Upload
anoop-nayak
View
145
Download
2
Tags:
Embed Size (px)
Citation preview
Anoop Nayak
No SQL Databases
Introduction
“Years have passed by and the data we use are very different”-Anonymous quote from a blog
Bits and bytes
Business Logic and Records
Photos, Videos and what not
But Database remained same…
1960’s Navigational
DBMS
1970’s Relational DMBS and
SQL
1980’s SQL
2000’s SQL
?
Technology
Traditional Databases
DB1
DB2
DB3
NoSQL Databases
DB1
DB2
DB3
Technology(Continued)
• Developed in several layers which can be replaceable unlike Traditional SQL counterparts
• Does not follow standards and hence it uses the latest technologies. Eg: Javascript is the language used in MongoDB
Performance
0 5000000 10000000 15000000 20000000 25000000
MySQL
MongoDB
Insert 50,000 records
Column1 Time Taken (ms)
Performance (Continued)
0 50000 100000 150000 200000 250000 300000 350000
MySQL
MongoDB
Selecting 50,000 from 2,00,00 records
Time Taken(ms)
Types of NoSQL Databases
• Document Store
• Graph
• Key-Value Store
• Multivalue databases
• Object Databases
• Tabular Databases
• Tuple Store
Document Store
What comes in your mind???
Graph
Heard this name somewhere before?
Key-Value Store
Stores as key-value pairs and hence
#fastest
Object Database
Stores stuff as objects as in the object oriented programming language
Tabular Databases
The normal stuff. But again NO SQL
Tuple Store
Anybody knowing python would know what I mean by this
FYI
Features, Features and Features
No indexing = Faster Searching
Very fast lookups. Lack of flexibility
Many NoSQL databases have caching
Redbull for database. More resources used
Meant for the data today
Easier to maintain Learning Curve
Meant to SCALE
Scalability in DNA Scalable setup is
cumbersome at times
SQLor
NoSQL
So now its your turn
Disclaimer: 1.All the above content is what I inferred from my experience and knowledge from various blogs and sites on internet. This is a relatively new topic but again is being used in the industry and hence is maturing fast.2. I have no hatred towards SQL eventhough this presentation is biased towards NoSQL. But todays applications force us developers/engineers to move onto alternatives.
Thank you