Scalability is the capability of a system, process or a network to manage/handle the load of growing demands. A distributed system will be considered as scalable system, if it can able to continuously evolve to handle increasing amount of works.
So as per the above definition now we can say that the system has to scale because of many reasons like increased amount of transactions, increased data volume (where you can scale by sharding/partitioning the data).
Note: Designing/Architecting Scalable System should not cause any loss into performance of the system.
A distributed system will be considered reliable if it keeps delivering services even when one or more components fails (software or hardware components).
Reliability in distributed system can be achieved by having redundant software and hardware and of course to build reliable and resilient system you have to pay extra cost.
We can take an example of any e-commerce application where any running transaction should not get cancelled because of the failure of the machine that is running the transaction. For example adding an item to shopping cart, the expectation is system should not loose the item and in case of any failure another running server should able to replace it.
Copyright @ b-transform.com. All Rights Reserved