QuestionQ76

VerifiedStoring the data

Your company is running its first dynamic campaign during the holiday season, delivering different offers by analyzing real-time data. The data scientists are gathering terabytes of data that grows rapidly every hour throughout their 30-day campaign. They use Google Cloud Dataflow to preprocess the data and store the feature (signals) data required by the machine learning model in Google Cloud Bigtable. The team observes suboptimal read and write performance for their initial 10 TB data load. They want to improve this performance while minimizing cost. What should they do?

Explanation

Cloud Bigtable requires reads and writes to be distributed evenly across a table's row space for good performance. A row-key schema that spreads access across the key space prevents hotspots and improves throughput without requiring additional capacity. Sequential numeric IDs and a single frequently updated row concentrate traffic and cause hotspots.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!