PROFESSIONAL-CLOUD-DATABASE-ENGINEER · Question #203
Your IoT monitoring platform ingests sensor data from 100,000 devices, generating 2 million writes per second during peak hours and 500,000 writes per second during off-peak hours. Each record is…
The correct answer is A. Implement row key sting wth hash prefix, configure automsted sealing with an optimal min max. Adding a hash (salt) prefix to the row key evenly distributes writes across tablets and nodes, eliminating hotspotting. Enabling autoscaling with appropriate min/max limits adjusts capacity to peak (2M writes/sec) and off-peak loads while keeping read latency under 10 ms…
Question
Your IoT monitoring platform ingests sensor data from 100,000 devices, generating 2 million writes per second during peak hours and 500,000 writes per second during off-peak hours. Each record is 1KB with a row key format: deviceId#timestamp. Your current 12-node Bigtable cluster experiences hotspotting on specific device ranges, with three nodes handling majority ofthe traffic while others remain underutilized. Read queries typically access the last 24 hours of data for specific devices. Storage costs are becoming a concern as you retain two years of historical data (500TB total). You need to optimize performance, eliminate hotspotting, and reduce storage costs while maintaining read latency under 10ms. What should you do-
Options
- AImplement row key sting wth hash prefix, configure automsted sealing with an optimal min max
- BSplit the table by device type, use separate Bigtable instances for different geographic regions,
- CChange the row key format to deviceId#timestamp, implement automated scaling with an
- DIncrease cluster size to 20 nodes, implement row key salting with hash prefix, and enable
How the community answered
(20 responses)- A80% (16)
- B10% (2)
- C5% (1)
- D5% (1)
Explanation
Adding a hash (salt) prefix to the row key evenly distributes writes across tablets and nodes, eliminating hotspotting. Enabling autoscaling with appropriate min/max limits adjusts capacity to peak (2M writes/sec) and off-peak loads while keeping read latency under 10 ms. Offloading cold data (older than 6 months) to Cloud Storage reduces Bigtable storage costs without affecting the 24-hour operational queries. Exam Questions, Study Guides, Practice Tests. Lead the way to help you pass any IT Certification exams, 100% Pass Guaranteed or Full Refund. Especially Cisco, Microsoft, CompTIA, Citrix, EMC, HP, Oracle, VMware, Juniper, Check Point, LPI, Nortel, EXIN and so on. Our Slogan: First Test, First Pass. Help you to pass any IT Certification exams at the first try. You can reach us at any of the email addresses listed below. Any problems about IT certification or our products, you could rely upon us, we will give you satisfactory answers in 24 hours.
Topics
Community Discussion
No community discussion yet for this question.