DDNMLog Scheme

The design criteria of the newly NTUH HIS logging system can be defined as: 1) To provide

an application framework for logging website usages with a caching and database approach;

2) Not to interfere with normal NTUH HIS Web operation traffics or performances. In the

DDNM portal, in order to keep the favourite links for the NTUH users, we can log users’

behaviour, DDNMLog, and obtain users popular function links accessed by adapting LRU

(Least Recent Used) algorithm. These links will be collected and implemented as “my

favourite”. Therefore, users can quickly retrieve the links they frequently select. In addition,

the pre-fetched links can be cached in advance to improve navigating performance.

DDNMLog, as depicted in Figure 6, allows NTUH HIS website administrators to record and

analyze clinician usages of HIS online resources. The application includes four components:

Favourite Links Generation, Logging, Queuing & Caching, and Back-end HIS Database.

Favourite Links Generation and Logging components are embedded in NTUH portal. The

other components are integrated in NTUH HIS.NET environment.

In the diagram, the Auth-WS validates user’s authentication, authorization and generates an

access key for the user. DDNMLog recognizes the same individual no matter where he/she

might be located in office, in lab, at home (via NTUH Virtual Private Network), or behind

the NTUH Intranet. To facilitate and expedite the clinician operations, the Favourite Links

Generation Module constructs and extracts the most recently executed function linkages

retrieved (via.Net Remoting technique) from the FavouriteTable, implemented as.NET

DataTable, in Queuing & Caching Module. The Module is implemented as a daemon, i.e.,

Window Service, resided in Portal Server.

The DDNMLog presents a caching and database approach implemented as an embeddable,

plug-in, service, invoked by the medical practitioners. The FavouriteTable is periodically

restored into the HIS Database for synchronization. The table is solely pre-fetched from the

Database by the daemon, i.e., Queuing & Caching Module, after each re-start. Under normal

operations, the daemon re-start rarely happens.

The DDNMLog database schema is described as followings. The MyFavouriteTable, in HIS

Database, consists of 4 attributes: 1) ID (User ID); 2) LinkItemNum (function linkage

number); 3) AccessTime (function linkage access timestamp); 4) Rank (LRU priority based

on 3). The FavouriteTable attributes, in Queuing & Caching Module, are matched with the

MyFavouriteTable attributes correspondingly.

42 Web Intelligence and Intelligent Agents

Fig. 5. An example of dynamic, hierarchical DDNM

A Dynamic Healthcare Portal Design and Enhancements 43