Score yourself

Give yourself 0, 1 or 2 points per row.

Area	0 points	1 point	2 points
Requirements	Started drawing	Stream to many	Spike speed, latency target, no stalls, score sync and redundancy agreed
Estimates	None	Numbers, not used later	Egress, request rate, origin load and API spike drove the design
API and data	Skipped	Playlist only	Play API with entitlement + signed URL, rolling playlists, score with match time
High-level design	Servers stream video	Encoder + CDN	Redundant upstream, shields, ISP caches, API tier and score path walked
Deep dive	Hand-waved	One topic	Latency trade-offs, request collapsing, pre-scaling or spoiler-safe score done properly
Trade-offs	None named	Named when asked	Raised unprompted (segment length, bitrate caps, push vs poll, degrade features)
Communication	Long silences	Mostly clear	Thought out loud, checked in, managed time

12 to 14: strong hire signal. 8 to 11: solid, polish the weak rows. Under 8: redo the round in a week, timer on.

The bar: fresher vs SDE-2

	Fresher	SDE-2
Video path	Encoder + CDN	Redundant ingest and encoders, packager, rolling playlists, shields
Latency	"Low latency"	Segment length vs buffer, LL-HLS trade-offs, numbers
Peak	"Autoscale"	Pre-scaling, load tests, warm caches, graceful degradation
CDN	"Use a CDN"	Request collapsing, multi-CDN, ISP caches, bitrate caps
Score	WebSocket push to all	CDN-cached polling or in-stream metadata, spoiler-safe by match time

Common mistakes

Relying on reactive autoscaling for a spike that arrives in minutes.
Forgetting the API tier (sign-in, entitlement, home page) while designing only the video path.
One encoder, one origin, for the biggest event of the year.
Pushing every ball to every socket without considering cost.
Showing the score ahead of the video.

Quick check

Five questions on this chapter. Pick an answer, then check it.

quiz

1[
2 {
3  "q": "How do you lower the delay behind live?",
4  "options": [
5   "Longer segments",
6   "Shorter segments, accepting a smaller buffer and more requests",
7   "A bigger origin",
8   "Higher bitrate"
9  ],
10  "answer": 1,
11  "explain": "Shorter segments, shorter delay, smaller cushion."
12 },
13 {
14  "q": "Millions request the same new segment at once. What protects the origin?",
15  "options": [
16   "Nothing",
17   "Request collapsing at the CDN/shield: one request goes upstream",
18   "Rate limiting viewers",
19   "Bigger disks"
20  ],
21  "answer": 1,
22  "explain": "Millions ask, one request goes upstream."
23 },
24 {
25  "q": "When should you scale up for a big match?",
26  "options": [
27   "When CPU alarms fire",
28   "Before the expected spike (pre-scale), since autoscaling reacts too slowly",
29   "After the match",
30   "Never"
31  ],
32  "answer": 1,
33  "explain": "Scale before the walk-out, not after it; switch off the extras when it's tight."
34 },
35 {
36  "q": "Why must a live score widget be careful about timing?",
37  "options": [
38   "Scores are secret",
39   "It must not show a wicket before the delayed video does",
40   "It's slow",
41   "For ads"
42  ],
43  "answer": 1,
44  "explain": "Poll a cached score, and never show what the video hasn't shown yet."
45 },
46 {
47  "q": "In a live streaming system, which tier is usually YOUR problem at peak?",
48  "options": [
49   "Video delivery (the CDN handles it)",
50   "The API tier: login, entitlements, play requests",
51   "DNS",
52   "Encoders only"
53  ],
54  "answer": 1,
55  "explain": "Video is a CDN problem; the API tier is your problem."
56 }
57]

41. Design Live Streaming at Scale (Live Cricket)

Score yourself

Score yourself

The bar: fresher vs SDE-2

Common mistakes

Quick check

Platform

Company

Legal