Lecture 14 -NoSQL Databases(1).pptx - Fordham University Gabelli School of Business NoSQL Databases Fordham University Gabelli School of Business A bit

Lecture 14 -NoSQL Databases(1).pptx - Fordham University...

This preview shows page 1 - 9 out of 32 pages.

Fordham University Gabelli School of Business NoSQL Databases
Image of page 1
Fordham University Gabelli School of Business A bit of History: Non SQL databases existed for a long time Initial data organization were not based on Hierarchical organization (IBM’s IMS) 2 Database – each cabinet keep records for a region Cabinet – each drawer keep records for a city Drawer – each drawer keep records for a store Drawer – each drawer keep records for a Customer Example Sales database How would we find record for a customer?What if we need to find all records for sale of an item?What would be the right access mechanism
Image of page 2
Fordham University Gabelli School of Business 3
Image of page 3
Fordham University Gabelli School of Business From Hierarchical Systems to Networks Use of pointers to identify location of records Pointer may be “record number” or the physical location of the record 4 Store XX Customer 1 Customer 3 Customer 2 Customer 4 Product zz
Image of page 4
Fordham University Gabelli School of Business Relational Model Every things as normalized tables Each table represents objects of a given type Each record describes on member of that type All members of the type have same descriptors Each descriptor is described by an attribute, represented as a column of the table Relational databases worked extremely well to support operational activities Easy to enforce ACID ( Atomicity , Consistency , Isolation , Durability ) conditions Worked well for data mining applications 5
Image of page 5
Fordham University Gabelli School of Business Relational Databases Based on relational algebra Underlying framework: Nornamization Focus on operational databases Normalization allows transactions to be captured with bare minimum data Allowed for data distribution via partitioning and replication 6 FOREIGN KEY 1 (Key of participating entity 1) FOREIGN KEY 1 (Key of participating entity 2) FACT (Transaction Specific data)
Image of page 6
Fordham University Gabelli School of Business Limitations of RDMS Poor for managing unstructured data Emergence of unstructured data in form of web pages Blogs Chats Audio and video clips Explosion in data Joins across tables can be time consuming Restructuring can be a cumbersome process 7
Image of page 7
Fordham University Gabelli School of Business NoSQL 8
Image of page 8
Image of page 9

You've reached the end of your free preview.

Want to read all 32 pages?

  • Spring '14
  • WullianallurR.Raghupathi
  • Fordham University, GABELLI SCHOOL OF BUSINESS

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture