Server Architecture
Overview
The server architecture for a SobekCM library will generally scale according to the size of the library. As
the number of pages (and hence necessary resource files), items, and titles increases the current server load
should be analyzed. As necessary, adjustments can be made to existing servers, or additional servers can be added to split the load.
When considering your server architecture, be aware that there are seven major components to the SobekCM system which should be considered when creating a server architecture:
- The IIS Web Server will run the main SobekCM digital library software and will be publicly accessed by all of your users.
- The database contains basic information needed to run the digital library and also contains all the metadata for all items within the library.
- The builder service application must run on a server to allow bulk loading and to perform some basic scheduled maintenance tasks on the digital library.
- Full text indexes are contained within a Solr/Lucene instance running on Apache and two Solr/Lucene indexes. This is actually optional if you do not plan to load any full-text to your library.
- All of your files must be web accessible and reside on a file server or file system. This scales directly with the number and complexity of the items within your digital library.
- To view zoomable images, some JPEG2000 zoomable server software must also be installed and be web-accessible.
- Finally, you can run an optional caching server if your library becomes very large and handles an extreme amount of traffic.
Despite this apparent complexity, the system can easily run on two servers.
Additional Information
The following documentation should be helpful in determing the architecture that best fits your needs and budgets:
