Latency Numbers

Access Type	Time	Converted Time
CPU cycle	.3 nano secs	1 secs
CPU L1 cache	1 nano secs	3 secs
CPU L2 cache	3 nano secs	9 secs
CPU L3 cache	13 nano secs	43 secs
Main memory (RAM)	120 nano secs	6 mins
SSD	150 micro secs	6 day
HDD	10 milli secs	12 months
San Franscio to New York	40 milli secs	4 years
San Fransisco to Australia	183 milli secs	19 years

Desk

Latency Key Takeways

• Avoid network cache whenever possible
• Replicate data across data centre for disaster recovery as well as performance
• Use CDNs to reduce latenc
• Keep frequently accessed data in memory if possible rather than seeking from disk, caching

Map Reduce IRL