Merikanto

一簫一劍平生意,負盡狂名十五年

Notes on System Design

In this post, we will cover a few notes on scaling the system, including Load Balancer & Reverse Proxy, Vertical & Horizontal scaling, and links to some architectures.


Load Balancer & Reverse Proxy

A reverse proxy accepts a request from a client, forwards it to a server that can fulfill it, and returns the server’s response to the client (which means a server behind the Reverse Proxy can communicate with somewhat different features of a protocol or a different protocol).

A load balancer distributes incoming client requests among a group of servers, in each case returning the response from the selected server to the appropriate client.


When you refer to a load balancer you are referring to a very specific thing - a server or device that balances inbound requests across two or more web servers to spread the load. A reverse proxy, however, typically has any number of features:

  • load balancing: as discussed above

  • caching: it can cache content from the web server(s) behind it and thereby reduce the load on the web server(s) and return some static content back to the requester without having to get the data from the web server(s)

  • security: it can protect the web server(s) by preventing direct access from the internet; it might do this through simple means by just obfuscating the web server(s) or it may have some more active components that actually review inbound requests looking for malicious code

  • SSL acceleration: when SSL is used; it may serve as a termination point for those SSL sessions so that the workload of dealing with the encryption is offloaded from the web server(s)

Also, a reverse proxy is specific to web servers. Load balancers however can deal with a lot of other protocols. While the web (HTTP) is the big idea nowadays, things like DNS, mail (SMTP, IMAP), etc. can be load balanced as well. It’s just nowadays when most people think “Internet” or “IP network” they think of the web. There’s a bunch more stuff out there that may be more obscure, or more of a niche.


Vertical & Horizontal Scaling

Vertical Scaling

Vertical scaling, or improving the capabilities of a node/server, gives greater capacity to the node but does not decrease the overall load on existing members of the cluster. That is, the ability for the improved node to handle existing load is increased, but the load itself is unchanged. Reasons to scale vertically include increasing IOPS, increasing CPU/RAM capacity, and increasing disk capacity.


Horizontal Scaling

Horizontal scaling, or increasing the number of nodes in the cluster, reduces the responsibilities of each member node by spreading the keyspace wider and providing additional endpoints for client connections. That is, the capacity of each individual node does not change, but its load is decreased. Reasons to scale horizontally include increasing I/O concurrency, reducing the load on existing nodes, and increasing disk capacity.


Architectures