Group Title: Department of Computer and Information Science and Engineering Technical Reports
Title: Database research at the University of Florida
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00095327/00001
 Material Information
Title: Database research at the University of Florida
Series Title: Department of Computer and Information Science and Engineering Technical Reports
Physical Description: Book
Language: English
Creator: Su, Stanley Y. W.
Lam, H.
Sharma, M.
Kamel, N.
Publisher: Database Systems Research and Development Center, University of Florida
Department of Computer and Information Science, University of Florida
Place of Publication: Gainesville, Fla.
Gainesville, Fla.
Copyright Date: 1995
 Record Information
Bibliographic ID: UF00095327
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.

Downloads

This item has the following downloads:

PDF ( PDF )


Full Text








Database Research at the University of Florida


Database Systems Research and Development Center
CIS Department and EE Department
University of Florida, Gainesville, FL 32611
Email: {su, hlam, sharma, kamel, hanson}@cis.ufl.edu


1 Introduction
This report gives a brief description of the current
database research activities in the Database Systems
Research and Development Center of the University of
Florida. For each project, we describe the objective
and approach, the current status, the source of support
and investigatorss, and the recent publications of
the project. Interested readers should contact the
investigators for further information and reprints. Most
of the reports are available via anonymous ftp to
ftp.cis.ufl.edu under /cis/tech-reports.

2 OSAM*.KBMS Project
2.1 Objective and Approach
This project deals with the design and implementation
of an active object-oriented knowledge base manage-
ment called O',.\\I'.KBMS for use in advanced appli-
cations. The system is founded on an object-oriented
semantic association model O'.\\I' which allows the
structural abstractions of any object class to be de-
fined in terms of its various types of associations with
other object classes, the behavioral abstraction in terms
of system- and user-defined operations, and knowledge
abstraction in terms of knowledge rules with triggers.
Multiple inheritance of structural properties, opera-
tions and knowledge rules is supported by the system.
To ease users' tasks in accessing and manipulating a
knowledge base, a high-level, non-procedural, object-
oriented query language OQL has been designed and
implemented. It is an "association- or patte in-I-. I
query language instead of a more traditional "attribute-
.-..I language. It allows data search conditions of
different types of complexity to be specified as patterns
(linear, tree and network structures) of object class asso-
ciations instead of comparing values of keys and foreign
keys. Query processing and optimization is based on the
mathematical properties of a set of algebraic operators
of an association algebra called A-algebra. Methods as-
sociated with object classes are written in a knowledge
base programming language called K. A set of graphic
tools are provided for the definition, editing, browsing,
and querying of knowledge bases. The above features


and facilities offer a desirable environment for object-
oriented design and development.

2.2 Current Status
Several versions of the system described above have
been implemented and demonstrated to several orga-
nizations. The latest version of O'.\ \I '.KBMS demon-
strated at SIG \IOD '93 contains the following key com-
ponents: a set of graphic tools, a transaction manager,
a query processor, a trigger and rule processor, and an
object manager implemented on top of ONTOS which
provides the low-level storage management supports.
The implemented version of K is called K.1. Programs
written in K.1 are first compiled into C++ code and
translated by the C++ compiler into executable code.
Presently, the KBMS runs on a network of SUN work-
stations and IBM RS6000 workstations. It is a single
user system without recovery and concurrency control
features.
Support and Investigators: The project was ini-
tiated in 1988 with the support of NSF during 1988-
90. It subsequently receive support from the US West
Advanced Technologies, the Florida High Technology
and Industry Council and presently IBM for different
parts of the project. The PI of this project is Stan-
ley Su (-' .. .- ..11 i 1..) and the CO-PI is Herman Lam


[1] Alashqur, A.M., Su, S.Y.W., and Lam, H., "OQL:
A Query Language for Manipulating Object-oriented
Databases," Proc. of the 15th I,, I Conf. on Very Large
Data Bases, 1989, pp. 433-442.
[2] Guo, Mingsen, Su, S.Y.W., and Lam, H., "An Asso-
ciation Algebra for Processing Object-oriented Databases,"
Proc. of the 7th IEEE I,, / Conf. on Data Engineering,
1991; received Best Paper Award, pp. 23-32.
[3] Kamel, N. Wu, P., Su, S. Y. W., "A Pattern-
based Object Calculus," paper submitted in 1992 and
accepted for publication in the VLDB Journal.
[4] Lam, Herman, Su, S.Y.W., Ruhela, V., Pant, S., Ju,
S.M., Sharma, M., and Prasad, N., "GTOOLS: An Ac-
tive GUI Toolset for an Object-oriented KBMS," It I









Journal of Computer System Science and Engineering,
Vol. 7, No. 2, April 1992, pp. I.'i--"
[5] Su, S.Y.W., Krishnamurthy, V., and Lam, H., "An
Object-oriented Semantic Association Model (O',.\ \')
for Modeling CAD/CAM Databases," Chapter 17 in Ar-
tificial Intelligence: Manufacturing Theory and Prac-
tice, S.T. Kumara, A.L. Soyster, and R.L. Kashyap,
(eds.), Institute of Industrial Engineers, Industrial En-
gineering and Management Press, Norcross, GA., 1989,
pp. 463-494.
[6] Su, S.Y.W. and Lam, H., "An Object-oriented
Knowledge Base Management System for Supporting
Advanced Applications," Proc. of the 4th 1,, / Hong
Kong Computer Society Database Workshop, 1992.
[7] Su, Stanley Y. W., Guo, Mingsen, and Lam, Herman,
"Association Algebra: A Mathematical Foundation
for Object-oriented Databases," to appear in IEEE
Transactions on Knowledge and Data Engineering, Oct.
1993.
[8] Yaseen, Rahim, Su, Stanley Y. W., and Lam,
Herman, "An Extensible Kernel Object Management
System," Proc. of the Conference on Object-Oriented
Programming Systems, Languages, and Applications
(OOPSLA '91), Phoenix, AZ., October 6-11, 1991, pp.
247-263.

3 Parallel Architectures and
Algorithms for Object-oriented
Knowledge Base Management
3.1 Objective and Approach
This project aims to study the parallel architectures
and algorithms for efficient processing of large object-
oriented knowledge bases. The approach taken in this
study is to represent a large knowledge base as a schema
graph at the intensional level and as object graphs at
the extensional level. Graph-based models are then used
for 00 query processing and optimization, knowledge
rule processing, and transaction management. A graph-
based query processing model supports two parallel
multiple wavefront algorithms, namely identification
and elimination algorithms. Query processing is divided
into two phases. In the first phase, objects of interest
or of no interest in the knowledge base are identified
or eliminated depending on the algorithms used. In
the second phase, system/user defined functions are
executed on the objects selected during the first phase to
produce the final result. In our research, the processing
and optimization of multiple queries is being studied
and heuristic rules are applied. In the graph-based rule
processing model, a control graph models the variety of
structural and semantic relationships and dependencies
among rules. This provides a mechanism to process the


triggered rules according to the application semantics
and also clearly defines the scope of parallelism that
can be exploited among rules. Compatible to the
control graphs of rules, we also use a control graph
representation for transactions. This representation
of a transaction is more general and powerful than
traditional flat and tree-based transaction models and
can uniformly embed the rule control graphs as part of
a transaction.
3.2 Current Status
We have presented and proven the correctness of mul-
tiple wavefront algorithms for both tree-structured and
cyclic queries. A unique data structure for support-
ing multiple wavefront algorithms has been designed
and implemented. A parallel query processor has been
implemented on a multiprocessor Transputer network.
Based on the experience, we built a query processing
evaluation system, which can process multiple queries,
on a 64-node nCUBE 2 parallel computer. We have im-
plemented a graph-based parallel scheduling algorithm
on the nCUBE2 parallel computer. Other supporting
modules of the OOKBMS (query processor, transaction
processor, lock manager, etc.) are simulated to create
an environment for evaluating architectures and algo-
rithms.
Support and Investigators: This project was ini-
tially supported by NSF with Stanley Su as the PI and
Herman Lam as the Co-PI. It is presently supported by
Fujitsu, Ltd., with Stanley Su as the PI.
[1] Lee, C., Lam, H. and Su, S. Y. W., "An Object
Flow Computer for Database Applications: Design
and Performance Evaluation," Journal of Parallel and
Distributed Computing, 17, 1993, pp. 298-314.
[2] Su, S. Y. W., Chen, Y. H., and Lam H., \IIII,.I.-
Wavefront Algorithms for Pattern-based Processing
of Object-oriented Databases," Proc. of First I,, I
Conf. on Parallel and Distributed Information Systems
(PDIS), 1991.
[3] Thakore, A. K., Su, S.Y.W, Lam, H., and Shea,
D.G., "Asynchronous Parallel Processing of Object
Bases using Multiple Wavefronts," Proc. of the It I
Conf. on Parallel Processing, Aug 1990, pp. 127-135,
[4] Thakore, A. K., Su, S. Y. W., and Lam, H., "Algo-
rithms for Asynchronous Parallel Processing of Object-
oriented Databases," to appear in IEEE Transactions
on Knowledge and Data Engineering.

4 KBMS-based Evolutionary
Prototyping of Software Systems
4.1 Objective and Approach
The objective of this project is to develop a high-level
prototyping language and a computing environment for









supporting software system development. The project
takes a knowledge-base modeling approach to system
development. In this approach, a prototyper would
first model a system he/she wants to build based
whatever knowledge he/she has about the system. The
model of the prototype would consist of object classes
which define the structural, behavioral and knowledge
abstractions of the system components, their structural
relationships, and the data these components process.
Methods associated with these classes are either real
programs written in the prototyping language, or
simulated code which takes some input and generates
some legal output (by table lookup, for example), or
modeled by a control structure of program segments,
each of which is again modeled by an object class. The
control structure which models a method can be used
by the system to automatically generate executable code
once the lowest level of program segments are written
in real or simulated code. As the prototyper gains
more knowledge about the system, the model of the
prototype is refined and decomposed into more detailed
components. At each stage of modeling, the model of
the prototype is executable to verify its functionality
and performance. The final system is the last version of
the evolved and tested prototype.

4.2 Current Status
We are extending the language K to include associa-
tion types for modeling methods and the structural re-
lationships of software components. The concept of time
and the management of versions are important for sys-
tem development. They are being incorporated into the
data model and the language. O'.\ \I'.KBMS and its
graphic tools are being extended to model and test the
executable models of software systems.
Support and Investigator: This is a three-year
project funded by NSF (1992-4). The PI is Stanley Su.
[1] Chen, H. H., and Su, S. Y. W., "TA-Algebra: A
Temporal Association Algebra for Processing Temporal
Object-oriented Knowledge Bases," paper in prepara-
tion.
[2] Shyy, Y. M. and Su, S. Y. W., "K: A High-level
Knowledge Base Programming Language for Advanced
Database applications," AC if SIC ilOD '91, Denver,
CO., May 29-31, 1991, pp. 338-347.
[3] Shyy, Y. M., Arroyo, J., Su, S. Y. W., Lam, H., "The
Design and Implementation of K: A High-level Knowl-
edge Base Programming Language of Os.\ \I '.KBMS,"
submitted for publication, 1992.
[4] Su, S. Y. W. and Shyy, Y. M., "A Universal Knowl-
edge Base Modeling Approach to Evolutionary Proto-
typing of Software Systems," submitted for publication,
1992.


[5] Su, S. Y. W. and Shyy, Y. M., "An Object-oriented
Knowledge Model for KBMS-supported Evolutionary
Prototyping of Software Systems," to appear in Ad-
vanced Database Concepts and Research Issues, N. R.
Adam and B. Bhargava (eds), Springer-Verlag, 1993.
[6] Su, S. Y. W. and Chen, H. H., "A Temporal
Knowledge Representation Model O.\\ I'/T and Its
Query Language OQL/T," Proc. of I,, / Conf. on
VLDB, Barcelona, Spain, 1991, pp. 431-442.
[7] Su, S. Y. W. and Chen, H. H., "Temporal Rule Speci-
fication and Management in Object-oriented Knowledge
Bases," to appear in the Proc. of the First L,, / Work-
shop on Rules in Database Systems, Edingburgh, 1993.

5 ORECOM Project
5.1 Objective and Approach
The objective of this project is to develop a system to
perform semantics-preserving schema translations in a
multimodel, heterogeneous database environment. The
approach taken in this project is the use of a neutral
data model called ORECOM as an intermediate repre-
sentation through which a source schema is translated
into a target schema. ORECOM provides a small num-
ber of general structural constructs for representing the
structural properties of high-level data models and a
powerful knowledge rule specification language for spec-
ifying semantic constraints and other semantic informa-
tion associated with the structural constructs. The se-
mantic properties of high-level modeling constructs and
constraints are decomposed into ORECOM's neutral
and primitive representations. The decomposed rep-
resentations are then used to precisely determine the
equivalence relationships between constructs and con-
straints of different data models. Furthermore, the dis-
crepancies found in the translation are explicitly speci-
fied by the primitive representations which can be used
for generating an explanation of the discrepancies or
used in the application development to account for the
missing semantics.
5.2 Current Status
The ORECOM schema translation system has been
implemented to support schema translations among
four data models: IDEF-1X, EXPRESS, NIAM, and
Os.\ 1. It is implemented in C running on both
UNIX-based Sun workstations and AIX-based IBM RS-
6000 workstations. Its graphical user interface (GUI)
is implemented in an X-Window environment using
OSF/Motif widget set. The input to the system is
a source schema defined in one of the four models,
which can be a pre-stored schema or an ad hoc schema
created by using a schema editor. When a translation
is completed, the system outputs a target schema
defined in a different data model and a discrepancy list.









ORECOM can also be used for heterogeneous schema
integration and schema re-engineering. A version of the
system has been demonstrated at several organizations
and also at the EXPRESS Users Group Conference -'
Support and Investigators: This project was initi-
ated in 1991 with the support of the National Institute
of Standards and Technology and has been supported
by the Florida High Technology and Industry Council
since 1992. The PI is Stanley Y. W. Su and the Co-PI
is Herman Lam.
[1] Su, Stanley Y. W., F .. S. C. and Lam, H., "An
Object-oriented Rule-based Approach to Data Model
and Schema Translation," CIS Technical Report TR-
92-015, University of Florda, May 1992; Submitted for
publication.
[2] Su, Stanley Y. W. Su and F... S. C., "ORECOM:
A Neutral Data Model for Data Model and Schema
Translations," CIS Technical Report, University of
Florda, May 1992; 1993.

6 Active Databases Sentinel
6.1 Objectives and Approach
Sentinel is a (re)active Object-Oriented DBMS cur-
rently under development using the open OODB Toolkit
from Texas Instruments, Dallas. The short-term objec-
tive is to integrate events and rules into an OODBMS in
a seamless manner. The long term objective [2] is to use
the resulting system: i) as a basis for a self-monitoring
(or adaptive) DBMS, ii) to provide support for cooper-
ative problem solving ( 8), and iii) to support multi-
media active DBMS for scientific applications ( 9).
6.2 Current Status
Snoop [5] is an expressive event specification language.
We have defined an event precisely and distinguished
among an event, an event expression (that has dura-
tion) and a condition. We propose the notion of an
event modifier (e.g., end-of) to map an event expres-
sion into an event. Primitive events are classified into
database, time, and external/abstract events. A num-
ber of event operators (disjunction, sequence, all, aperi-
odic and periodic with cumulative and non-cumulative
variations) have been defined along with the grammar
for constructing complex events. Parameter contexts
are defined to compute complex events and their pa-
rameters. One of the optimization techniques for the
Changes operator proposed in HiPAC was incremental
evaluation. Extended Relational Algebra (ERA) [4] pro-
vides a mathematical basis for evaluating changes to ar-
bitrary, non-aggregate expressions of relational algebra.
Incremental versions for select, project, join, union, and
difference were developed and their correctness shown.
The restriction on the chain rule developed in HiPAC
was relaxed to obtain a generalized chain rule. In [1]


we classify objects into passive, reactive, and notifiable
classes. In contrast to ADAM and Ode, we provide sup-
port for an external monitoring viewpoint which permits
notifiable objects to dynamically subscribe to monitor
changes to reactive objects and take appropriate ac-
tions. Both events and rules are supported as first class
objects. The proposed framework supports event and
rule specification even on the rule class. Our approach
[3] to AI and database integration is to support multiple
production rule systems on a shared database. Towards
this end, we developed a methodology for translating
an OPS5 class of production rule applications into rela-
tions and triggers. Only disjunction of events and mul-
tiple triggers capability for a relation is required for the
above. Work on the concurrency control and recovery
issues is underway to execute OPS5 applications con-
currently on a shared database.
In addition to the above, we are currently: i) imple-
menting a nested transaction model for concurrent rule
evaluation and ii) developing an optimizer for process-
ing queries, deductive, and active rules [6].
Support and Investigators: This project is directed
by Sharma Chakravarthy (-li.-lltII' i.- ufl -dli). This
project is supported by Sofrevia, NSF, and NRad.

[1] E. Anwar, L. Maugis, and S. Chakravarthy. A New
Perspective on Rule Support for Object-Oriented
Databases. In Proc. AC if SIG i\OD, pages 99-108,
1993.

[2] S. Chakravarthy. Architectures and monitoring
techniques for active databases: An evaluation. UF-
CIS TR-92-041. To appear in Applied Data and
Knowledge Engineering Journal, 1993.

[3] S. Chakravarthy and R. Blanco-Mora. Supporting
very large production systems using active dbms ab-
straction. UF-CIS TR-91-25, University of Florida,
Sep. 1991.

[4] S. Chakravarthy and S. Garg. Extended relational
algebra (era): for optimizing situations in active
databases. UF-CIS TR-91-24, University of Florida,
Nov. 1991.

[5] S. Chakravarthy and D. Mishra. Snoop: An
expressive event specification language for ac-
tive databases. UF-CIS-TR-93-007, University of
Florida, March 1993. (Revised and extended ver-
sion of UF-CIS-TR-91-23).

[6] S. Chakravarthy and X. Zhang. An open architec-
ture for optimizing active and deductive rules. UF-
CIS-TR-93-013, University of Florida, April 1993.









7 Real-Time and Temporal Databases
7.1 Objectives and Approach
The objectives of this project are to investigate: time-
constrained real-time transaction processing, ECA rules
having all of the temporal semantics used in temporal
databases, and the preservation of multiple pasts using
one and two-dimensional time notions. For real-time
databases hard, soft and firm deadlines are considered.
7.2 Current Status
In [1], we have clarified the notion of temporal validity
by using interpretation-based definition that supports
multiple characteristic functions for validity. Further,
we have shown the adequacy of three time concepts
(event, transaction, and valid time) for preserving
multiple past state of a temporal database. Finally,
modeling and querying of a temporal database using
one and two-dimensional time representations has been
developed [2]. For real-time transaction processing, a
scheduling algorithm has been synthesized taking both
static and dynamic costs into account. A pre-analysis
of transactions is used. The proposed cost conscious
approach (CCA) has been compared to the EDF-HP
approach [3]. Simulation of the scheduling policy has
shown that it adapts well to the fluctuations in the
system load.
Support and Investigators: This project is directed
by Sharma Chakravarthy. Prof. Ted Johnson is a
collaborator on the real-time database part.

[1] S. Chakravarthy and S.-K. Kim. Semantics of
time-varying information and resolution of time
concepts in temporal databases. In Proc. of the I,, I
Workshop on Infrastructure for temporal Databases,
pages G1-G13, June 1993.

[2] S. Chakravarthy and S.-K. Kim. Temporal
Databases with Two-Dimensional Time: Its In-
terpretation and Implementation. UF-CIS-93-020,
University of Florida, July 1993.

[3] D. Hong, T. Johnson, and S. Chakravarthy. Real-
time transaction scheduling: A cost conscious ap-
proach. In Proc. AC if SIG 1[OD, pages 197-206,
1993.

8 Distributed/Cooperative Databases
8.1 Objective and Approach
For distributed databases, we are developing a collection
of core modules for the distributed database design.
The novel aspect of our approach is that the design
space is partitioned and algorithms developed for
each task with well-defined interfaces so that the
algorithms can be mixed and matched in a number of
ways. We are also investigating the systems support


required for cooperative problem solving. In our view,
recent advances in database technology (e.g., active
and heterogeneous DBMSs) and maturation of other
concepts (e.g., temporal databases) provide us with
techniques and abstractions for formulating a viable
solution to the above problem.
8.2 Current Status
In the work on distributed databases, an objective func-
tion for vertical partitioning that generalizes and sub-
sumes earlier work has been derived [3]. This objec-
tive function is used to compare previously proposed
algorithms for vertical partitioning (such as bond en-
ergy, BVP, and Ra's algorithm). We have shown how
the objective function can be extended to include addi-
tional information, such as transaction types, integrity
constraints, different local and remote accessing costs
and replication. Distribution of data in the presence of
rules is currently being investigated. For the coopera-
tive part, we have identified its key underlying charac-
teristics. We have proposed a database-centered solu-
tion [2] by combining and extending techniques and ab-
stractions to support the characteristics identified. We
have a multi-staged approach that covers both immedi-
ate and long-term solutions for supporting cooperative
problem solving. We have proposed preliminary solu-
tions to capability modeling and activity specification
and management [1].
Support and Investigators: This project is directed
by Sharma Chakravarthy. Profs. Ravi Varadarajan and
Sham Navathe have collaborated on this project. Parts
of this project have been supported by US West, NSF,
and Bellcore.

[1] S. Chakravarthy, K. Karlapalem, S. B. Navathe,
and A. Tanaka. Database supported cooperative
problem solving. UF-CIS TR-92-46, University of
Florida, Dec. 1992. (Submitted for publication).

[2] S. Chakravarthy, S. B. Navathe, A. Tanaka, and
K. Karlapalem. The cooperative problem solving
approach: A database-centered approach. In S. M.
Deen, editor, Cooperative Knowledge Based Sys-
tems, pages 30-52. Springer-Verlag, 1990.

[3] J. Muthuraj, S. Chakravarthy, R. Varadarajan, and
S. B. Navathe. A formal approach to the vertical
partitioning problem in distributed database design.
In Proc. I,, / Conf. on Parallel and Distributed
Information Systems, pages 26-34, 1993.

9 Federated/Multimedia Information
Systems
9.1 Objectives and Approach
This project has identified two critical problems in pro-
viding database management system support for medi-









cal research and teaching environments [1]: i) support-
ing a federated multimedia database using currently ex-
isting ("legacy") systems and ii) providing transparent
access to the multi-media database to meet the data re-
quirements of various groups. Based on our analysis of
the current setup and future requirements, we propose
an architecture consisting of a cluster of existing systems
as federated server and two types of clients- regular and
real-time -with different functional and throughput re-
quirements. The mediator/middleware concept is used
for supporting a federated server.
9.2 Current Status
A logic-based approach to federated databases has been
taken in [3, 4] for: schema integration, inferring inter-
relational correspondences, mapping global queries on
to local schemas, and query optimization. Currently
a p.1.I ,..,. I database" is being implemented as a
client multi-media DBMS for the Radiology Department
at U.F. This database system is being developed to
assist researchers, physicians, instructors and students
in documenting interesting medical cases and to provide
an effective graphics-oriented browsing environment for
teaching purposes. On a different aspect of this project,
a diagnostic expert systems has been prototyped for
radiation therapy [2] by extracting rules (knowledge
engineering) from experts who prescribe the therapy. As
this prototype has been deemed a success by the experts,
this system is being extended to include chemotherapy
and other aspects of pediatric neuro-oncology care.
Support and Investigators: This project is directed
by Sharma Chakravarthy. Profs. B. L. Maria, J.
Honeyman, and D. Dankel are active participants of
this project. This work is supported by the grants of
Profs. Maria and Honeyman.

[1] S. Chakaravarthy, B. L. Maria, V. Krishnaprasad,
Z. Tamizuddin, and F. A. Lambay. A federated
multi-media dbms for medical research: Architec-
ture and functionality. In Proc. of the I,, / Conf. on
New Generation Technology and Information Sys-
tems, Israel, June 1993.

[2] B. L. Maria, F. A. Lambay, S. Chakravarthy,
S. Tufekci, and D. Dankel. Xneo: A diagnostic
expert system for pediatric neuro-oncology. To
appear in Proc. of 4th I,, / Conf. on Management
and Technology. University of Miami, March 1994.

[3] W. K. Whang, S. Chakravarthy, and S. B. Navathe.
Relational schema integration: Dealing with inter-
relation correspondences and querying over compo-
nent relations. Computing Systems, 6(3), Aug. 1993.

[4] W. K. Whang, S. B. Navathe, and S. Chakravarthy.
Logic-based approach to realizing federated informa-


tion systems. In I,, / Workshop on Interoperability
in Multidatabase Systems, pages 92-100, 1991.

10 Bio: An Interoperability
Environment for Remote Molecular
Biology Databases and Analysis
Tools
10.1 Objective and Approach
The purpose of the XBio project is to apply a new ap-
proach for the interoperable access of remote heteroge-
neous databases and analysis tools. The new approach
is based on the formal syntax description of the reports
generated by remote database application programs and
analysis tools, as well as their inputs. The Abstract Syn-
tax Notation One (ASN.1) and the Standard General-
ized Markup Language (SG \ I ) are used as the basis for
the integration. Because these reports are provided as a
service to the public, they represent the most stable part
of the databases. We argue that these reports represent
the most natural basis for building a stable integration.
The main advantages of this approach are that both
databases and analysis programs are uniformly treated,
data can flow easily from one application to the next
without having to worry about format conversions or
data extraction explicitly, it requires no global schema,
and it is easily extended and customized by the user.
10.2 Current Status
The XBio project has resulted so far in a graphically-
oriented system which is being further developed and
extended using the proposed approach. The project is
being developed as a collaborative work between the
University of Florida, the Human Genome Center for
Chromosome 22, and the Naval Postgraduate School.
Support and Investigator: Support for XBio is cur-
rently pending from the National Science Foundation.
The work done so far has been done as departmen-
tal research in the Database Center. Nabil Kamel
(kn.aml'.i. i- ufl -Ii), who is a member scientist in the
Center, is the leader of the XBio project.
[1] Kamel, N., Song, T. and Kamel, M., "An Approach
for Building an Integrated Environment for Molecular
Biology Databases and Software Tools," The It /
Journal of Distributed and Parallel Databases, Vol. 1,
No. 2, 1992, pp. 303-327.
[2] Kamel, N. N., "A Profile for Molecular Biology
Databases," Computer Applications in the Biosciences,
Vol. 18, No. 4, 1992.
[3] Kamel, N. N., Song, T., and Kamel, M., I........I. ,I-
ing GUI in Integration (.1 \l..I ..lI Biology Databases,"
The 3rd IEEE Workshop on Future Trends of Dis-
tributed Computing Systems, Taiwan, 1992, pp. 404-
410.









[4] Kamel, M. and Kamel, N. N., "The Federated
Database Management Systems an Architecture for
Distributed Systems for the 90's," Proc. of the
Second IEEE Workshop on Future Trends of Distributed
Computing Systems, Egypt, 1990, pp. 346-352.
[5] Reddy, G.R., Dame,J.B., Chakrabarti, D., Almira,
E.C., Schuster, S.M. Ferl, R.J., Yang, T.P., Rowe, T.C.,
Laipis, P.J., Kamel, N.N., Cockburn, A.F., \ 111 ....
Genome Project: A Novel Approach to Sequence and
Analyze the Genes of The Malaria Parasite, Plasmod-
ium Falciparum, Utilizing the Genease Activity of Mung
Bean Nuclease (abstract)," Genome Sequencing and
Analysis Conference IV, 1992.
[6] Su, S. Y. W. and Kamel, N. N., "An Object-oriented
Knowledge Base Management Technology for Support-
ing Scientific Research and Applications," Proc. of The
,.1 I, / Conf. on Bioinformatics, Supercomputing,
and Complex Genome Analysis, 1992.

11 Active Database Rule Condition
Testing Optimization
11.1 Objective and Approach
The goal of this work is to make testing of conditions of
active database production rules as fast as possible using
optimization. A discrimination network, such as a Rete
network, has a structure very close to that of a relational
database query plan. The approach being taken
extends the traditional approach to query optimization
by taking into consideration additional information
about update frequency, which is a key variable in
discrimination network optimization but is not relevant
for query optimization. Recent results include the
design of a generalized TREAT/Rete discrimination
network called Gator, and the development of a Gator
network optimizer that takes as input the description
of a rule condition, a database, and update patterns,
and produces an optimal Gator network for the rule.
Early experiments with this optimizer show it can
result in orders of magnitude speed-up compared with
unoptimized Rete networks.
11.2 Current Status
A sensitivity analysis of the optimizer is being done.
When this is completed, the next step will be to
implement a Gator network rule condition matching
mechanism in the Ariel active DBMS.
Support and Investigator: Support is pending from
NSF, AFOSR; investigator is Eric N. Hanson,
[1] Hanson, E., "Gator: A Discrimination Network
Suitable for Optimizing Production Rule Matching,"
IJCAI-93 Workshop on Production Systems and Their
Innovative Applications, August 1993; longer version is
UF CIS TR-009-93, Feb. 1993.


[2] Hanson, E., "The Design and Implementation
of the Ariel Active Database Rule System," IEEE
Transactions on Knowledge and Data Engineering,
1993; accepted for publication.

12 Reliable and Convenient Interaction
Between Active Databases and
Applications
12.1 Objective and Approach
Active database systems now typically support a capa-
bility whereby the action of a rule in the DBMS can
invoke some computation in an application program, or
set of programs, running outside the DBMS, typically on
a client workstation, (e.g., put a pop-up window on my
display if the price of my favorite stock looks cheap).
However, current implementations of this suffer from
a reliability problem, i.e., requests from the DBMS to
applications may be lost (we call this the lost depen-
dent operation (LDO) problem), or requests based on
non-committed updates may be processed by applica-
tions (we dub this the dirty dependent operation (DDO)
problem. We are developing solutions to the LDO and
DDO problems and extending Ariel to support reliable
interaction with application programs.
12.2 Current Status
The Ariel system is now being extended with a data
communications system and durable queues to support
a recoverable application/active database communica-
tion mechanism.
Support and Investigator: Support is pending from
NSF; Eric N. Hanson is the PI.

13 Active Databases for Parallel
Shared-nothing Architectures
13.1 Objective and Approach
We are investigating implementation of production-
system-style active database systems on parallel shared-
nothing architectures, with the goal of increasing the
power and speed of active database rule systems. The
approach is to develop a general rule language that
supports pattern matching, user-defined matching, and
aggregation, a parallel discrimination network, and
effective partitioning strategies for data, rules, and
discrimination networks.
13.2 Current Status
The project is in the early design stages.
Support and Investigator:
Support is by NCR/Teradata; Eric N. Hanson is PI.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs