Prime Infrastructure is just a tool that allows you to monitor your network and do some other things that may assist in improving availability. That would be things such as managing deployments via a centralized console with workflow checks and balances built in vs. via multiple staff logging into individual devices and typing away at the command line.
Achieving an SLA of whatever level requires the system be designed implemented and operated according to those goals. Things like redundancy, using high quality equipment, restricting physical and logical access to the equipment, providing UPS systems etc. all go into achieving an SLA target.
You need to also define your terms. For example 99.99% ("four nines") means about 50 minutes of allowable downtime during one year. Does that include only unplanned outages or are planned system outages also included in your metric. If the latter, then redundancy at most layers is mandatory. Does the SLA mean every network drop must be available? what if one cable to one use is cut? For that user the network is unavailable. If it takes more than 50 minutes to repair it, your SLA is not met. That may be an extreme example, but what if the power to one IDF closet where a single access switch is lost? Would that count against your SLA?
