When deciding which server architecture to use for your environment, there are many factors to consider, such as performance, scalability, availability, reliability, cost, and management.
In this tutorial, you will learn about commonly used server setups, with a short description of each, including the pros and cons. Keep in mind that all of the concepts covered here can be used in various combinations with one another and that every environment has different requirements, so there is no single correct configuration.
One server setup is when an entire environment resides on a single server. For a typical web application, this includes the web server, application server, and database server. A common variation of this setup is a LAMP stack, which stands for Linux, Apache, MySQL, and PHP, on a single server. An example use case for this is when you want to set up an application quickly. A basic setup like this can be used to test an idea or get a simple web page up and running.
Unfortunately, this offers little in the way of scalability and component isolation. Additionally, the application and the database contend for the same server resources such as CPU, Memory, I/O, and more. As a result, this can possibly cause poor performance and make it difficult to determine the root cause. Using one server is also not readily horizontally scalable. You can learn more about horizontal scaling in our tutorial on Understanding Database Sharding. Learn more about LAMP stack as well in our tutorial on How To Install LAMP on Ubuntu 22.04. The following is a visual representation of using a single server:
The database management system (DBMS) can be separated from the rest of the environment to eliminate the resource contention between the application and the database, and to increase security by removing the database from the DMZ, or public internet.
An example use case is that this can get your application set up quickly and prevents the application and database from fighting over the same system resources. You can also vertically scale each application and database tier separately. This is possible by adding more resources to whichever server needs increased capacity. Depending on your setup, this may also increase security by removing your database from the DMZ.
This setup is slightly more complex than a single server. Performance issues, like high latency, can arise if the network connection between the two servers is geographically distant from each other. There can also be performance issues if the bandwidth is too low for the amount of data being transferred. You can read more on How To Set Up a Remote Database to Optimize Site Performance with MySQL. The following is a visual representation of using a separate database server:
Load balancers can be added to a server environment to improve performance and reliability by distributing the workload across multiple servers. If one of the load balanced servers fails, the other servers will handle the incoming traffic until the failed server becomes healthy again. It can also be used to serve multiple applications through the same domain and port, by using a layer 7 application layer reverse proxy. A few types of software capable of reverse proxy load balancing are HAProxy, Nginx, and Varnish.
An example use case is in an environment that requires scaling by adding more servers, also known as horizontal scaling. When you set up a load balancer, it enables an environment capacity that can be scaled by adding more servers to it. It can also protect against DDOS attacks by limiting client connections to a sensible amount and frequency.
Setting up a load balancer can introduce a performance bottleneck if the load balancer does not have enough resources, or if it is configured poorly. It can also present complexities that require additional consideration, such as where to perform SSL termination and how to handle applications that require sticky sessions. Additionally, the load balancer is a single point of failure, this means that if it goes down, your whole service can go down. A high availability (HA) setup is an infrastructure without a single point of failure. To learn how to implement an HA setup, you can read our documentation on Reserved IPs. You can read more in our guide on An Introduction to HAProxy and Load Balancing Concepts as well. The following is a visual representation of setting up a load balancer:
An HTTP accelerator, or caching HTTP reverse proxy, can be used to reduce the time it takes to serve content to a user through a variety of techniques. The main technique employed with an HTTP accelerator is caching responses from a web or application server in memory, so future requests for the same content can be served quickly, with less unnecessary interaction with the web or application servers. A few examples of software capable of HTTP acceleration are Varnish, Squid, Nginx. An example use case is in an environment with content-heavy dynamic web applications or many commonly accessed files.
HTTP acceleration can increase site performance by reducing CPU load on a web server, through caching and compression, thereby increasing user capacity. It can also be used as a reverse proxy load balance, and some caching software can even protect against DDOS attacks. Unfortunately, it can reduce performance if the cache-hit rate is low, and requires tuning to get the best performance out of it. The following is a visual representation of setting up an HTTP Accelerator:
One way to improve performance for a database system that performs many reads compared to writes, such as a CMS, is to use primary-replica database replication. Replication requires one primary node and one or more replica nodes. In this setup, all updates are sent to the primary node and reads can be distributed across all nodes. An example use case is increasing the read performance of the database tier of an application. Setting up a primary-replica database replication improves database read performance by spreading reads across replicas, and improves write performance by using exclusively for updates, with no time spent on serving read requests.
Some of the cons for primary-replica database replication are that the application accessing the database must have a mechanism to determine which database nodes it should send update and read requests to. Also, if the primary fails, no updates can be performed on the database until the issue is corrected. It also does not have built-in failover in a case of failure of primary node. The following is a visual representation of a primary-replica replication setup, with a single replica node:
It is possible to load balance the caching servers, in addition to the application servers, and use database replication in a single environment. The purpose of combining these techniques is to reap the benefits of each without introducing too many issues or complexity. Here is an example diagram of this type of server environment set up:
Here is a breakdown of the process when a user sends a request for dynamic content:
When a user requests static content, the following process applies:
This environment still has two single points of failure, the load balancer and the primary database server, but it provides all of the other reliability and performance benefits that were described in previous sections.
Now that you are familiar with some basic server setups, you should have a good idea of what kind of setup you would use for your own application(s). If you are working on improving your own environment, remember that an iterative process is best to avoid introducing too many complexities too quickly.
Thanks for learning with the DigitalOcean Community. Check out our offerings for compute, storage, networking, and managed databases.
This textbox defaults to using Markdown to format your answer.
You can type !ref in this text area to quickly search our full set of tutorials, documentation & marketplace offerings and insert the link!
Click below to sign up and get $200 of credit to try our products over 60 days!
And Google App Engine gives all of that for free, plus
And all of that automatically, you don’t worry how it works inside.
@dan / @kiru42: The easiest way to avoid having the load balancer be a single point of failure is to set up an active/passive HAProxy pair. This would require two HAProxy servers and a virtual/floating IP that can move between the two servers. The active HAProxy server would handle all of the requests unless it went down, at which point the passive HAProxy server would take over the requests. This is a high availability (HA) setup.
DigitalOcean, similarly to many other hosting providers, does not currently support floating IP addresses.
Some people suggest using round robin DNS to workaround not having floating IP addresses, but it is not a true HA setup for a variety of reasons. I do not recommend it because there are situations where it can actually decrease your availability. If you would like to experiment with it anyway, here is a tutorial: https://www.digitalocean.com/community/articles/how-to-configure-dns-round-robin-load-balancing-for-high-availability
@dlazerka whilst we’re at it, why don’t we suck all of the fun out of the web development as well, and just go for a prebuilt system? Recommending people use something out of the tin rather than learning how it works is setting them up for failure. Live a little, learn a little.
#5 could be it’s own article.
lovely article though.
I’m starting in this ultra immense world and I’ve got to say this is, by far, the best post I’ve found.
By the way, can anyone tell me where can I find that network topology icons? I would love to use them in my projects.
Thanks in advance.
PS: Again, GREAT post.
@dlazerka nothing is ‘free’ - yes perhaps Google App Engine might offer it for free to start with, but then as soon as you actually need to scale, and start handing hundreds of thousands of queries per hour, you will soon start paying. Knowing how to set this up using bare metal hardware or cloud instances might save you a lot of money, Google charge a premium for all of this of course.
Then there is the fact that you’re locked into their system, so once things start costing a lot then you will have to jump through hoops to get your app and data migrated and its never easy.
I’m guessing you’ve probably never actually built anything that needs any kind of scale anyway :)
Well explained, thanks!
Informative article. Thanks
Very good article. While researching the best configuration for our production server, I reached this article. So I’m putting my requirement here and see if any expert here could suggest me a best economic hosting plan either in Digital ocean or else where.
We are planning to launch our application in production once all the testing is completed after deployment to our development server. We are using Jenkins CI automation to deploy the code to the development server and we will do the same in production server too, but we are currently doing a feasibility study to understand which Digital Ocean or other hosting options are good for Django Web and Mobile eCommerce app. This app is meant for both consumers and businesses for consumer shopping and lead generation to businesses. That means traffic is anticipated heavily if the app gets popular hopefully.
If I want to implement no. 3, the loadbalancer reverse proxy, and the website is just WordPress based e-commerce website, what should be the app-2?
Assuming the app-1 is the WP website itself.
Where should i install the caching plugins, etc? app-1 or app-2?