Design Principles

Type	Definition	Examples
Service Level Objective (SLO)	Objectives your team must hit to meet that agreement. What are the goals you are trying to accomplish for either ourselves or the customer	Durability of disks should be 99.9% Availability of service should be 99.95% Service should successfully serve 99.999% of requests/transactions 99.5% of requests will be completed in 5ms
Service Level Indicator (SLI)	Real numbers (metrics) on system performance. What were chosen to measure progress towards the final goal	Requests latency Availability/uptime of the service Failures per requests Consistency and durability of the data
Service Level Agreement (SLA)	The agreement that you make with your clients. Consequences of failing to achieve your defined SLO	Partial refund of service subscription fee Additional subscription time added for free If 99% of the system requests aren't completed in 5ms, client will get a refund

Availability is the % of time a service or infrastructure is accessible and functioning normally.

$Availability \% = \frac{(Total Time - Total Downtime)} {Total Time} * 100$

Availability % (Nines)	Downtime (Year)	Downtime (Month)	Downtime (Week)
90% (one nine)	36.53 days	72 hours	16.8 hours
99% (two nines)	3.65 days	7.20 hours	1.68 hours
99.9% (three nines)	8.77 hours	43.8 minutes	10.1 minutes
99.99% (four nines)	52.6 minutes	4.32 minutes	1.01 minutes
99.999% (five nines)	5.25 minutes	25.9 seconds	6.05 seconds
99.9999% (six nines)	31.56 seconds	2.59 seconds	604.8 milliseconds
99.99999% (seven nines)	3.15 seconds	263 milliseconds	60.5 milliseconds
99.999999% (eight nines)	315.6 milliseconds	26.3 milliseconds	6 milliseconds
99.9999999% (nine nines)	31.6 milliseconds	2.6 milliseconds	0.6 milliseconds

Availability Type	Overall Availability Trend	Formula	Example
Sequential	Decreases	`AvailabilityServiceA ∗ AvailabilityServiceB`	`99.9% * 99.9% = 99.8%`
Parallel	Increases	`1 - (1 - AvailabilityServiceA) * (1 - AvailabilityServiceB)`	`1 - (1 - 99.9%) * (1 - 99.9%) = 99.9999%`

Reliability is the probability of a service working as expected for a certain time, even under different conditions.

$Mean Time Between Failures (MTBF) = \frac{Total Elapsed Time - Sum Of Downtime}{Total Number Of Failures}$

$Mean Time To Repair (MTTR) = \frac{Total Maintenance Time}{Total Number Of Repairs}$

Power of Two	Approximate Value	Name	Conversion
10	Thousand (1000)	1 KB (Kilobyte)	8 KB
20	Million (1,000,000)	1 MB (Megabyte)	8 KB * 10³
30	Billion (1,000,000,000)	1 GB (Gigabyte)	8 KB * 10⁶
40	Trillion (1,000,000,000,000)	1 TB (Terabyte)	8 KB * 10⁹
50	Quadrillion (1,000,000,000,000,000)	1 PB (Petabyte)	8 KB * 10¹²

Type	Formula	Example
Bandwidth	`KB/message * message/day * days/sec` 10^-5 days/sec	Example: `2.5 KB/message * 100M message/day * 1 day/sec` (`24 hours * 60 min * 60 sec = 86400 sec = 10^-5 day/sec`) Calculate the number: `2.5` Calculate zeros: `0 (from KB) + 2 (from 100M) + 6 (from millions) - 5 (from days/sec) = 10^3` Check the "Power of Two" table: `10^3` Combine all together: `2.5 MB/sec`
Memory	`KB/message * message/day * 20%` 80%/20% cache hits rule: 80% of egress is served by 20% of data stored on the cache servers	Example: `2.5 KB/message * 100M message/day * 20%` Calculate the number: `2.5 (from KB/message) * 2 (from %) = 5` Calculate zeros: `2 (from 100M) + 6 (from millions) - 1 (from % because 20% is 0.2) = 10^7` Check the "Power of Two" table: `10 * 10^6` Combine all together: `2.5 * 10 = 25 GB/day`
Storage	Memory Storage = Throughput (bytes/day) * Retention Period (days) `KB/message * message/day * days * years`	Example: `2.5 KB/message * 100M message/day * 365 days` Calculate the number: `3 (round 2.5) * 3 (365 reduce to 300 and remove zeros) = 9` Calculate zeros: `0 (from KB) + 2 (from 100M) + 6 (from millions) + 2 (from 300 which is a rounded 365 days) = 10^10` Check the "Power of Two" table: `10 * 10^9 = 10 * 10^9` Combine all together: `9 * 10 (that is left from the zeros) = 90 TB`

Uncompressed	Compressed	Compression Ratio	Savings (%)
100 GB	40 GB	`100 / 40 = 2.5` or `2.5:1`	`100 * (1 - 40/100) = 60%`