Consistent Hashing medium
When adding a node to the ring, data must move. What happens to latency during this transfer?
- A Latency drops immediately as the cluster capacity increases, resulting in a sudden burst of improved response times.
- B Latency remains entirely unaffected because data movement occurs within a dedicated, background high-speed fiber ring.
- C Latency typically increases as the network and disk I/O are saturated by the 'bootstrap' process of streaming data to the new node.
- D The cluster becomes temporarily read-only for all geographic regions until the tokens have fully converged across the gossip layer.
Answer: C. Latency typically increases as the network and disk I/O are saturated by the 'bootstrap' process of streaming data to the new node.
Scaling out a distributed database is resource-intensive. The new node must 'stream' historic data from its clockwise neighbor to become functional. This 'Bootstrap' process competes with production traffic for disk bandwidth and network throughput, often causing a measurable latency spike. High-performance systems throttle this background streaming to protect user experience.
One system design question a day, with spaced repetition so the ones you miss come back.
Get the appRelated questions
- In a 'Smart Client' architecture (e.g., modern Cassandra drivers), how does the client find the right node?
- When a new node (Node X) joins the ring, where does it acquire its initial data from?
- Google's Maglev hashing algorithm (used in their Network Load Balancers) differs from standard Consistent Hashing rings. How does it generate the lookup table?
- Virtual Nodes improve balance but increase metadata size (Memory). How does 'Multi-Probe Consistent Hashing' achieve good balance without VNodes?
- In the Chord DHT protocol (a P2P consistent hashing implementation), naive lookups are O(N). How does Chord optimize this to O(log N)?
- Which of these systems is a famous user of Consistent Hashing?