Five Things You Need to Know About a Carrier Grade Server
The hardware and software that powers the network must be able to perform under some very extreme conditions
Jul. 8, 2013 08:00 AM
The telecom industry provides mission-critical services, which means that the hardware and software that powers the network must be able to perform under some very extreme conditions. Servers, in particular, must be extremely reliable and function during and after everything from streams of heavy traffic to massive earthquakes. Carrier grade servers have to go through rigid testing to receive certification, and they have several characteristics that make them very important.
They Are Part of the National Infrastructure
Telecommunications, as a whole, is considered an important part of the national infrastructure. This means that any disasters that derail the service are a huge problem for everyone involved - from emergency response personnel to government officials to everyone who just wants to call and check on the people they know. This infrastructure has to be operational before, during, and after an event, which is why these servers are tested so ruthlessly. Carrier grade servers should never be the weak link that drops the ball in an emergency.
NEBS Is Tough but Important
The Network Equipment Building System (NEBS) guidelines are extremely stringent and not subject to any kind of leniency. It will take time for the tests to be completed on any new carrier grade server, and the equipment will probably be destroyed in the process. (This happens when one of the tests is to see how well it reacts to fire.) However, these guidelines are based on some directive from the FCC, and when the products can pass these tests, providers can confidently use them and have a reasonable expectation for almost constant uptime.
They Should Be Able to Shake Off an Earthquake
It may sound strange, but these products can't be considered carrier grade until they can handle an 8.2 earthquake. These seismic tests are usually done pretty early in the process, and most products seem to pass the first time without much trouble. Nevertheless, designers must consider many different elements, from taught cables to power failures, as they iterate the next generation of servers.
Compatibility with Legacy Infrastructure Is Still Possible
Carrier grade systems are often assumed to be proprietary technology that are designed for specific purposes and therefore incompatible with other servers already at work. While this has been true, there are new systems that are compatible with legacy servers and operating systems, which make them a more cost-effective option for providers. These new OEM servers provide the necessary levels of reliability, performance, and customization while still keeping costs low enough to be a valid option for emerging and rural markets.
Extreme Uptime Requirements
Industries that provide mission-critical services can suffer a lot when their servers go down. Some estimates claim that a server outage could end up costing a million dollars for every hour that it's down. Given that a normal, average server is capable of hitting 30 to 300 minutes of unplanned downtime a year that could translate into millions of dollars every year.
Carrier grade servers are designed to cut out downtime as much as possible. The requirements published by Telcordia, in fact, only allow for three minutes of downtime per year from planned and unplanned causes. There's not a lot of room there for accidents, and the goal is to always maintain around 99.999% uptime.