SearchBlox Replication
The Replication feature in SearchBlox (available in all ENTERPRISE Editions) keeps search indexes in multiple instances of SearchBlox across multiple physical machines synchronized. The SearchBlox instance that makes the indexes available for replication is called the Indexing Server and the one or more instances of SearchBlox that get the search indexes from the Indexing Server are called the Search Servers. The Indexing Server and the Search Servers use HTTP to communicate with each other. When changes are available on the Indexing Server, the Search Server downloads (Pull Replication) the changes, again over HTTP. Using SearchBlox Replication, a high-availability search architecture can be quickly and easily built as shown below:
Benefits
Full Server Replication: With SearchBlox Replication, it is just not the search indexes that are kept synchronized. Any stopwords lists, stylesheets and server settings updated on the Indexing Server are also automatically downloaded by the Search Servers.
Seamless Index Reloading: When the Search Server receives an updated version of a search index from the Indexing Server, it first loads the new version of the search index. With the old and new version of the search index running simultaneously, the Search Server switches search request handling from the old search index to the new search index. After the switchover, the old version of the search index is deleted from the Search Server. This mechanism ensures that search requests are continually proccessed during the transition from the old to the new version of the search index.
Automated Operation: SearchBlox Search Servers automatically check with the Indexing Server for changes. The time interval between these checks can be configured for fine-grained control. When changes are available on the Indexing Server, the Search Server automatically downloads these updates over HTTP and loads them.
High Scalability: Using replication, search applications using SearchBlox can be scaled horizontally. High scalability is achieved by merely adding more Search Servers to the architecture.
Simple Configuration: Enabling replication in SearchBlox is simple and only requires configuring details of the Indexing Server in the Search Servers. As all communication and data downloads are over HTTP, no special network configurations are required. Replication can be instantly disabled in any Search Server and the Search Server can then act as an Indexing Server.