPUs
4GB RAM 2 CPUs
4GB RAM 2 CPUs
4GB RAM 10,000
Users 3 CPUs
4GB RAM 3 CPUs
4GB RAM
+ RDBMS 3 CPUs
4GB RAM
+ RDBMS 3 CPUs
8GB RAM
+ RDBMS 20,000
Users 4 CPUs
6GB RAM 4 CPUs
6GB RAM
+ RDBMS 4 CPUs
6GB RAM
+ RDBMS 4 CPUs
8GB RAM
+ RDBMS 4 CPUs
8GB RAM
+ RDBMS 60,000
Users 4 CPUs
6GB RAM
+ RDBMS 4 CPUs
6GB RAM
+ RDBMS 8 CPUs
12GB RAM
+ RDBMS 8 CPUs
16GB RAM
+ RDBMS 10 CPUs
16GB RAM
+ RDBMS 100,000
Users 4 CPUs
8GB RAM
+ RDBMS 8 CPUs
16GB RAM
+ RDBMS 10 CPUs
16GB RAM
+ RDBMS 10 CPUs
16GB RAM
+ RDBMS 400,000
Users 10 CPUs
32GB RAM
+ RDBMS 10 CPUs
32GB RAM
+ RDBMS 10 CPUs
32GB RAM
+ RDBMS +RDBMS: Separate database servers are recommended where noted. See the Deployment on an external database (RDBMS) section of the PaperCut MF and NG manual for more information.
Geographic Considerations Having a single main PaperCut server is ideal for having a single dashboard to track printing worldwide. That being said, for multinational customers that want to roll out PaperCut we recommend having a PaperCut server for each continent or country.
Why?
At the moment you can only set one currency to use on your PaperCut server. Keep user and print job metadata in-region to comply with local regulations. For super large customers with hundreds of thousands of users or thousands of devices, having vast geographic distances between sites can mean that calls to the server take longer, leaving threads open for a greater amount of time, which can compound the load on the server. For these customers that need reports and data that transcend national boundaries, look into our Central Reports solution.
Comments
0 comments
Please sign in to leave a comment.