Querying and Online Analytical Processing of Large-scale Information Networks

Monday, Feburary 13, 2012
4:00 p.m.
Davis Marksbury Theater
Refreshments Served 3:30p.m.

Dr. Peixiang Zhao, University of Illinois at Urbana-Champaign

Querying and Online Analytical Processing of Large-scale Information Networks

Abstract:
Social and technical information systems usually consist of a large number of interacting physical, conceptual, and human/societal entities. Such individual entities are interconnected to form large and sophisticated networks, which, without loss of generality, are often referred to as information networks. Noteworthy examples of information networks include communication networks, research collaboration networks, biological networks, social networks and ultimately, the Web. Information networks are ubiquitous and have formed a critical component of modern information infrastructure.

In this talk, I will discuss a series of fundamental queries of practical value arising in real world information networks, and explore the corresponding effective and scalable query processing solutions in the case of large-scale information networks. I will be mainly focused on two pieces of my recent work. SPath is a high performance graph indexingmechanism to address the general sub-graph queries in large information networks. Graph Cube is the first of its kind warehousing and online analytical processing (OLAP) tool for multidimensional information networks. In Graph Cube, we jointly considered both the multidimensional metadata together with the graph structure of information networks for query processing, thus generating a set of structure-enriched and concisely summarized aggregate networks as query results. It has been widely believed that the marriage of information network analytics and query processing technology will bring many exciting opportunities for future study, and I will briefly mention some of the promising research topics within and beyond this direction as well in the talk.