Answer Key: Bigtable

Back to Exercises

Exercise 1: Design Keys

Question: Design row keys for a time-series application. How do you avoid hot spots?

Answer

Goal: Design keys for even distribution and time-series queries.

Key Design

Good Design: {metric}:{hash}:{timestamp}

Example:

cpu_usage:abc123:20240101100000
cpu_usage:def456:20240101100000
memory_usage:abc123:20240101100000

Components: - Prefix: Metric name (logical grouping) - Hash: Hash of device/entity ID (distribution) - Timestamp: Time component (ordering)

Benefits: - Even distribution (hash component) - Supports time-range queries (timestamp component) - Co-locates related data (prefix component)

Answer: Use {metric}:{hash}:{timestamp} format. Hash component ensures even distribution, timestamp enables time-range queries.

Exercise 2: Handle Hot Spots

Question: Your table has hot spots. How do you fix them? What key changes do you make?

Answer

Hot spot causes: Sequential keys, time-prefixed keys, small key space.

Solutions:

  1. Add Hash Component: {prefix}:{hash}:{suffix}
  2. Reverse Timestamp: Use reverse timestamp for recent data distribution
  3. Add Random Component: Add random component to keys
  4. Redesign Keys: Redesign key structure for better distribution

Example: Change from {timestamp}:{device} to {device}:{hash}:{timestamp}.

Answer: Add hash component to keys, reverse timestamp ordering, redesign key structure for even distribution.

Exercise 3: Optimize Performance

Question: Your queries are slow. How do you optimize them? What column families do you use?

Answer

Optimization strategies:

  1. Optimize Column Families: Group related columns together
  2. Design Keys: Design keys for even distribution
  3. Monitor Compaction: Monitor and tune compaction
  4. Optimize Queries: Reduce data scanned, use appropriate filters

Example: Use separate column families for frequently vs rarely accessed columns.

Answer: Optimize column families, design keys for distribution, monitor compaction, optimize query patterns.