Nest Camera Backend Design: Video Streaming, Cloud Storage and Motion Alerts

Quick Overview

A technical product manager question on designing the cloud architecture and technology stack behind a Nest-style home security camera. It tests tracing live video, recordings and motion alerts end to end, rough sizing of storage and bandwidth, and trade-offs in privacy, reliability and cost.

Nest Camera Backend Design: Video Streaming, Cloud Storage and Motion Alerts

Company: Cisco

Role: Product Manager

Category: System Design

Difficulty: medium

Interview Round: Other

As a product manager for a connected home security camera like the Nest camera, you are asked to explain how you would design its cloud architecture and technology stack: everything between the camera mounted at the front door and the app on the owner's phone. Describe the main components, how video and alerts move through them, and the technical choices you would make. ```hint Follow two paths Trace live video and a motion alert separately, from the moment the camera captures them until they reach the owner's phone. Each path has different latency, cost and storage needs. ``` ### Clarifying Questions - Does the camera record around the clock, save only clips around detected events, or offer both as plan tiers, and how long is footage kept? - Are the cameras wired or battery-powered? A battery camera cannot stream or record all day. - Where should detection of motion, people and packages run: on the camera, in the cloud, or split between them? - What scale should the design target (number of cameras, countries), and are there privacy requirements such as end-to-end encryption or keeping data in the user's region? ### What a Strong Answer Covers - A complete component map: device firmware, a secure device connection, video ingest, storage, event processing, notifications and the apps. - Separate treatment of live view, recorded history and alerts, with latency targets and trade-offs such as peer-to-peer versus cloud relay and event clips versus continuous recording. - A rough estimate of upload bandwidth and storage, and how it drives cost and plan design. - Security and privacy: device identity, encryption, household sharing, and where machine learning runs. - Reliability and operations: cameras that go offline, firmware updates, and monitoring of alert latency and camera uptime. ### Follow-up Questions - How would the architecture change if most detection moved onto the camera itself? - The owner's home internet goes down for an hour during a break-in. What does the system still capture, and how does the owner find out? - Storage is the largest cost line. Which levers would you pull, and how would they map onto subscription tiers?

Overview: A technical product manager question on designing the cloud architecture and technology stack behind a Nest-style home security camera. It tests tracing live video, recordings and motion alerts end to end, rough sizing of storage and bandwidth, and trade-offs in privacy, reliability and cost.

Cisco logo
Cisco
Dec 26, 2024
mediumProduct ManagerOtherSystem Design
0
0

As a product manager for a connected home security camera like the Nest camera, you are asked to explain how you would design its cloud architecture and technology stack: everything between the camera mounted at the front door and the app on the owner's phone. Describe the main components, how video and alerts move through them, and the technical choices you would make.

Clarifying Questions Guidance

  • Does the camera record around the clock, save only clips around detected events, or offer both as plan tiers, and how long is footage kept?
  • Are the cameras wired or battery-powered? A battery camera cannot stream or record all day.
  • Where should detection of motion, people and packages run: on the camera, in the cloud, or split between them?
  • What scale should the design target (number of cameras, countries), and are there privacy requirements such as end-to-end encryption or keeping data in the user's region?

What a Strong Answer Covers Guidance

  • A complete component map: device firmware, a secure device connection, video ingest, storage, event processing, notifications and the apps.
  • Separate treatment of live view, recorded history and alerts, with latency targets and trade-offs such as peer-to-peer versus cloud relay and event clips versus continuous recording.
  • A rough estimate of upload bandwidth and storage, and how it drives cost and plan design.
  • Security and privacy: device identity, encryption, household sharing, and where machine learning runs.
  • Reliability and operations: cameras that go offline, firmware updates, and monitoring of alert latency and camera uptime.

Follow-up Questions Guidance

  • How would the architecture change if most detection moved onto the camera itself?
  • The owner's home internet goes down for an hour during a break-in. What does the system still capture, and how does the owner find out?
  • Storage is the largest cost line. Which levers would you pull, and how would they map onto subscription tiers?