- Enterprise Search
- Knowledge Management
- Social Intranet
- Content Analytics
- Further Application
- Best Practice
Reliability and Scalability
An important point with the conversion of global searching solutions in big organizations is the consideration of the edge terms given by the available IT infrastructure. These are stamped by the organizational structure, business processes, role and rights concepts as well as of the range of the network connection between different locations.
In addition, the number of systems included in the search and the number of indexed iinformation objects plays a crucial role with the specification of the concrete implementation strategy for a search solution.
Consider in particular the following factors:
- Total quantity and growth rates of the to be indexed information objects
- Number of users at the same time accessing the search solution
- Bandwidths between possible different locations to be indexed systems and distribution of user about the locations (as a rule a mainly local indexation of the systems) is derived from it)
- Data source-covering access and right structures as well user’s groups
- Little adminstration of updates with a defined time slot for an if necessary necessary new indexation
In the result of the evaluation of these marginal conditions a concrete implementation concept originates including the topology of the search and the hardware resources required for it.
Besides, several factors can lead to the fact that a cluster search solution must be realised.
Need of a Cluster Search
A cluster search can (like any cluster solution) have different objectives:
- Reliability through redundancy
- Scalability: Control of the the amount of data through parallelization (when indexing and / or search)
- Increase of access speed (Near-Real-Time-Indexing and / or better response time behaviour)
- Supply of a homogeneous system access for a complicated overall system existing of several components
The supply of efficient servers has a positive influence on the minimisation of the number of the indexation servers and/or search servers.
Often several integrator instances are required by the topology of the organization or the amount of data.
The cluster search makes sure with the fact that about a homogeneous user interface all search sources are bundled up again.