Management data integration

Effective systems management solutions require the challenging integration of management information coming from a diverse set of sources, e.g., management products, personnel records, accounting systems and the like. Examples of such management challenges are:

• Internet-scale management in cloud computing, managing applications that use services and resources from multiple providers distributed over the cloud;
• Providing "situation-specific", on-demand remote assistance to operators where in-depth expertise is not locally available; and
• Management applications where sources of data are diverse, such as energy-efficient (green) computing where there is a mix of IT and non-IT data.

Standards and initiatives are emerging to facilitate the integration of management information, but they are not effective in keeping up with the increasing diversity and changes in the sources of information as well as the management requirements. If information relevant to management tasks can be extracted easily from the cloud of information, it will pave the way to building powerful management applications and solutions efficiently.

The goal of the Fusio project is to enable simple and cost-effective creation of integrated management solutions. The project is building a data integration platform for rapidly building custom management applications with the following benefits:

• Loose coupling of data across diverse data sources
• Flexible, lightweight, and fast integration
• Discovery of new relationships between data
• Enabling of new types of management solutions
• “Google-like” (but focused) search
• Context-aware applications
• Internet-style, ad-hoc data mash-ups

The ability to pull together information from any relevant data source, e.g., management products, calendars, databases, directories etc., and to extract and establish relationships between data within and across sources makes it possible to build management solutions with different perspectives and capabilities in a quick and easy fashion.

The architecture of Fusio is shown in Figure 1.

In many cases, management data sources do not share the same or compatible data model, nor do they employ the same methods to access data. Fusio therefore assumes a heterogeneous environment: in response to a query, it uses source specific methods to collect data, converts collected data into a semantic Web representation with little or no manipulation of the original data, and uses rule-based inferencing to refine and relate data from multiple sources. Text-based search and a Fusio data-model are used to focus and simplify this process.

Fusio does not introduce an additional central repository of management information, as such a store would potentially be huge and would not be suitable to retain the needed transient data, e.g., CPU load right now. Therefore all information in Fusio is fetched on demand in a selective fashion driven by the user query, using partial caching for efficiency.

Fusio is based on semantic Web, linked data and text-based search technologies.