Upload
others
View
16
Download
0
Embed Size (px)
Citation preview
Copyright © 2017 NTT. All Rights Reserved.
Partitioning tables across multiple servers to scale out PostgreSQL
PostgreSQL sharding with foreign data wrappers
NTT, in cooperation with a development community*1, is committed to PostgreSQL development in order to expand its usage to enterprise areas by scaling out a table to multiple servers. PostgreSQL now has the capabilities of load-balancing for a large amount of simple queries and of pushing down analytic queries to off-load the query execution to multiple distributed servers.
Features ■ Open source technology. Developed as a built-in feature of open
source DBMS*2 which leads to TOC*3 reduction. ■ Distributed table. Transparent access from your application to
tables distributed in multiple servers the same as a single table. ■ Various data sources. Manage various external data sources from
PostgreSQL using foreign-data wrapper, e.g., Hadoop, Oracle and MySQL.
Application Scenarios ■ Adjust to your business growth. Seamless performance scale-out
by adding servers as your business grows. ■ Large data analysis. Speed up your analytic queries by pushing
down the query execution to multiple distributed servers that store large-scaled data expanding with daily business operation.
■ Simple access to multiple data sources. A single point of access for a batch execution searching among various kind of multiple data sources in large enterprise systems.
Collaboration Partner The PostgreSQL Global Development Group
*1 The PostgreSQL Global Development Group *2 DBMS: data base management system *3 TCO: total cost of ownership
〈Contact〉[email protected]
Scaling out
PostgreSQL
PostgreSQL
結果 出力
JOIN SORT
SQL Query
結果 出力 Result
Foreign-data wrapper
PostgreSQL
PostgreSQL
Relational operations (sort, join, …)
JOIN SORT
JOIN SORT
Sharded table Sharded table Sharded table
PostgreSQL
JOIN SORT
Sharded table
Load balance Distribute select and update queries to sharded servers
A B C D
A B C D
Table alias
Query pushdown Execute at remote
server
Open source developers worldwide
Technologies to enhance networks
D-20