The following table lists the recommended server configurations for a variety of client and server scenarios (with Read performance being one of the most significant requirements).
In the table, the term ‘object’ refers to a database item whereas ‘point’ refers to a specific type of database item that stores I/O (a database point). Historic points are database points that have the Historic enabled.
| Application | Server Requirement |
|---|---|
|
Small stand-alone client-server with up to 10,000 objects and 1-2 remote clients |
Quad-core processor, 8GB RAM, 50GB free disk space. Desktop class hardware and operating system. |
|
Server with 10,000 objects, 5,000 historic points stored for 2 years online, 3 or more clients |
Quad-core server class processor, 8GB RAM, 25GB free space on high performance disks for database configuration, log files, and so on. A separate 250GB high performance disk system for historic data. Server class hardware and 64-bit operating system. |
|
Server with 50,000 objects, 25,000 historic points stored for 2 years online, multiple clients |
Quad-core server class processor, 8GB RAM, 25GB free space on high performance disks for database configuration, log files, and so on. A separate 500GB high performance disk system for historic data. Server class hardware and 64-bit operating system. |
|
Server with 100,000 objects, 75,000 historic points stored for 2 years online, multiple clients |
Quad-core server class processor, 16GB RAM, 50GB free space on high performance disks for database configuration, log files, and so on. A separate 1TB high performance disk system for historic data. Server class hardware and 64-bit operating system. |
|
Server with 175,000 objects, 100,000 historic points stored for 2 years online, multiple clients |
Quad-core server class processor, 16GB RAM, 50GB free space on high performance disks for database configuration, log files, and so on. A separate 2TB high performance disk system for historic data. Server class hardware and 64-bit operating system. |
|
Server with 250,000 objects, 150,000 historic points stored for 2 years online, multiple clients |
2 x quad-core server class processor, 32GB RAM, 100GB free space on high performance disks for database configuration, log files, and so on. A separate 4TB high performance disk system for historic data. Server class hardware and 64-bit operating system. |
|
Server with 450,000 objects, 250,000 historic points stored for 2 years online, multiple clients |
2 x quad-core server class processor, 48GB RAM, 100GB free space on high performance disks for operating system, installed applications, database configuration, and so on. A separate 50GB high performance disk system for log files. A separate 4TB high performance disk system for Historic data. A separate 4TB high performance disk system for Event Journal data. Server class hardware and 64-bit operating system. |
|
Server with 1,000,000 objects, 500,000 historic points stored for 2 years online, multiple clients |
2 x octa-core server class processor, 96GB RAM, 100GB free space on high performance disks for operating system, installed applications, database configuration, and so on. A separate 100GB high performance disk system for log files. A separate 8TB high performance disk system for Historic data. A separate 8TB high performance disk system for Event Journal data. Server class hardware and 64-bit operating system. |
NOTE: Quad-core processors have four cores; octa-core processors have eight cores.
There are many factors that can affect disk space use, including the amount of historic data your system generates and the overheads and mechanics of the file system. In many cases, you can calculate a suitable disk size by following these steps:
- Estimate the approximate number of historic and event records you expect your ClearSCADA to generate over a year. Each historic record is 32 bytes and each Event Journal record is 768 bytes.
- To allow for overheads such as clusters, double the amount you estimated in step 1.
- For efficient disk usage, double the total from step 2.
This should provide you with an approximate value for the amount of disk space your system requires for a period of one year. Using this information, you can determine what size of disk you need (for a year). If you want to replace disk drives more or less often, adjust the amount of time in your calculation in step 1.
With regard to the disk space required to store log files, we recommend that you calculate the amount of disk space needed to store at least 72 hours of logging (ideally longer), regardless of the size of your system. The storage required for log files will vary directly with the logging options selected and the duration of log file retention. We recommend that you review the log file storage requirements regularly during system operation.
For more information on the required desktop and server class operating systems, see Virtualization.
For information on ViewX and WebX client recommendations, see Client PC Requirements.
NOTE: The typical point to object ratio is around 1:2. So for a system with 50,000 points, this equates to around 100,000 objects. However, ratios vary per system and this should only be regarded as an estimate. Some ratios may require lower or higher specification computers, depending on the actual configuration.
Further Information
Disk architectures: Disk Drives and RAID Arrays.
For further information on PC requirements specific to your particular operating system, see the relevant system requirement pages at http://www.microsoft.com
System Performance Guidelines.
For further information about log files, see Logging and Monitoring in the ClearSCADA Guide to Server Administration.