Foundations Of Highly Available System Design Part 1 - Achieving 5 9's of Availability

Systems which are 99.999 %(5 9s) available or operational throughout the year are called highly available systems. This basically means that your system is only down for -

• 5.26 minutes in a year.
• 1.31 minutes in a quarter.
• 26.30 seconds in a month.
• 6.05 seconds in a week.
• 864.00 milliseconds in a day.

The basic way to achieve a highly available system is by eliminating single points of failure.

If you see the image attached then we have -

1. Multiple Servers - If one fails then the request can be redirected to the another one. But who will monitor if server is up and running and who will redirect the request?
2. Comes load balancer into picture which is responsible for monitoring the servers and redirecting the traffic to the another available server. But what if the load balancer itself fails?
3. Comes a stand-by load balancer to rescue which monitors the health of primary load balancer and comes into picture if the primary one fails. But how this fallover to stand-by load balancer works?
4. Here comes the need of rapid IP remapping mechanism so that load is quickly transferred to the standby load balancer, so floating IP comes into picture here.

Floating IP is a static virtual IP that is assigned to active primary load balancer, if it fails then the virtual IP is assigned to standby load balancer, so basically this virtual IP floats b/w the two load balancers hence called Floating IP.

This is a very basic foundation to achieve a highly available system will add more parts soon. If you find the article useful and want more such articles than subscribe the newsletter and follow Kartik S.

Cheers

Kartik Sapra