DONKO DONJERKOVIC

Phone: (650)-654-7661

Mail: donko2 AT yahoo DOT com

http://www.cs.wisc.edu/~donjerko

AREAS OF EXPERTISE

Business intelligence server internals and architecture, online analytical processing (OLAP), analytics on relational systems (ROLAP), database query optimization and evaluation, database statistics and data integration.

EDUCATION

Ph.D. in Computer Sciences, University of Wisconsin - Madison, August 2000. GPA: 3.96

Thesis: Techniques for Partial Query Evaluation

Advisor: Prof. Raghu Ramakrishnan

M.S. in Computer Sciences, University of Wisconsin - Madison, May 1996

M.S. in Physics, University of Wisconsin - Madison, May 1995. (GPA: 4.0)

B.S. in Physics, University of Zagreb, Croatia, July 1992. (GPA: 3.9)

EXPERIENCE

  1. Principal Member of Technical Staff, Oracle  (06/01/2006 - Present).
    Engineering Lead / Architect for Aggregate Navigator, a central module in Oracle BI server. The task of the Navigator is to rewrite high level multidimensional queries into standard SQL using the most optimal set of materialized views.     Designed and implemented  following features of Oracle BI Server version 11:
  1. Parent-child hierarchies with API functions such as: IsAncestor, IsLeaf, and IsRoot.
  2. First and Last time period computations; useful for inventory style measures (API: LastPeriod).
  3. Moving average computations over time dimension (API: PeriodRolling).
  4. A framework that allows a transactional store and a data warehouse to share the same unified data model. This feature enables dynamic switching between real time and traditional analytics.
  5. Reduced Oracle BI server startup time from more than 10 minutes to 1 minute (on very large models).

  1. Lead Engineer,  Siebel Systems, Acquired by Oracle, (2/16/2005 -06/01/2006).
    Designed and implemented following features of Siebel Analytics (Oracle BI) Server version 10:
  1. Per-measure predicates in the form of: Filter (⟨measure⟩ By <predicate>)
  2. Time-shifted computation of measures: Ago (<measure>, <time step>, <#steps>)
  3. Computation of measures up to the current time member: ToDate (<measure>, <time grain>)

  1. Senior Member of Technical Staff, Oracle (6/17/2002 - 2/15/2005).
    Researched, designed, and implemented following features in Oracle 10g SQL execution engine:
  1. Right-outer, right-semi, and right-anti join algorithms.
  2. Several improvements to hash join execution code that resulted in significant performance gain on TPC-D database performance benchmark.
  3. Caching mechanism for deterministic PL/SQL functions.

  1. Senior Software Engineer, Desana Systems (8/21/2000 - 3/15/2002).
    Project lead for content processing of XML and HTTP data. Designed, implemented, and evaluated algorithms for fast XPath rule evaluation on streaming XML data.

  1. Research Assistant, Computer Sciences, University of Wisconsin - Madison (6/95 - 8/00).
    Partial querying is common in decision support systems as a means of exploring data. Two typical partial queries are: Top N, in which users ask for a subset of ordered answers, and First N, in which users do not have any preference.
    To effectively evaluate Top N queries, I developed a probabilistic query optimization framework in which predicate selectivities (and cardinalities of intermediate results) are modeled as random variables. In addition to being useful for Top N queries, this framework can be applied to query optimization with other resources, such as main memory and network bandwidth.
    When processing queries for First N answers, I achieved significant speed improvements by better memory utilization. Database statistics are instrumental in choosing what tuples should be kept in memory.

    In the course of my research I also designed and implemented a columnar SQL processor called DTE. It provided a unified SQL interface across disparate data sources such as file systems, web, and relational databases.

  1. Member of Technical Staff / Summer Intern, Database Principles Research, Bell Laboratories, (5/99 - 8/99).
    Worked on Aqua, an approximate query answering system from Bell Labs. Co-developed a new form of database statistics called Quantized Samples, which has higher precision than traditional samples, given the same space.

  1. Teaching Assistant, Physics, University of Wisconsin - Madison (1/95 - 6/95).
    Taught college physics for undergraduate engineering students.  

REFEREED CONFERENCE PUBLICATIONS

  1. Donko Donjerkovic, Yannis E. Ioannidis, Raghu Ramakrishnan. Dynamic Histograms: Capturing Evolving Data Sets, Proceedings of the 16th International Conference on Data Engineering, San Diego, California, USA, February 2000.
  2. Donko Donjerkovic, Raghu Ramakrishnan. Probabilistic Optimization of Top N Queries, Proceedings of 25th International Conference on Very Large Databases, Edinburgh - Scotland, September 1999.
  3. Raghu Ramakrishnan, Donko Donjerkovic, Arvind Ranganathan, Kevin S. Beyer, Muralidhar Krishnaprasad. SRQL: Sorted Relational Query Language, Proceedings of 10th International Conference on Scientific and Statistical Database Management, Capri, Italy, July, 1998.
  4. Miron Livny, Raghu Ramakrishnan, Kevin S. Beyer, Guangshun Chen, Donko Donjerkovic, Shilpa Lawande, Jussi Myllymaki, R. Kent Wenger.DEVise: Integrated Querying and Visual Exploration of Large Datasets, Proceedings of ACM SIGMOD International Conference on Management of Data, Tucson, Arizona, USA, May 1997.

TECHNICAL REPORTS

  1. Donko Donjerkovic, Swarup Acharya, Phillip Gibbons, Viswanath Poosala. Quantized Samples: Improving Accuracy by Blurring Values, Bell-Labs Technical Report, August 1999.
  2. Donko Donjerkovic, Yannis Ioannidis, Raghu Ramakrishnan. Dynamic Histograms: Capturing Evolving Datasets, UW-Madison Technical Report 1396, March 1999.
  3. Donko Donjerkovic, Raghu Ramakrishnan. Histogram Guided Partial Query Evaluation, UW-Madison Technical Report 1419, August 2000.

PATENTS

  1. Bolsius Roger, Venkatasubramanian Raghuram, Ni Ling, Donjerkovic Donko, Chowdhury Saugata. Systems and Methods for Providing Value Hierarchies, Ragged Hierarchies and Skip-Level Hierarchies in a Business Intelligence Server. United States Patent Application 20110295836, December, 2011

SKILLS

  1. Languages: C++ / C, Java, SQL, MDX, XML / XPath, PHP, Perl
  2. APIs: ODBC, JDBC
  3. Protocols: HTTP, TCP/IP