William T. O'Connell

Optimizer and parallel engine extensions for handling expensive methods based on large objects

Object-relational database systems allow users to define new types and functions. This typing extension allows structural semantics to be applied to new data types (non-SQL92 data types). However, defining new, large, complex objects within a tuple is not always practical. Practicality alone requires that large object data structures be used. This implies that predicate evaluation can be performed on tuple columns that are physically stored outside of a tuples physical record. This out-of-line storage mechanism is commonly used today for column values that are physically large (implemented with a large object descriptor in the tuple pointing to a large object data structure). This type of predicate-based evaluation presents new optimizer and run-time challenges to the database system on clustered and massively parallel processor shared-nothing architectures.

Trends in data warehousing: a practitioner's view

This talk will present emerging data warehousing reference architectures, and focus on trends and directions that are shaping these enterprise installations. Implications will be highlighted, including both of new and old technology. Stack seamless integration is also pivotal to success, which also has significant implications on things such as Metadata.

Where is business intelligence taking today's database systems?

Academia has provided very performant and storage efficient technologies for fundamental Business Intelligence (BI) objects: cubes, instigated research in stream technologies resulting in renewed interest in continues and temporal queries, supplied further data mining and data exploration algorithms and research query optimizations for complex queries with variety of histograms. Database industry either incorporated into their SQL engines some of these algorithms, or tried to integrate better stand alone BI engines such as online analytical processing (OLAP), or provided their own unique solutions for BI. The innovation in BI technologies within the database offerings made business community apply relational engines to their problems. This application provided valuable feedback on performance, functionality, manageability, and integration of BI features in the Relational Database Management Systems (RDBMs). Consequently, it gave a raise to new trends in BI technologies. As a result, these new trends and issues are quickly emerging as they are being driven by the continued acceptance of the Intranet for business infrastructures.

Extreme streaming: business optimization driving algorithmic challenges

Organizations are striving for competitive advantage. As a result, business optimization is being pushed to new heights in terms of volume and speed. Areas such as customer profitability, campaign profitability or customer insight for better service are driving new analytical challenges as well as new algorithms over large volumes of data. Using Telecom as an example, this talk will discuss business demands which are driving an evolution of analytics over extremely high volumes of streaming data being ingested into a warehouse -- this business direction is forcing algorithms to evolve. In this environment, we are forced to deal with (1) massive continuous ingest rates, data changes/updates while the same data is being consumed by applications, users and processes, (2) dealing with scalability without disruption, and (3) overcoming the physical limits of the infrastructure (e.g., IO). This talk will address both how these technical challenges are being addressed at a high-level, as well as known problem areas for further research. This talk will also highlight issues of new algorithmic approaches arising from the business needs and the research issues associated with them. Examples include (1) streaming analytics, (2) social analytical algorithms integrated into data mining approaches within these continuously streamed environments and (3) new data types such as voice and textual analytics for customer calling pattern and churn analysis.