Giselle

Technical Background
Nowadays, services that provide geographical information over the web are commonplace. These services allow a user to submit a query and the sought data is then returned in the form of a map that is marked with features of interest. Some services also let the user do various analysis on the resulting map, such as:

• Pick two points and measure the distance between them.
• Draw a polygon and measure its area.
• Show/hide layers with information about various features, drawn atop of the map.

However, the objectives addressed by Giselle are much more ambitious in that they also allow the user to edit GIS information apart from only viewing it. Such editing accounts for the following circumstances:

• Concurrent use. Several users can edit the same GIS objects simultaneously. Such attempts must be detected, and reasonably handled using an appropriate locking/merging strategy.
• Modification auditing. It is desirable that modifications to the GIS database are recorded so as to allow rolling-back to a previous version, as well as to determine who changed what.
• Constraint enforcement. Editing is subjected to constraints. For example, some polygon edges should not be editable (e.g., parcel borders).
• Bandwidth consumption. The editing is done over the web, potentially using a low-bandwidth connection. Thus, the amount of network traffic must be kept minimal.

Technologies used in Giselle

The technologies involved with Giselle in meeting the objectives stated above fall in the following categories:

Persistence. GIS data can be stored either in spatial databases or in files. Spatial databases are applicable where large amounts of data are to be stored, and where efficient queries, enabled by database's spatial indexing capabilities, are required. Today, several spatial databases exist, which conform to the OpenGIS Simple Features for SQL specification. Among them are Oracle Spatial, MySQL 4.1  and PostgreSQL.

Files are useful for exporting limited amounts of data (e.g., all parcels owned by a particular user) on a medium such as the user’s hard disk or a CD. Standardized file formats exist, for example Geography Markup Language (GML) and ESRI Shapefile.

Bitmap data, such as aerial and satellite images, are stored as standard image files. When high quality and fast manipulation of data is required (e.g., extraction of a portion of an image), formats such as Tagged Image File Format (TIFF) and Portable Network Graphics (PNG) are used. In cases where small footprint is of prime importance, lossy compression techniques are employed (e.g., JPEG and JPEG2000). Bitmap files are typically associated with meta-data that specifies what is the content of the bitmap file (geographical location, spectrum, scale, etc.).

Transport. GIS data has to be transferred from where it is persisted to where it is needed for analysis or display. One way to achieve this is to accompany the application with the data stored in files. Frequently, however, the data are too large or change too frequently, and centralized storage is desired. In such scenarios, the data can be made available through files on a networked file system, or with specialized GIS servers. Standardized protocols for communication with the server exist (e.g., the OpenGIS series of Web Service specifications).

Scalability and Load Balancing. When many concurrent users request services, the software must allow to scale the underlying network topology and hardware to a large number of simultaneous interactions. Giselle libraries provide for load balancing techniques that distribute network load over several server computers.

Presentation. A common method of distributing GIS data is via web applications, where the users select the area and layers they are interested in, and their browser requests the appropriate image from the web server for display. Also, software libraries exist that can be embedded in more complex applications that run not only on desktop computers, but also on handheld ones (e.g., Microsoft MapPoint,).

There are already many solutions that allow editing of GIS data, for example the ESRI ArcEditor applications, and DeLORME XMap/GIS Editor. These applications are highly suitable for controlled environments, where the choice of operating system, network connectivity, and user skill match those for which the software was built. They are not, however, useful for distributed deployment as a service to a wide community of users, which is the main advantage of Giselle.

Front page | Technical Background | Deployment Requirements | Giselle Architecture | Interoperability

For more details, contact :