What is isolated stack in Cloud 🌨️

Yes world is more and more interconnected. Users and application are seperate different geographic location. Latency is one the most crucial part here and we have to keep that in mind while creating architectural deployment. So latency optimized based solution is required. Nowadays Users are moved from desktop to mobile they have gotten accustomed to having access to their applications and data anywhere with quick response time.

So Running an application across multiple region gives low latency and provides high availability. It improves the BCM/DR.

What is isolated stack

Where user are routed to the region where their data resides that particular request is fully processed in that region. If the application data is partitionable into separate databases, then one deployment option some applications have considered is to partition or shard the application and data across multiple regions into separate isolated stacks.

So each stack would use the Primary Region with Failover Region approach. So user’s data is resides to a single regional deployment based on the sharding or partition.

Here we have two region AA and BB. Host resolution for region AA aa.example.com and for region BB BB.example.com
So analysis says two close region will have 10–20 ms round trip and some far away continents may have 100–200 ms round trip.

Applications that are dedicated to run within a single region, this gives potentially higher availability by sharding user data across multiple regions and better end-user latency for the users close to the regions hosting their data.

Sharding the user data across regions can reduce the number of users impacted due to an application change that affects just a single region.

Adding to this, this model can be used to meet jurisdictional requirements for keeping data within a given region.

Core issue with isolated stacks

1. have to deal with failover and loss of availability when a single region has issues.
2. distributed denial of service (DDoS) attacks and large traffic spikes being directed at a specific region, since user requests go to a specific region and cannot be load-balanced across multiple regions
3. user experience as a function of where their data is located (i.e., if a user’s data resides in region AA, then it is always routed to region AA no matter where the user is in the world)