Transcript No Slide Title - Microsoft Research

Laguna Beech
Extract of the important and unimportant topics
Bernstein
DeWitt
Gawlick
Gray
Maier
Schek
Stonebraker
Hans-J. Schek,
Institute of Information Systems
ETH Zürich and UMIT Innsbruck
Bernstein
Pro
• Distributed system administration (hundreds of servers,
millions of clients)
• TP application schemas (e.g. versioning, compensation in
nested transactions)
• Automatic data translation
• Active databases
Contra
• Database machines
• Real extensible database systems
DeWitt
Pro
• DBMS for managing scientific applications
• CASE support by DBMS
• Optimisation of queries over complex object hierarchies
• Active DBMS – whatever that means
Contra
• Processing general recursive queries
• Hardware sorters
• Hardware filters
• Anything related to concurrency control
• Object-oriented DBMS that mention encapsulation – no
parallization
Gawlick
Pro
• Productivity and operations
• Technology for transaction processing
• Interdisciplinary communication
• Access patterns
Gray
Pro
• Procedures in database systems
• Automatic DB physical design
• Disaster recovery – data & application replicated
• Ten years continuous operation
• Large or exotic databases – 1012 records, video, fax,
sound
• Specialty databases – case, cim, geo,…
Maier
Pro
• Single-object constraints
• Physical representation language
• Update semantics for logic DBs
• Constructive type theory
Contra
• Storing DML as strings in DB – not compositional
• Behavior-only object models
Schek
Pro
• Systematics on semantic data models, knowledge
representation, complex objects, nested relations, power
sets, fixpoints,...
• Optimisation, mapping to kernel operations
• Host language coupling with external objects
• Tight DBMS cooperation with applications
Contra
• More recursive query processing – missing applications
Stonebraker
Pro
• Integration of 4GLs programming languages and DBMS
• 1000 node distributed DBMS
• Abolition of IR systems as one-offs – efficient text in
general purpose DBMS
• End-user usable application development environment
Contra
• Recursive query processing
• Interface between prolog and dbms