Data and storageIntersections
10 million users a day. Why can’t HyperLogLog count who came both days?
0%
Page 39
More questions.
Data and storageLive drivers
Assume 500,000 drivers ping every 4 seconds. Where do their locations live?
0%
Data and storageHot keys
One cricket match’s chat hits 5,000 messages a second. What breaks?
0%
Data and storageBRIN
How can a 6 MB index stand in for a 44 GB one?
0%
Data and storageSQL vs NoSQL
DynamoDB has transactions. PostgreSQL shops run sagas. So is BASE just NoSQL?
0%
Data and storageLong transactions
A two-hour report only reads. Why can’t the database clean up behind it?
0%
Data and storageConnection pools
Why can moving to more databases triple the connections on the old ones?
0%
Data and storageBuffer pool
Why is a 99% cache hit ratio ten times slower than 99.9%?
0%
Data and storageNear real-time
Elasticsearch acknowledged your new product. Why can’t anyone search for it yet?
0%
Data and storageOptimism
When 90% of tries collide, why does a commit average 10 tries?
0%
Data and storageHybrid search
Why does vector search hand you an iPhone 16 case for “iPhone 15 case”?
0%
Data and storageDoc values
Why can’t the index that finds Khadi Co kurtas sort them by price?
0%
Data and storagePage cache
Why should a 64 GB search node keep its heap under 30 GB?
0%
Data and storageArchitecture
Search finds the kurta. Why does the database still quote its price?
0%
Data and storageData tiers
Why do 90 days of Bazaar’s logs need 37 TB of disk, not 180?
0%
Data and storageCrawling
How does Google crawl the web without knocking small sites over?
0%
Data and storageClick bias
Why does a ranker trained on clicks keep whatever sat first on top?
0%
Data and storageScale
How long would one grep over all of GitHub’s code take?
0%