InterviewBit Problems   Design Search Typeahead

About the Design Search Typeahead category (1)
How can we modify the trie so that reads become super efficient? (5)
What would be the schema for storing the trie ? Would we use SQL or no-SQL databa (5)
Search Typeahead How does our client know which application servers to talk to. How does it know which application servers have gone down and which ones are still working? (2)
How many queries per second should the system handle? (6)
Why do we need Trie? Can we simply shard on the basis of whole word? (1)
Search Typeahead How do we handle the case where our application server dies? (2)
Unclear numbers crunching (1)
How can no of queries per day be 100 billion / day (3)
Using redis as ds for typeahead feature (3)
What if I use a separate trie for updates and copy it over to the active one periodically? (4)
How much data would we need to store? (6)
Can frequent writes affect read efficiency? (4)
Offline update? (2)
How would a typical write work in this trie? (5)
Can we use sampling? (5)
Capitalization and special characters (1)
How does client know when update query is to be called? (2)
But How would we store Trie in DB? (4)
Atomic "copy" of the Trie (1)
Search Typeahead What would the API look like for the client? (4)
Would all data fit on a single machine? (2)
Why would a writer thread need to take a lock on a trie in order to avoid inconsi (2)
This is not a good design (2)
(more points): 2. You could send the common part to the client itself, including (3)
Search Typeahead Is Latency a very important metric for us? (1)
Search Typeahead How important is Consistency for us? (1)
Search Typeahead How important is Availability for us? (1)
Search Typeahead How would a typical read query look like? (1)
Search Typeahead How would a typical write query look like? (1)