When it comes to data center availability, the Tier standard is the first response for many people. Many people refer to Uptime's definition of Tier to evaluate their data centers. But what exactly is the standard? "Explained Hu Jiaqing, head of the North Asia Division at the Uptime Institute." Tier1 is the basic ability to meet data centers, mainly single loops, UPS, parallel, and so on. The user's network can be maintained, and the Tier2 will add a portion of the backup device, such as a few UPS and generators, which can be replaced at any time in case of damage. But basically a single loop, but the downtime is short and can take turns, but the trunk may still have to stop for maintenance. Tier3 is a user who can perform network maintenance and repair in any case without downtime. For the data center, the ideal situation is to allow errors. Even fires, earthquakes, tsunamis, etc., do not affect the normal functioning of IT devices. This is what is known as the highest level, that is, Tier4."
At home, in many cases, less than Tier4 of the data center is not redundant UPS or generators, but the lack of two real power sources. Many domestic data center builders believe that as long as electricity from two different power plants is available, two different sources of electricity can be considered. But that's not the definition of Uptime. The Uptime institute belongs to the 451 group, 451 group CEO Martin V.McCarthy explained: "two sets of power does not meet the standard of Tier4, the key problem is not from the two power plants, it is a source of electric power in the world have such a misunderstanding, is not confined to china. In fact, Tier4 is not only the requirement of the power, but for the entire data center to limit fault tolerance mechanism, allowing the fire, allowing short circuit, allowing the emergence of various problems, but does not allow the power data center." However, the domestic electricity is very stable, in some cases compared with foreign countries, can adjust the Tier? Martin answer is simple and firm, the Tier4 standard is in the most harsh conditions can still ensure the normal operation of the data center, the most stringent conditions may never happen, but does not mean that we can ignore the possibility of their occurrence. That's the standard. The user's choice is not to change standards, but to choose standards that are acceptable to them.
Tier4 specifies multiple redundancy, which greatly increases the backup settings while improving availability, leaving little room for energy savings. But the user really need to cut electricity costs, how to balance the two? Martin said, in the world we are not recommended to use the Tier4 standard to the construction of the data center, the user needs to determine their own data center tasks, and the application running on it. These applications are used to determine which standard of Tier is used. It can be Tier3, or Tier2, or even Tier1. Uptime can now authenticate individually at different locations in a data center. It should be noted that if there is a common part, it is necessary to reach a higher standard.
沒有留言:
張貼留言