The principle factor which decides the uptime for a colocated server is power.sudden power can damage hardware parts of the servers like mother board,RAM and hard disk power failure can knock down network at times only about 5% colocation service providers have right power equipments most them have equipments which lack redundancy in case one unit fails multiple units manage the load but if it lacks redundancy it will over load the other units and chance of their failure are also possible every thing including ups,power distribution unit,generators,circuit breakers must be redundant and run below capacity .
General problems with colocation service provider
Insufficient generator capacity
Generator capacity should more 1.5 times more than actual load so that it insure safe working without overloading colocation services should be provide with a redundant back up generator in case of main generator fails there should be place where switching of generator can be done fast and easily most of the colocation service provider use low capacity generators at start and with growth of company they keeps on adding new generators this makes multiple points for generators and mostly don’t have common access this increases the shut down time and generators should be well maintained and regularly serviced.
Transfer switch failure
Most of the colocation service provider use mechanical transfer switches to save cost and these are not trustworthy as circuit breakers there should be redundant transfer switches because if transfer switch fails means network down.
Lack of non Redundant UPS
UPS is power provider until generator is is switched on for power supply if UPS did not work properly and quickly even with quality UPS failure is an issue so you must make sure that no UPS gets overloaded and runs below 50% of its capacity and max to max load should 75% and all units should be serviced regularly.separate power grids and multiple UPS should be designed .
There should redundant parts used in colocation services providing companies so that life of equipment increases and well-maintained and redundant components running below capacity at every part of the system it don’t effect network uptime
Comment