Katkat
Back

Design a topic-based news subscription system

OptiverSoftware Engineer · Online Assessment
HardSoftware EngineerOnline AssessmentSystem Design

Last updated

This question evaluates system design skills for building a topic-based news subscription service, focusing on streaming ingestion, relevance filtering, sliding-window rate limiting, de-duplication, ordered delivery, and data structure choices for in-memory operation.

Design and implement a topic-based news subscription service for a news aggregation platform. Provide a class NewsProvider supporting the following operations and constraints:

Operations

  • AddSubscription(subscriberId, minInterest, maxNewsPerSecond, topics) -> bool • Create or update a subscriber’s configuration. • Return true on success; return false on invalid input or violated constraints.
  • RemoveSubscription(subscriberId) -> bool • Remove a subscriber; return false if the subscriberId does not exist.
  • NewsReceived(newsId, timestamp, interest, topics) -> bool • Ingest a news item with its timestamp, interest score, and topics. • Return false if newsId already exists; otherwise true.
  • Publish(nowTimestamp, maxAge) -> dict[int, list[int]] • For each subscriber, return the list of newsIds to deliver at nowTimestamp, subject to the rules below.

Publish rules

  • Eligibility: news interest ≥ subscriber.minInterest.
  • Topic filter: news topics intersect subscriber.topics.
  • Freshness: nowTimestamp - maxAge ≤ news.timestamp ≤ nowTimestamp.
  • Rate limiting: per-subscriber sliding window allowing at most maxNewsPerSecond items in any 1-second rolling window.
  • De-duplication: the same newsId must not be delivered to the same subscriber more than once.
  • Ordering: higher interest first; if ties, older timestamp first; if still tied, smaller newsId first.

Deliverables

  • Specify data structures for: subscriptions, news storage, delivered-tracking per subscriber, and per-subscriber sliding-window counters.
  • Describe algorithms for filtering, rate limiting (sliding window), de-duplication, and multi-key sorting under high throughput.
  • Analyze time and space complexity of each operation and Publish.
  • Discuss concurrency, idempotency, and persistence considerations.
  • Explain how you would scale this to millions of users and a large, continuous news stream (partitioning, sharding, backpressure, batching, and fault tolerance).