"I think we can analyse the data that's collected especially the users who are requesting it.if we find that there are set of users who regularly use this service then we may cache the data for those users to speed up processing and better the experience. The fresh data (which will be a delta on the cached data can only be calculated) and the stale data can be left out thereby updating the cache"
Nilanjan D. - "I think we can analyse the data that's collected especially the users who are requesting it.if we find that there are set of users who regularly use this service then we may cache the data for those users to speed up processing and better the experience. The fresh data (which will be a delta on the cached data can only be calculated) and the stale data can be left out thereby updating the cache"See full answer
"I started off with defining the users of the system. From them i dervied the functional requirements and categories into MoSCoW. Then for the DB part, Userprofiles, App data to be stored in a RDBMS. The binary apps to be stored in a scalable, distributed DB such as S3. The app data from RDBS to have a link to S3 URLS.
Then i touched upon Partition tolerance while scaling, Replication, Caching at multiple levels, Load balancers at multiple levels.
This being my first System design interview i"
Anand K. - "I started off with defining the users of the system. From them i dervied the functional requirements and categories into MoSCoW. Then for the DB part, Userprofiles, App data to be stored in a RDBMS. The binary apps to be stored in a scalable, distributed DB such as S3. The app data from RDBS to have a link to S3 URLS.
Then i touched upon Partition tolerance while scaling, Replication, Caching at multiple levels, Load balancers at multiple levels.
This being my first System design interview i"See full answer