Redundancy — How to Design Zero Downtime Application Services
Scaling your startup? or wanted to get rid out of nasty feedback on Internet about your product at the time when it goes down the service? Looks scary? Definitely not, Redundancy is there to help.
I always would love to talk more about Scaling your Services so that you can better architect your instances and monitor them on the go without being engaged with other Microservices classes and instances. This time I came up with a new thing and that is ‘Redundancy’. But what exactly is Redundancy and how to it will turn our services and instances to go Zero downtime?
Wait, Let me explain.
In the simplest terms, Redundancy is duplicating the active-server nodes and keeping them on standby mode so if any of the node crashes down, the passive node takes the place of active node without impacting your application.
This instance setup approach is also known as Active-Passive HA mode. An initial nodes are on active mode. Passive nodes are on backup side, if any of the active node fails, the passive node quickly replaces the active node by almost making your application going Zero downtime.
There are systems like Banking, GPS, Air-crafts and Defense where there is no place to compromise on service availability factor. In such case, Redundancy can really help your system backing up your active server nodes at times of failures and crashes.
Getting rid of Single Point of Failures
Other benefit (besides increasing availability factor and zero downtime) is that your system gets rid out of Single Point of Failures due to large number of nodes are working in conjunction to achieve a single synchronous application state. When so many redundant nodes are deployed than there is no chance of Single Point of Failure in system. In case a node goes down, redundant node takes it place and the system remains whole unimpacted.
Monitoring and Automation
Monitoring helps your system to stay out of Single Point of Failures. Automation gives the system the power of self-healing without human intervention. Also the system becomes smart enough to add or remove instances on the go, Kubernetes is such a good example of this.
That’s it.
More coming on the way. Stay tuned.
