Typical Commercial Linux Support Costs
* Typical Linux System Failure Schedules
* 1% - 2% die during assembly and initial power up
* 2% - 5% die during 30-day high-load burn-in tests
* 1% - 2% die during first year of service
* Avoidable Linux Support Issues (all are 95% preventable)
* bad admin (doesn't know how to debug simple problems)
* releasing untested systems, software and procedures
* bad assembly (loose screws, loose wires, metal-to-wire vibrations
* loose wires on the patch panel, loose power cords
* ambient temperature too high (should be 65F or less)
* inadequate fan provision which causes the system to overheat
* bad (cheap) memory ... use high-quality memory
* dead power supply ... (bad vendors/manufacturers, overworked PSU)
* random failures ... bad parts bought from the wrong vendors
* General RELATIVE Linux Support Costs
* £1 == Relative Cost to fix something in R/D stage
* £2 == Relative Cost to fix something at Manufacturing stage
* £4 == Relative Cost to fix something at Final QA/Shipping/Testing
* £8 == Relative Cost to fix something at Customer site
* £16 == Relative Cost to fix something after receiving returns
* Ancillary Linux Support Costs
* Airline, Hotel, meals, cars
* Costs of support servers, testing scripts
* Costs of managers, meetings, paperwork, etc
* Costs of Offices, office admin, invoicing, overhead