Jagannathan S / Oct 2010

streakconvertingSoftware and s/w Development

Dec 13, 2013 (3 years and 8 months ago)

99 views

Jagannathan S / Oct 2010

Content presented here do not represent the views of Hewlett Packard

Company. These are the perspectives and views of the presenter

Why a DBMS ?


Provides an easy way for


Modeling/designing structured data


Accessing this structured data


Query language and
indexing for fast search


Multiple users to edit shared data without
compromising data integrity


ACID Transactions


Eliminates redundancies in storing shared data


Evolution of DBMS

Wave 1

File
-
Based

Hierarchical

Network

Wave 2

Relational

Multi
-
dimensional

In
-
memory

Wave 3

Object
-
Relational


XML

Wave 4

DBMS In the “Cloud”

DBMS


The Big Players

Commercial DBMS Offerings

Open Source DBMS Offerings

Source: IDC Embedded DBMS Forecast


Jan 2010

Most common Operating Environments for DBMS are : Windows, Unix and Linux

Demands from


Application evolution


Data Security and
Protection


Data Growth


High availability and
Internet adoption

Enablers

Evolution of


Server / Computing


Storage Technologies
and Hardware


Networking Speeds


DBMS Technologies

Challenges / Drivers

DBMS Feature Evolution


Backed up by a proven
mathematical model


Business Applications


The de facto standard for backend
of internet applications


SQL Standard / Programmer
community


SQL Abstraction Paradigm


JDBC,
Hibernate etc


Simple Applications


Backup & Restore

Enablers


Early Microprocessors


End
-
user terminals


Direct
-
attached
Storage


Standalone tape
drives


Early LAN

Challenges / Drivers

DBMS Features



SQL Execution



ACID Transactions



Import / Export



Business applications


High availability


Basic security


High performance


Distributed use
models


Growth in data

Enablers


32 bit CPUs


SATA Disk Drives /
RAID


Tape autoloaders


Volume Management


Mainstream LAN


SQL Programmer
community



Challenges / Drivers

DBMS Features



Basic Clustering


Basic Security / Audits


Distributed queries


On
-
line backup


Row
-
level Locking



Internet adoption


Online commerce /
24X7 availability


Java


Compliance needs
and data archival
needs


Early content
management

Enablers


64 bit CPUs


Disk Arrays


Encryption and
compression


Tape libraries


Early NAS


LAN/ WAN/ 1Gbps


OOAD


Multi
-
pathing



Challenges / Drivers

DBMS Features



Stored Procedures, Triggers


Advanced Security


Database Replication


Parallel Query


Non
-
structured Data Types


64
-
bit / Large memory support


Object Relational DBMS


Table Partitioning


Advanced backup/recovery


Information Lifecycle
management


Explosive digital
content growth


Virtually zero backup
windows


Shared hardware


Global organizations


Early
SaaS

/ Cloud
applications

Enablers


Multi
-
core CPUs / Grid
computing


Server and storage
virtualization


SAS Disk arrays


QoS


Deduplication /
D2D2T


Thin provisioning


10Gbps n/w





Challenges / Drivers

DBMS Features



Java and XML support


Advanced Clustering


Table compression


OLAP Support


Grid Computing


Column Partitioning


Parallel backups/recovery


Advanced Manageability


Large database support


Unstructured Data
Explosion


Mainstream
SaaS

/
Cloud usage models


Real time BI / DSS


Zero Admin


Software Appliances


“Green” requirements

Enablers


Cloud hardware


Solid State Drives


Tiered Storage and
Storage Pooling


CDP


iSCSI


100Gbps n/w


Heuristics / learning/
AI


Green IT






Challenges / Drivers

DBMS Features



Datawarehousing

Databases


Key Value Databases


Cloud Service Databases


Move from ACID to BASE


Non
-
SQL query language


Did you know ?


Oracle was started in 1977


it was the name of a project that
Larry Ellison completed even after it was scrapped!


The recording density for data


aka capacity


has increased
60,000,000
-
fold in 50 years!


The amount of worldwide information is projected to be 988
exabytes

by end of 2010! An Exabyte is a million terabytes


The magnetic HDD is 50 years old. In 1956 IBM introduced 305
RAMAC which was the size of a refrigerator, and stored a total
of 4.4 megabytes on 50 doubled
-
sided, two
-
foot
-
diameter
disks. The disk had a purchase price of $10,000,000 per
Gbyte

and weighed over a
tonne
!


Global digital information doubles every 18 months!


FaceBook

is one of the largest
MySQL

installations world
-
wide!


OLTP
vs

OLAP Databases

OLTP

OLAP

Designed for real
-
time
business operations

Designed for analysis of
business measures

Optimized for a common set
of transactions


usually
modify

operations

Optimized for bulk loads and
complex, unpredictable
queries

Optimized for validation

of
incoming data during
transactions

Pre
-
loaded with validated
data, requires no real
-
time

validation

Supports

large number of
concurrent users

Supports few concurrent
users

DBMS
vs

RDBMS


DBMS


Needs to be Persistent


Needs to provide uniform interfaces to applications


Need not impose ACID constraints of the database



RDBMS also


Needs to support a tabular structure to comply to
“Relational” theory


Needs to enforce relationships between tables


Needs to enforce ACID constraints on the database



Datawarehouse

DBMS Magic Quadrant


Gartner
2010