Comparison analysis of DBMSs and development of data store module. Сoncept for financial analysis needs

1.

«Comparison analysis of DBMSs and
development of data store module
concept for financial analysis needs»
Made by the Р9-204 group student:
Grinenko A. M.
Research advisor:
Pakhomov A. P.
January, 18 2016г

2.

Aims and goals
Research actuality:
New data technologies give instruments for data storage
improvements but they are not actually used.
Aims and goals:
• Compare different DBMS and data storage approaches
• Measure and explane better performance
• Develop a data store concept for financial analysis
2

3.

Case study highlight
Improved performance = better decision-making:
3
l
Huge amount of information from different organizations
l
Different data types.
l
Unstructured data parsing.
l
Detailed, accurate search.
l
Quick indexes update.

4.

DBMS
Relational DBMS
Object-oriented DMBS
DataBase
Management
Systems
NoSQL Search Engines
In-memory databases
4

5.

Relational vs NoSQL
Compare productivity using two metricks:
Queries per second;
Queries per second per thread;
5

6.

Estimations
l
MongoDB could handle
more complex queries faster
than Oracle;
l
Oracle performs better when
deleting data when
MongoDB performed better
during insertions;
6

7.

ElasticSearch vs Mongodb
Indexing
MongoDb
14.6 sec
ElasticSearch
0,7 sec
Aggregation
MongoDb
42,116 sec(first try) - 3.779 sec(second
try)
ElasticSearch
0,902 sec
Insertion
MongoDb
522 sec -> ~12,159 documents/second
7
ElasticSearch
600 sec -> ~10,597 documents/second

8.

Data storage prototype
8

9.

Results
l
9
Different DBMS and data storage
approaches were compared
Better performance was measured and
explained
The data store concept for financial
analysis workspace was developed

10.

Thank you for attention!
10