What is the difference between reverse proxy and load balancer?

I always hear the two terms are synonyms, while they actually have a difference.

Answers given to this question in the above link seems like load balancer has more in term of capabilities than reverse proxy. Can anyone please clarify this? A reverse proxy is only a reverse proxy. A load balancer is just a generic term used to describe a device that can perform load balancing (distribution of incoming requests based on factors such as content/protocol, application, server load, security etc). Reverse proxies provide both load balancing and caching/proxy functionality.

Reverse proxies: Forward requests to backend servers for. A specific domain, IP, port or protocol. Redirect HTTP requests to HTTPS for. Forwared request based on rules. Server/Application based. Load balancers: Distribute incoming requests to backend. Servers based on a set of load. Balancing rules, or algorithms. Network/Interface based. A reverse proxy is a software component that takes a request from one IP/domain and delegates it to another. A load balancer typically takes incoming requests from a larger number of IPs and distributes them out to multiple backends. A load balancer will handle connections between the front-end and the backend.

What is load balancer architecture?

The load balancer is a kind of network device used in high-load, highly-demanding applications to distribute incoming network traffic across a cluster of servers.

A load balancer is responsible for allocating traffic to servers in a cluster based on the rules set by the administrator. To ensure reliability, a load balancer typically supports a round-robin approach that distributes incoming network traffic evenly among multiple servers in a cluster, or it uses more advanced techniques, such as weighted round-robin and least connections.

Why to use a load balancer? Load balancers come in a wide variety of different sizes, performance, price, and features. Each of them comes with its own advantages and disadvantages. The purpose of using a load balancer is to distribute traffic evenly among several servers. To achieve this objective, the load balancer uses a number of protocols to determine which server(s) a request should be routed to. These protocols include round-robin, least connections, weighted round-robin, and others.

Round-robin. Round-robin is the simplest and most basic protocol. The load balancer receives a request and sends it to each server in turn. The server that gets the request next time will receive another one, until all requests are completed. Round-robin works fine when you don't have many servers.

Pros: Simple. No complex configuration. Cons: Not suitable for applications with a large number of servers. Least connections is another simple protocol. This protocol uses a queue, and whenever a new request arrives, the queue shifts one element to the head. The next request is sent to the server from the end of the queue. This method ensures that all servers receive equal shares of requests. However, the request might be routed to a server that is busy, so the protocol might cause a backlog of pending requests on the queue.

Suitable for applications with a large number of servers. May cause a backlog of pending requests. Weighted round-robin is similar to least connections in that it ensures that all servers receive equal shares of traffic. The difference is that weighted round-robin takes into consideration the workload of each server. The heavier a server is working, the more likely it will be selected.

Can we use load balancer and reverse proxy together?

Recently I'm experimenting with nginx reverse proxy and found it very powerful.

But I wonder if we can use it in front of Kubernetes? Can we use nginx to handle the incoming requests and create the pods in a cluster as a load balancer does? I don't have a very good idea about it, and I didn't find anything in the official documentation about this use case. The most similar is Webinoss: nginx Reverse Proxy In Front Of Kubernetes which seems not complete.

I also saw some articles where people use kube-proxy in front of Kubernetes cluster and use nginx in front of it to handle the incoming requests and create the pods, but I don't know if it's still possible now because kube-proxy has been deprecated now. You can use nginx as an ingress controller on your Kubernetes cluster. Ingress controllers are meant to route incoming traffic to backend services. When you forward incoming requests to a Kubernetes service, the request will be passed to a backend Pod in the Kubernetes cluster that's associated with that service.

Ingress controllers typically perform this routing using nginx (or some other web server), as a reverse proxy. This setup allows you to route requests from outside the Kubernetes cluster to Kubernetes services.