1. Presented by
Norman Kutemperor
Gensonix –A cost effective solution
to ‘impossible’ SQL problems
Scientel Info. Tech.
Email: norm@scientel.com
Website: www.scientel.com
2. WHO IS SCIENTEL?
• International software development
company of DBMS Systems and
applications.
• Creators of sophisticated custom crafted
solutions
• Solved several Data Intensive problems
Utilizing NOSQL Databases since 1977
3. PLATFORM
• Gensonix DBMS is a very
powerful, database
management system.
• Supports relational,
hierarchical/vertical, network,
and column schemas
• Gensonix by itself can
address all of today's
information management
tasks
4. Gensonix NSQL
• Gensonix program modules are
smaller/efficient
• Gensonix compiler optimizes to native level
• Gensonix optimized objects execute faster
• Gensonix NSQL syntax is carefully designed to be
extremely simple
• Provides high reliability
• Extremely Short Learning curve
• Eliminates the need for Middleware
• Needs only HTML/JS
5. Gensonix DB Organization
• GENSONIX data store - base level storage
• Enhances speed tremendously
• Simple Smart Procedural syntax for better flow
• Relies less on other software agents
• Minimizes ETL
• Minimizes Map/reduce
• Indexed and non-indexed data logging
• Ideal for Business AND Non-commercial apps.
• Exports to Mysql for SQL queries if needed
6. Gensonix NSQL Capabilities
• In-memory Arrays and management
• In-memory Indexing
• In-memory Tables
• Exact and precise procedural Syntax
• Pre-determined execution control
• Extensive processing w/o JOINS!
• Raw data tables with no indices
• Large data types - including images
7. NoSQL Databases
• Not a new concept –Gensonix since 1977
• May or may not conform to SQL standards
• Generally smaller footprints
• Therefore low system overhead
• Requires comparably less hardware
• High Speed
• High Scalability
8. SQL Databases
• Stored in Relational Databases
• Allows query of all or any items from all or any
tables
• Large band width required to drive such queries
• Band width and resources reserved/declared up
front to satisfy all potential queries
• Detailed queries are possible but not always
needed
• More and More Businesses adopting to deal
with concise summaries
9. Heavy Requirements of SQL
• Requires large hardware
• Often large foot print Databases
• Complex queries - Difficult to construct
• Difficult to translate by some connectivity
agents
• Difficult to understand by some humans
due to non-procedural syntax
• Duplication of resources – Join tables
10. SQL MODEL I
CUSTOMER MASTER
ID NAME ADDRESS CITY STATE ZIP
1 COMQ 3507 PARK LANE OAKPARK IL 48569
2 DELLDATA 1600 BEACH DR COLUMBUS OH 46398
400 BYTES X 2 = 800 BYTES
11. CUSTOMER STATS
ID LAST ORDER AMT NO of orders AVG AGE
1 01/10/05 10.00 2 30
2 05/11/06 20.00 5 45
200 BYTES X 2 = 400 BYTES
12. JOIN TABLE
TOTAL 1200 BYTES
ID NAME ADDRESS CITY STATE ZIP LAST ORDER AMT NO of orders AVG AGE
1 COMQ 3507 PARK LANE OAKPARK IL 48569 01/10/05 10.00 2 30
2 DELLDATA 1600 BEACH DR COLUMBUS OH 46398 05/11/06 20.00 5 45
#OF records = 4 JOIN TABLE SIZE 2400 Bytes
#OF records = 8 JOIN TABLE SIZE 4800 Bytes
In this scheme WITH EACH RECORD ADDED THE JOIN TABLE SIZE INCREASES BY THE SUM OF
THE TWO TABLES NOT COUNTING INDEX TABLE AND CONTROL TABLE
13. THE GENSONIX MODEL I
CUSTOMER MASTER CUSTOMER STATS
Direct
ID NAME ADDRESS CITY STATE ZIP
operation LAST NO of
ORDER Terms AMT orders AVG AGE
3507 PARK OAKPAR
1 COMQ LANE K IL 48569 01/10/05 4 10.00 2 30
DELLD 1600 BEACH COLUMB 05/11/06 2 20.00 5 45
2 ATA DR US OH 46398
RELATIONAL LAYOUT – NETWORK METHOD
NO JOINS No JOIN TABLES
SAVES MEMORY
SAVES DISK
SAVES TIME
SIMPLE OPERATION
SPEEDY DELIVERY
SAVE MONEY
14. SQL MODEL II
CUSTOMER TABLE ORDER TABLE
CUST ID ORDER # ORDER DATE AMT
ID NAME ADDRESS CITY STATE ZIP
3207 OAK 48569
1 1153 11/02/02 20.00
1 COMP Park lane PARK IL
1 1245 07/03/03 20.00
1600 COLUMBUS OH 46398 2 1358 05/04/05 40.00
2 DELLDATA BEACH DR
2 1478 03/11/07 50.00
579 LORR PARK
3 JET WAYS DR CITY UT 78901 3 1698 02/12/08 60.00
3 1698 02/09/09 70.00
JOIN TABLE
ID NAME ADDRESS CITY STATE ZIP ORDER # ORDER DATE AMT
3207 OAK 48569
1 COMP Park lane PARK IL 1153 11/02/02 20.00
3207 OAK 48569
1 COMP Park lane PARK IL 1245 07/03/03 20.00
1600 COLUMBUS OH 46398
2 DELLDATA BEACH DR 1358 05/04/05 40.00
1600 COLUMBUS OH 46398
2 DELLDATA BEACH DR 1478 03/11/07 50.00
PARK
3 JET WAYS 579 LORR DR CITY UT 78901 1698 02/12/08 60.00
PARK
3 JET WAYS 579 LORR DR CITY UT 78901 1698 02/09/09 70.00
PROCESSED ESSENTIALLY THE SAME WAY AS METHOD I
15. GENSONIX MODEL II
CUSTOMER TABLE
ID NAME
ADDRES
S CITY STATE ZIP
RELATIONAL LAYOUT
3207 OAK 48569 HIERARCHICAL METHOD
1 COMP Park lane PARK IL
1600 COLUMB
BEACH US OH 46398
2 DELLDATA DR
579
LO
3 JET WAYS
RR
DR
PARK
CITY UT 78901 ORDER TABLE
CUST ID ORDER # ORDER DATE AMT
1 1153 11/02/02 20.00
1 1245 07/03/03 20.00
NO JOINS - SUBSTANTIALLY
DIFFERENT PROCESSING { 2
2
1358
1478
05/04/05
03/11/07
40.00
50.00
METHOD 3 1698 02/12/08 60.00
3 1698 02/09/09 70.00
ONLY ACCESSES PART OF
BOTH TABLES
16. GENSONIX COLUMN DATABASE MODEL
QTY SOLD
YEAR JAN FEB MAR APRIL MAY JUNE JULY AUG SEP OCT NOV DEC
2011 10 11 10 12 10 13 25 30 9 7 10 10
2010 25 30 9 7 10 10 0 0 7 9 5 4
2009 10 0 0 25 30 9 7 10 10 10 9 9
2008 3 4 45 20 25 30 9 7 10 10 12 11
COLUMN PROCESSOR ROUTINES
FIRST INTRODUCED IN 1997
SALES STATISTICS IMPROVED BY A FACTOR OF 900%
REDUCED STORAGE
INCREASED LOOKUP TIME
17. Examples of –Products
developed in Gensonix
DBIS Intranet DBIS e-Commerce
For the Enterprise On the Web
18. An Enterprise problem -
A recursive Gensonix Solution
• A manufacturer wishes to create Multi-level
Assembly for building a very large End product
Ex: An airplane
• An average of 10 items at 75 levels
• Each level includes at least 1 sub-assembly
• Therefore the product utilizes 750 items
• The problem: They need to know the real time
cost of the end-product
19. DBIS Intranet
• Scientel's Enterprise Solution - (DBIS)
Distribution Business Information
System
• Web-like Supply Chain Management
system
• Designed to run on your in-house
intranet.
• Diverse enough in its range and depth
of applications
• Manages the entire distribution
operation.
23. A Web problem – A Slick Gensonix solution
• A Supplier Sells over 100,000 items through a catalog
• They operate multiple branches
• They are located throughout North America
• They have a 1000+ page catalog
• The problem: Sell items on-line
• Display 1000+ page catalog on-line
• Surfers to be able to Interact with catalog
• Have excellent response time
• Have periodic sales/promotions
• Have live inventory
• Accept orders right from the catalog
• Route on-line orders right to the appropriate branch
real-time
• Automatically post to their in-house DB
• Print shippers on the shop floor
26. Why Gensonix?
• Gensonix is designed with the following criteria
• 1 Small foot print
• 2 Affordable cost
• 3 Easy to learn
• 4 High speed of operations
• 5 Easy to Administrator
• 6 Allows Efficient systems design process
• 7 Highly stable and Reliable End-product
• Resulting in the 5S’s
• STABILITY
• SCALABILITY
• SPEED
• SIMPLICITY
• SOUNDNESS
27. Conclusion:
Gensonix Enterprise Today
• GENSONIX Enterprise Highly Scalable Server Environment
• Achieving Multi-Million Transactions per minute
• Scales to very large #of Nodes
• Supports Up to 17 TB/Db node
• Redundant node support
• Multi-terabyte+ scalability
• High Uptime
• High reliability
• Indexed and Non indexed Table structures
• Intake speed ~ 1TB/hr
• Supporting OLTP - No Unloads – Housekeeping on the fly
• Custom ultrahigh speed or Commodity HW