In a Hypermedia system a link server is responsible for the management of hypermedia information, such as links and anchors associated with each resource. A link server is a server that has both a set of clients and a link database for storing and linking hypermedia data items.
In the world of hypermedia, a link is an item which relates to another, typically in a document or other graphical form. This can be a single document or a sequence of documents. HyTime (the standard SGML format for hypermedia links) is a well established technology which allows this relationship to be expressed in a simple, readable way and stored together in a link database.
There are a number of different link databases, which can be selected by the end user. These are known as contexts and each context has a specific selection of links. This selection can be customised to suit the end users particular needs, and there is an interface utility which allows the end user to carry out a range of linking activities from one place.
When a user selects data of interest, Link Server International they can choose from a set of link functions (start link, follow link, show link etc). This information is encoded into an HTTP message which is sent to the link server for processing.
Once the request has been received, the link server is able to return the relevant information from its context database, or to compile the links into an HTML document which can then be delivered as a web page. The link service can also be used to hard wire the links into a document in a chosen format, if the user has chosen this option from the interface utility menu.
The link service is implemented by an interface agent which runs on a local computer, and is able to respond to the various linking functions by capturing the data being selected, if possible, and parcelling this into an HTTP request for processing at the link server. It is important to remember that this process is asynchronous, and it is necessary to be careful not to leave the interface agent running for extended periods.
It is essential that a link is secured and that there are no unauthorised users accessing the database linked to the other database, as these could lead to significant security issues. These vulnerabilities can be mitigated by ensuring that the appropriate authentication and authorization mechanisms are in place, and that there is a robust backup plan in place for failover scenarios.
For SQL Server based links, the database name should be a four-part name containing the product name of the OLE DB provider and the name of the SQL Server instance. Using an OLE DB provider requires that the account under which the SQL Server service is running has read and execute permissions for the directory and all subdirectories in which the provider is installed.
Similarly, the connection string for the link is required to contain an authenticating user name and password. These may be either a global user or a fixed user, and they must be a user on both the source and destination databases of the link.
Comments
Post a Comment