Solutions Architect Series – Part 5: Avoiding anti-patterns in solution architecture

This is my learning note from the book Solutions Architect’s Handbook written by Saurabh Shrivastava and Neelanjali Srivastav. All the contents are mostly distilled and copied from the book. I recommend you to buy this book to support the authors.

Another series: Fundamentals of Software Architecture: An Engineering Approach

Often, the teams can drift away from best practices due to timeline pressure or the unavailability of resources. You always need to give special attention to the following architecture design anti-patterns:

In an anti-pattern (an example of a poorly designed system), scaling is done reactively and manually. When application servers reach their full capacity with no more room, users are prevented from accessing the application. On user complaints, the admin finds out that the servers are at their full capacity and starts launching a new instance to take some of the load off. Unfortunately, there is always a few minutes’ lag between the instance launch and its availability. During this period, users are not able to access the application.

With anti-patterns, automation is missing. When application servers crash, the admin manually launches and configures the new server and notifies the users manually. Detecting unhealthy resources and launching replacement resources can be automated, and you can even notify when resources are changed.

With anti-patterns, the server is kept for a long time with hardcoded IP addresses, which prevent flexibility. Over time, different servers end up in different configurations and resources are running when they are not needed. You should keep all of the servers identical and should have the ability to switch to a new IP address. You should automatically terminate any unused resources.

With anti-patterns, an application is built in a monolithic way, where all layers of architecture including web, application, and data layers are tightly coupled and server dependent. If one server crashes, it brings down the entire application. You should keep the application and web layer independent by adding a load balancer in between. If one of the app servers goes down, the load balancer automatically starts directing all of the traffic to the other healthy servers.

With anti-patterns, the application is server bound, and the server communicates directly with each other. User authentication and sessions are stored in the server locally and all static files are served from the local server. You should choose to create an SOA, where the services talk to each other using a standard protocol such as HTTP. User authentication and sessions should be stored in low latency-distributed storage so that the application can be scaled horizontally. The static asset should be stored in centralized object storage that is decoupled from the server.

With anti-patterns, a single type of database is used for all kinds of needs. You are using a relational database for all needs, which introduces performance and latency issues. You should use the right storage for the right need, such as the following:

  • NoSQL to store user session
  • Cache data store for low latency data availability
  • Data warehouse for reporting needs
  • Relation database for transactional data

With anti-patterns, you will find a single point of failure by having a single database instance to serve the application. Wherever possible, eliminate single points of failure from your architectures. Create a secondary server (standby) and replicate the data. If the primary database server goes offline, the secondary server can pick up the load.

With anti-patterns, static content such as high-resolution images and videos are served directly from the server without any caching. You should consider using a CDN to cache heavy content near the user location, which helps to improve page latency and reduce page load time.

With anti-patterns, you can find security loopholes that open server access without a fine-grained security policy. You should always apply the principle of least privilege, which means starting with no access and only giving access to the required user group.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.