In modern cloud-native systems, managing state across multiple API calls can be tricky. Imagine an airline ticket purchase:
  1.  List flights & prices
  2.  Reserve a flight
  3.  Complete payment

If the price changes between steps, how do you ensure the user can still complete the process reliably?

Traditional 𝗦𝘁𝗶𝗰𝗸𝘆 𝗦𝗲𝘀𝘀𝗶𝗼𝗻𝘀 can help by routing a user to the same instance, but they come with drawbacks: tight coupling, failover risks, and uneven load distribution.

A more robust solution? 𝗦𝗵𝗮𝗿𝗲𝗱 𝗦𝘁𝗮𝘁𝗲 using Redis or NoSQL:
  • Store temporary state on each API call
  • Persist final state to the main database
  • Handle concurrency with 𝗢𝗽𝘁𝗶𝗺𝗶𝘀𝘁𝗶𝗰 𝗟𝗼𝗰𝗸𝗶𝗻𝗴
  • Ensure 𝗮𝘁𝗼𝗺𝗶𝗰 𝘂𝗽𝗱𝗮𝘁𝗲𝘀 using transactions or Lua scripts
  • Use 𝗶𝗱𝗲𝗺𝗽𝗼𝘁𝗲𝗻𝗰𝘆 𝗸𝗲𝘆𝘀 to prevent duplicate operations

In large distributed systems, adding a 𝗖𝗼𝗿𝗿𝗲𝗹𝗮𝘁𝗶𝗼𝗻 𝗜𝗗 alongside state makes debugging and tracing end-to-end workflows much easier.

Proper TTL, failure handling, and atomicity ensure your solution is 𝘀𝗰𝗮𝗹𝗮𝗯𝗹𝗲, 𝗿𝗲𝘀𝗶𝗹𝗶𝗲𝗻𝘁, 𝗮𝗻𝗱 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻-𝗿𝗲𝗮𝗱𝘆.

Memcached is a distributed, high-speed caching system that stores simple key-value pairs in memory. It’s ideal for quick, transient data storage without persistence or complex features.

Redis is a distributed, open-source, in-memory data store used as a cache, database, and message broker. It supports advanced data structures and features like persistence, replication, and clustering.

Yes, you can configure IDistributedCache in .NET to use either Redis or Memcached as the backend cache provider. This allows you to use the distributed caching infrastructure of .NET while leveraging the power of Redis or Memcached to store the data.