What is Load Balancer?

A load balancer is a way to have multiple servers perform the same operations. The main purpose of this device/software is to make sure to avoid the burden on a single application by districting workloads. Many new webservers deal with thousands if not millions of users and return the correct text, image, or video, all in a fast and reliable manner.

Load balancing categories

• Sender initiated – it is where sends locates the best target
• Receiver initiated – receiver selects the best target.

Benefits of Load Balancing

Some of the benefits of load balancing are:

• Reduce downtime
• Scalability
• Redundancy
• Flexibility
• Efficiency

Static load balancing

This type of load balancing divides traffic equally or based on rules. In this process, actions are always processed on the assigned target and it lacks scalability. Round robin is an example of static load balancing.

Round Robin

Round Robin load balancing rotates connections between servers in the order that requests are received.

Sample how this works

• 1^st request goes to 1^st servicing node
• 2^nd goes to 2^nd
• 3^rd goes to 3^rd
• 4^th goes to 1^st
• 5^th goes to 2^nd
• 6^th goes to 3^rd
• 7^th goes to 1^st
• As request comes in it rotates through servicing nodes

Dynamic load balancing

Dynamic load balancing works on the present traffic because it transfers workload to underutilize machines in real-time. It is far more complex than static but modular because there is no assignment of dedicated nodes to work distribution therefore it has a modular architecture. Also, it has the ability to scale.

Examples of Dynamic load balancing are:

Randomized

This works as the name implies. It generates a random number by using an underlying random number generator. If any nodes face a high number of requests, a Random algorithm generates a number to evenly distribute traffic.

Centrally Managed

In this process, all decisions are based on intelligence and it may vary based on current service load, time of day, and specific kind of request.

Threshold-Based

This type of request sends on the particular server until a certain amount of use. If the number exceeds let’s say more than 10% of requests send to another server.

References:

https://avinetworks.com/what-is-load-balancing/

https://avinetworks.com/what-is-load-balancing/

https://www.f5.com/services/resources/glossary/load-balancer