Faculty Member, Computer Science
College of Engineering
S. Rao Kosaraju
About
WORK IN DATABASES: A NEW TYPE OF DATABASE MANAGEMENT SYSTEM
Web site for more documentation on ILE: http://progeny.isu.edu/~vkantabu/ILE/
Link to NSF CDI grant information:
http://www.nsf.gov/awardsearch/showAward.do?AwardNumber=0941371&am
When we study complex systems in any field, we often want to store and process information about entities of various sorts and their interrelationships. The best existing ways to do that at the present time are (1) Relational Databases, (2) Object-Oriented Databases, including variations like Object-Relational Databases, and (3) XML. None of these is really satisfactory for complex systems. Relational Databases have a high level of built-in data redundancy that invites errors and inconsistencies. These redundancies can make the design of a good database schema difficult, and can even make the simple act of data entry very annoying. OO databases and the like only directly support simple relationships and can be hard to use, accounting for their lack of popularity. XML imposes an hierarchy on the entities, and only allows limited breakaway from the hierarchy. Additionally, pointers indicating the non-hierarchical relationships are represented as text rather as true pointers.
I am developing a new kind of database management system called Intentionally-Linked Entities, or ILE. In ILE, relationships among entities will be represented directly as true links among them. Thus general graphs (as in Graph Theory), and in fact more (to be explained below), can be represented naturally. The data model will be similar to the Entity/Relationship data model, which was never implemented very well in the past due to the lack of good programming tools. (The most valiant attempt in the past was the flopped Network Databases.) However, at the present time sufficient tools and programming languages have been developed so that complex linked data structures are now in more widespread use. Such complex linked data structures are used in operating system kernels, for example. Interestingly enough, they have not been used in the database field except in index structures. The main idea behind the ILE database system is simply to use modern linked data structures in the main arena of database storage to the fullest extent possible.
What was meant above by saying that we can represent more than just general graphs in ILE? In a graph, an edge represents a binary relationship, that is, a relationship between two nodes, where the nodes commonly represent entities. In ILE, relationships with arities greater than two are possible, and in fact are convenient to create and naturally represented. Thus ILE data structures are more powerful than general graphs. In fact, in ILE, we can also store a new kind of attribute that pertain not to entities in a static way, but that pertain to the entities as they enter a specific relationship. These extra capabilites of ILE are important in the application of ILE to complex networks such as the ones to be referred to in the last paragraph. (Sorry for the forward reference. I'll correct this when I have more time.)
There is another important property of ILE relationships that is important for the representation of complex systems. Each role in a relationship can have more than one (actually zero or more) entities! Here is how this property can be useful. Consider a relationship called "transaction," referring to a commercial transaction in the Spanish mercantile network we are studying. There are two roles in the transaction, the server (also called "agent") role and the client role. Each role, however, can have more than 1 person serving in it. In fact, there are situations where there may be 0 clients. This is where, for example, there was a notarized document created for some agent(s) but where there was no client involved.
The situation in the previous paragraph is where ILE has a clear advantage over Relational and other kinds of database systems. Suppose that in a transaction relationship there are 2 servers and 5 clients. A Relational system would need 10 rows to represent this, inviting redundancy and spelling errors. ILE represents this as only 1 relationship.
I am working with Dr. J. B. "Jack" Owens of our university's History Dept. in a project in which we will use ILE to store and process information on 16th century mercantile networks in Spain.
Web site for more documentation on ILE: http://progeny.isu.edu/~vkantabu/ILE/
WORK IN VLSI DESIGN: U.S. PATENT AVAILABLE FOR LICENSING.
Dr. Vitit Kantabutra also had several innovations in the area of VLSI Design/Computer Engineering, with a concentration on highly efficient computer arithmetic circuits.
Follow this link for information on a U.S. patent currently available for licensing:
http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HIT
All U.S. patents with Dr. Kantabutra as inventor or co-inventor:
http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HIT
WORK IN ALGORITHMS AND COMPUTATIONAL GEOMETRY:
I am also published in these areas. Details will follow shortly.
OTHER INTERESTS:
I also have intense interests, though not professional, in music and in photography. My photos are of general nature. For links to my photos online try the following:
1. Galleries on pbase:
http://www.pbase.com/vkantabu/
The biggest of these galleries is the one for Pocatello:
http://www.pbase.com/vkantabu/pocatello
2. Pictures on Google
Baltimore Inner Harbor
http://picasaweb.google.com/vkantabu/May252005HopkinsFamily#5032308605
http://picasaweb.google.com/vkantabu/May252005HopkinsFamily#5032308708
http://picasaweb.google.com/vkantabu/May252005HopkinsFamily#5032308829
http://picasaweb.google.com/vkantabu/May252005HopkinsFamily#5032309898
http://picasaweb.google.com/vkantabu/May252005HopkinsFamily#5032308326
http://picasaweb.google.com/vkantabu/May252005HopkinsFamily#5032309713
Pocatello Photo Blog (Haven't put anything new there for a while)
http://picasaweb.google.com/vkantabu/PocatelloPhotoBlog?authkey=Gv1sRg
Concert at ISU, April 24, 2009
http://picasaweb.google.com/vkantabu/ConcertIdahoStateCivicSymphonyCam
NWF photo entries 2009 (closes in July, I am change some of the photos)
http://picasaweb.google.com/vkantabu/NWFEntries2009AllPocatelloAreaPho
Rescue on Chinese Peak, April 27, 2009
http://picasaweb.google.com/vkantabu/RescueOnChinesePeak?feat=directli
My photos on flickr.com
http://www.flickr.com/photos/96517640@N00/
My photos on trekearth.com
http://www.trekearth.com/members/vkantabu/photos/
Contact Information
College of Engineering: Computer Science
Idaho State University
Box 8060
Pocatello, Idaho 83209
U.S.A.
001-208-251-4368
