Building a scalable data platform at Hotstar
Ғылым және технология
In this talk, Jayesh Sidhwani - former head of the data infrastructure at Hotstar - shares the patterns learned in the process of building a scalable data platform. The session covers:
1. Ingestion Patterns
2. Unified Ingestion Proxy
3. Schema Definitions
4. In-flight enrichments
5. Highly Available
6. Storage Patterns
7. Decouple storage and compute
8. Query Lineage & Optimization
9. Noisy Neighbour
10. Consumption Patterns
11. Single GUI and a programmatic interface. All the magic underneath
12. Parity between streaming and stationary data
Session details on hasgeek.com/fifthelephant/win...
Пікірлер: 14
You folks are doing awesome work !
This has been a great talk
Phenomenal talk! I like a lot of the concepts presented!
Great talk!
at 4:57 you are explaining that kafka is highly available and highly durable, you are seriously explaining kafka Characterstics that's it!!1
How much monthly AWS billing during IPL Season or World cup matches?🤔
@abhijithganesh2064
2 жыл бұрын
Must be in millions of dollars, their value spikes
@prashanthb6521
2 жыл бұрын
important question, which not many will answer. LYFT spends 100 million annually in US.
@abhijithalder4996
2 жыл бұрын
the plan we get as a common user, i guess is not the same for all these giant clients, so we can't guess the actual deal until we can become one such client or the budget manager :)
looking at your face it looks like you just woke up!!
The Diagram at 2:27 is not very informative
Amazing talk, loved the engineering. But hotstar got bad UI, cluttered 🤮
The image is too offensive