Carrier Grade refers to networks or infrastructure that have extremely high reliability, and is both well-tested, reliable, and redundant. There are five major attributes of carrier-grade: High availability/serviceability; Quality of Service/ Quality of Experience; Performance and Efficiency; scalability; and security. Many different kinds of networks can be “carrier grade,” including PSTN, VoIP, Ethernet, and even the hardware that makes up carrier grade networks.
The most important feature of a carrier-grade service is that it has “five nines” or even “six nines” reliability. This means there can be a maximum network unscheduled downtime of only a few minutes per year. This is also combined with fault-tolerant design and fast redundancy (less than 50 milliseconds to react to faults). With proper redundancies, some networks can legitimately lay claim to having 100% uptime. If a single component fails, there is a natural disaster, or if there is an attack, the network must be durable enough to withstand it. If a component does fail, it must be easy to debug or replace it. In the event of failure, nodes and other components automatically restart, recover, and reconfigure themselves in a self-healing manner.
A second feature of carrier grade networks is the quality of service. Voice and video technology must have as little lag and/or as much redundancy to prevent jitter or other artifacts which are, at best distracting, and at worst, make communication impossible. The term “quality of service” is also used for the algorithms that make sure that VoIP packets get priority over non-Real Time Traffic (RTT), which is why the term “quality of experience” is sometimes used instead.
The third feature of a carrier grade network is performance. The right combination of hardware and software make carrier grade equipment run at a level that is far above consumer grade electronics. Carrier grade servers and other infrastructure is always the newest processors, made from the finest materials, optimized to run with multiple cores, made to cool those cores, and run at peak performance levels as much is required. Amazingly, these servers are often about the size of two blu-ray players stacked on top of each other. The energy demands of modern computer equipment is extremely high, and that basic measurement is often doubled because it takes as much energy to run the fans and coolants. Engineers are always looking for new ways to make components smaller and cooler.
Carrier grade networks must be scalable. This can refer to an individual component that is sometimes, but not always, running at its highest levels; it can refer to a particular location that collocates equipment to balance the load; or it can refer to adding entire server farms on the other side of the country or the world in order to expand business opportunities and also to add redundancy.
The final component of carrier grade infrastructure is security. As with network reliability, the data and the service must be kept safe from DoS attacks, break-ins and theft, ID theft attempts, and viruses.
Because there is so much overlap within these categories, other sources may list the defining characteristics of carrier grade under different criteria. But the most basic definition is this: a carrier grade network has nearly 100% uptime, is durable enough to withstand natural and man-made attacks, and is scalable.