Historically, multimedia systems and databases emerged as separate disciplines addressing the needs of different application domains. However, as these application domains matured over the years, the scientific disciplines also moved closer. For instance, on the media management side, although MPEG1, MPEG2, and MPEG4 standards were primarily focused on video compression and decompression tasks, researchers concentrated on the video-content description and indexing issues while developing the MPEG7 standard (which is currently in the works). On the data management side, commercial database management systems that were once primarily targeting traditional business applications, are today including media and Web engines within their packages. New applications, such as digital libraries and dynamic Web content, necessitates a common understanding of both of these disciplines.

As a result, over the recent years, there has been a flurry of activities in the so called area of multimedia databases. These activities resulted in many techniques useful in extracting various features of media and providing means for indexing/querying based on these features. Despite these activities, however, multimedia and databases are never truly integrated. The activities on the media side was for a long time limited to feature extraction and only recently (through MPEG7 and related activities) is moving towards incorporating media semantics. Most existing media storage/retrieval systems, on the other hand, are mostly single-media based and they are still using traditional query optimization and processing techniques, ignoring the imperfections and subjectivity inherent in these objects.

Therefore, I use my expertise in these two disciplines, multimedia systems and data and information management systems, to move to the next level of multimedia-database research and education, where the database management systems are not only dealing with media data, but are in fact designed with multimedia semantics and requirements in mind. In short, my work addresses the data management needs of new application domains, most notably digital libraries and increasingly dynamic content of the Web.

Quality Quorums
Conventional methods for replicating homogeneous data use quorum systems. Traditional quorum systems assume that the data is homogeneous, the access cost to servers is uniform, and the quality of the data is fixed. Quality quorums handle various aspects of heterogeneous replication. More....

Query Optimization
Multimedia and web database queries have different characteristics from queries in traditional databases. For instance, due to the use of thresholding and top-k predicates and sub-queries the number of query results can also change depending on the query execution order. Furthermore, finding exact matches is not required. Due to these inherent differences, we develop novel query processing mechanisms. More....

Integration of Database and Internet Technologies
The content of many web sites change frequently. Web site performance, including system up-time and user response time, is a key differentiation point among companies that are eager to reach, attract, and keep customers. Therefore, we develop solutions for integrating Internet services, business logic, and database technologies, and for improving end-to-end scalability of e-commerce systems. More....

Indexing Multimedia Documents based on Time
The knowledge about the dynamic properties of multimedia documents, not only enables users to query and retrieve them, but many essential functionalities, such as (1) object prefetching for interactive document visualization, (2) result summarization/visualization, and (3) query processing for document retrieval, depend on its

• efficiency in extracting and representing dynamic information,
  
• speed in comparing two documents using this information, and
  
• capability of providing a meaningful similarity value as a result of the comparison.

In most non-trivial models, dynamic properties are declared as relationships between the media objects. Therefore, a general model must enable us to compare two documents based on the declared intentions of the document authors. More....