How can you maintain product scalability?

Simple is elegant. Keeping solutions modular, decomposable and simple when you design a solution one needs to keep in mind. Scalability != Complexibility.

Most important thing is cost, you can put 100s and thousands of machines, storage and memory. But cost constraints would make it difficult to scale and would require different strategies to scale. How much consistency is desired is User experience fine with a data that may have possibility of being in- consistent for 1 second , 60 second or 5 minutes The consistency and availability would define the cost of Scalability. Absolutely directly proportional. For example most content applications need not hit App servers and can be served from CDN cache while most fintech apps would need to hit App sever everytime. Thus content App scaling may not have huge cost difference.

Scalable architectures must not be conflated with very complex architectures. In today’s world, it is considered normal to design extremely complex systems from day one. We will have myriads of micro services developed by multiple orgs within the enterprise. These will be scaled independently and are designed to cater to “number of concurrent consumers on a Google scale”. In reality this kind of architecture increases complexity and leads to bloated infrastructure requirements. The key is to design it in such a way that you can evolve the architecture over time. Deployment architect is especially crucial.

Most Scalable architecture the severs and data stores should be stateless and be able to route perform required operation. Think about the topmost layers, how would they scale, for example how would load balancers scale ?? How would API gateways scale. Think about bottomes layers DB, what may be the rate of Scalability, would you want multiple masters or split between read and write between master slave is good enough. Think about scaling of, Data size itself, for NoSql dbs consistent hashing works without much hassle but ACID COMPLIANT DB one may have to worry about sailing of database layer to accommodate increase in size of db. There are very few ACID Compliant DB that provided horizontal scaling without redistributing.

Besides the above considerations, scalable code needs to decouple itself from deployments. It cannot assume a deployment topology since the latter will change with time to meet growing scalability requirements.

Scalable low level designs are very important, meaning your schema should be able to incorporate new data and without redoing original schema. Normalization is a way to go. The connection pulling, closing the unused connections. Size of each thread and objects stored in each thread. It's always better to use distributed caches to reduce redundant data in threads. The use of protocols to communicate across microservices should be thought. Persistent protocols like grpc, http2 , websocket are good for async call which helps in better utilization of resources. But at the same time large number of persistent connections can quickly become bottleneck.

Scalable user Interface, should allow UX team to be able to change user flow without making change on the backend. Employ right API strategy that gives control to front end to team to change data flow or add new data to UX. APIs request should allow capability to query all related schema for an object. Develop caching on UX and CDN caches, to reduce round trips. Token should not be tied to used flow rather the data being accessed updated.

Modular Design - Adopt a modular approach in product design. This allows for easy adaptation and scaling of the product to meet changing market demands or technological advancements. By creating components that can be easily upgraded or replaced, you ensure long-term relevance and flexibility of your product. Focus on User Feedback - Regularly gather and analyse user feedback. This insight is crucial for identifying areas for improvement and potential expansion. By staying attuned to user needs and expectations, you can adapt your product to evolving market trends, ensuring its scalability and sustained success.