Redhshift Architecture

 

#Do you know #Amazon Redshift architecture? #How it actually Works #Internally?
.
.

.
.
.
.
.
.
.
.
.
.
.

Lets understand the Amazon Redshift Architecture in Depth
==============================

🚀Amazon #Redshift Follows a #Master #slave Architecture

✍️Here in #Amazon Redshift :-

✏️Leader node is a master

✏️Compute nodes are slave

✍️We also have #one node cluster in #Amazon Redshift where same node acts as #Leader node as well as #Compute node.

✍️For a #Multinode cluster #Leader node is #Seperate.

Let's see how it works Internally:-
=========================
✍️When Client write a program/Query and run it .

✍️Leader node picks the query  which it received from the Client and parses it and create a execution plan

✍️Leader node based on execution plan assign the tasks to the  #slice nodes of #compute node

What are #slicenodes?
===================

✍️Compute node are internally partitoned into #slice node with equal number of resources distributed.

✍️Let's consider if we have a compute node with 12 GB memory  and 12 CPU cores then let's consider we have 4 slice nodes

✍️then each slice node will get 3 CPU cores each and 3gb memory

a)slice node1 --------------> 3 gb and 3 CPU cores

b)slice node2 --------------> 3 gb and 3 CPU cores

c) slice node3 --------------> 3 gb and 3 CPU cores

d) slice node4--------------> 3 gb and 3 CPU cores

like this be will be distributed.

4) #Slice nodes runs parallely and completes the tasks assigned and give the result back to #Leader Node.

5) #Leader Node aggregates the result  from different compute nodes and gives back it to client .

✍️This is what actually happen internally in Redshift Architecture internally

#Leader node acts as a bridge between compute node and Client

#We cannot direclty communicate with Compute node direclty.
.
.
.