5 Ways to Accelerate and Scale Out PostgreSQL€¦ · Scale Out PostgreSQL Denis Magda Apache...

5 Ways to Accelerate andScale Out PostgreSQL

Denis Magda

Apache Ignite Committer and PMC Chair

GridGain VP of Product Management

2019 © GridGain Systems @denismagda

Agenda

1

• Tapping into RAM with caching techniques

• Sharding and replication solutions

• Cache and scale out with in-memory data grids

• Q&A

2019 © GridGain Systems @denismagda2

Caching Techniques


Ultimate Purpose of Caching

3

Speed up operations by reducing

disk access and computation (i.e. CPU)


Computer Latency at Human Scale

4

System Event Actual Latency Scaled Latency

One CPU cycle 0.4 ns 1 s

Level 1 cache access 0.9 ns 2 s


Level 3 cache access 28 ns 1 min

Main memory access (DDR DIMM) ~100 ns 4 min

Intel Optane DC persistent memory access ~350 ns 15 min

Intel Optane DC SSD I/O < 10 µs 7 hrs

NVMe SSD I/O ~25 µs 17 hrs

SSD I/O 50-150 µs 1.5 - 4 days

Rotational disk I/O 1 – 10ms 1 – 9 months

Internet: SF to NY 65 ms 5 years


Computer Latency at Human Scale

5

System Event Actual Latency Scaled Latency

One CPU cycle 0.4 ns 1 s



Level 3 cache access 28 ns 1 min

Main memory access (DDR DIMM) ~100 ns 4 min

Intel Optane DC persistent memory access ~350 ns 15 min

Intel Optane DC SSD I/O < 10 µs 7 hrs

NVMe SSD I/O ~25 µs 17 hrs

SSD I/O 50-150 µs 1.5 - 4 days

Rotational disk I/O 1 – 10ms 1 – 9 months

Internet: SF to NY 65 ms 5 years


Basic Types of Caching in Postgres

6

• Query result caching

• Query plan caching

• Relation caching

– Data and indexes


Relation caching: Shared Buffer and OS Buffer

7

• Postgres Shared Buffer Cache

– Allocated and managed by Postgres

• OS Buffer (aka. Page Cache)

– Caches chunks (pages) of files

• Suggestions/considerations:

– No silver bullet – select and tune

– Possible duplication between shared and OS caches

– Limited by local RAM capacity

Postgres

Shared Buffer

OS Buffer

Disk

Data flow on

reads/writes


Horizontal Scalability


Defining Requirements for Solution

9

• Strong Consistency (ACID)

• Load Balancing

• High-Availability and Failover


Pgpool 2 for Read-Heavy Workloads

10

• Pgpool coordinator

• Primary for reads and writes

• Hot replicas for reads


– Good for load balancing of read-

heavy workloads

– ACID enforces sync replication

and limits a number of replicas

– Primary machine capacity defines

your total cluster capacity


Sharding With PostgreSQL-XL or CitusData

11

• Coordinator keeps metadata and distributes queries

• Data nodes store shards/partitions

• Supports data co-location and JOINs


– Suited for mixed workloads

– Total capacity is your cluster capacity

– Scaling and failover is not trivial

– Disk-based solution


Caching and Scaling With In-Memory Data Grids


Apache Ignite/GridGain In-Memory Computing Platform

Mainframe NoSQL HadoopIgnite Persistence

Persistent Layer

RDBMS

Machine and Deep Learning

EventsStreamingMessagingTransactio

nsSQLKey-Value

Service GridCompute Grid

Application Layer

Web SaaS SocialMobile IoTR

olli

ng U

pgra

des

Security

& A

uditin

g

Monitoring &

Manag

em

ent

Se

gm

en

tatio

n P

rote

ctio

n

Data

Cente

rR

eplic

ation

Netw

ork

Backups

Full,

Incre

menta

l, C

ontinuous B

ackups

Poin

t-in

-Tim

e R

ecovery

Hete

rogeneous R

ecovery

In-Memory Data Store

GridGain Enterprise FeaturesApache Ignite Features


Primary Ignite Deployment Modes

Enhance Legacy Architecture - IMDG Simplified Modern Architecture - IMDB

Ignite In-Memory Computing Platform

Application Layer

Web-Scale Apps Mobile AppsIoT Social Media

Ignite In-Memory Computing Platform

External Database

NoSQLRDBMS Hadoop

Application Layer

Web-Scale Apps Mobile AppsIoT Social Media

Ignite Persistence


How Postgres is Accelerated?

15


Distributed SQL

Persistent Store

ANSI-99 SQL

Compute Grid

JDBC ODBC

C++.NETJava RESTBinary Protocal

(Thin client)

In-Memory Data StoreIndexes on

RAM or Disk

Ignite Distributed SQL SupportCross-platform

Compatibility

DDL & DML

Support

SELECT, UPDATE,

INSERT, MERGE,

CREATE, DELETE

& ALTER

Dynamic

Scaling


Holy Grail of Distributed World: Affinity Collocation

17

• Related data is on the same node

– Countries and Cities

– Departments and Employees

• Collocated Processing

– Efficient Distributed JOINs

– Collocated Computations

– Reduced network traffic

– Performance boost!


Ignite SQL Queries

1. Initial Query

2. Query execution over local data

3. Reduce multiple results in one

Ignite Node

Canada

Toronto

Ottawa

Montreal

Calgary

Ignite Node

IndiaMumbai

New Delhi

1

2

23


Life Without Stored Procedures: Compute Grid

GridGain Cluster

C1

R1

C2

R2

C = C1 + C2

R = R1 + R2

C = Compute

R = Result

in T/2 time

Automatic Failover

Load Balancing

Zero Deployment


Persistent Store

Server Node


Persistent Store

Server Node


Transactions and Consistency

20

• Distributed Key-Value Transactions

– 2 phase commit protocol

– Spans to Postgres

• Transactional SQL (Beta)

– MVCC

• Strong or relaxed consistency

– Atomic and transactional tables

– Tunable Write-ahead-log settings


Consistency Across Postgres and Ignite/GridGain

• Coordinator writes to the database first

• Commits in the cluster afterwards

• The database must be transactional

– Postgres!


Demo


Apache Ignite – We’re Hiring ;)

23

• Rapidly Growing Community

• Great Way to Learn Distributed

Storages, Computing, SQL, ML,

Transactions

• How To Contribute:

– https://ignite.apache.org/

https://ignite.apache.org/community/contribute.html#contribute


-

50,000

100,000

150,000

200,000

Ap

r-1

4

Jun

-14

Au

g-1

4

Oct-

14

De

c-1

4

Fe

b-1

5

Ap

r-1

5

Jun

-15

Au

g-1

5

Oct-

15

De

c-1

5

Fe

b-1

6

Ap

r-1

6

Jun

-16

Au

g-1

6

Oct-

16

De

c-1

6

Fe

b-1

7

Ap

r-1

7

Jun

-17

Au

g-1

7

Oct-

17

De

c-1

7

Fe

b-1

8

Ap

r-1

8

Jun

-18

Au

g-1

8

Oct-

18

De

c-1

8

Apache Ignite Is a Top 5 Apache Project

Over 2M downloads per year

and 4M total downloadsTop 5 Dev Mailing Lists

1.

2.

3.

4.

5.

Top 5 User Mailing Lists

1.

2.

3.

4.

5.

Monthly Ignite/GridGain Downloads

From January 1, 2019 Apache Software Foundation Blog Post:

“Apache in 2018 – By The Digits”

A Top 5 Apache Software Foundation Project


Q&A

@apacheignite@gridgain

5 Ways to Accelerate and Scale Out PostgreSQL€¦ · Scale Out PostgreSQL Denis Magda Apache...

Documents

Transcript of 5 Ways to Accelerate and Scale Out PostgreSQL€¦ · Scale Out PostgreSQL Denis Magda Apache...