980
79
You are looking for a data center to host your critical environment. A, so called, 2N datacenter; redundant power feeds, redundant fiber routes that are physically separated and enough cooling capacity to guarantee a stable inlet temperature. With your environment up and running an outage appears whilst you thought you were safe, when choosing a data center.
One of the main reasons for choosing a data center, is to limit the risk of downtime and unavailability of the company's critical environment. A data center offers redundant power feeds, multiple power sources (main grid and emergency generators) and redundant fiber paths to make sure one feed/source and path will always be available. So far, a data center can guarantee a certain uptime. The uptime guaranteed often guarantees the availability of at least one feed/source or path; in terms of data center design: N. Does that ensure the uptime of your environment? To maximize your environment 's uptime, the resources a data center delivers must be used.
When choosing a high-standard datacenter, the equipment you will use in that data center must be able to use the safeguards that a data center offers. The infrastructure of power, fiber paths and cooling are as strong as the weakest link. For example, when using a server that is only connected to one feed, the guaranteed uptime on one of the two power feeds do not apply anymore for the power on that server. When using a fiber connection from one fiber path, that fiber path is the single point of failure, although the data center has two redundant fiber paths. The same for using an ATS, when using it, the power to the ATS is dual feed, however the power path behind the ATS will be the single-point of failure. To achieve the highest uptime, you must use the safeguards that a data center offers as most as possible.
Using a data center that guarantees uptime and connecting your devices correctly, so the safeguards are used, does that guarantee the uptime of your environment? Unfortunately, not. The server is now the single point of failure. The next step is optimizing your environment itself. Are hypervisors used in your environment over multiple servers? Are you using other clustering techniques? Are you sure there is no single point of failure within your storage system or within your network? To really use the advantage of the uptime guaranteed by a data center, your full environment must be designed so it can also guarantee the uptime of your environment.
Even if the design is right, the human error can make your design worthless. Connecting both power supplies of a server to the same power feed, connecting 2 connections of one fiber path to one server and two fiber connections of the other path to another server; those mistakes can happen and can cause unnecessary downtime. Procedures and cross-checks can help to reduce the risk of human error.
Companies can spend lots of money to expensive data centers and using servers with a single power supply or the use of one switch with a single power supply to connect servers with two power supplies. Take care of all the aspects within your infrastructure, even the procedures supporting cross-checks, and the highest uptime can be guaranteed to your environment.
