10 key elements in establishing an efficient IDC data center

With the advent of new technologies such as cloud computing and virtualization, the evolution of data centers may turn itself into a very different environment. However, any data center that operates smoothly and successfully always requires some basic elements. Whether the scale of the data center is like a vertical wardrobe or a plane, or even rumors that Google is building a yacht-type data center, these elements are critical.

1. Environmental control

A standardized, predictable environment is the cornerstone of any High-quality data center. This is not just about cooling the equipment and maintaining proper humidity (according to Wikipedia, the recommended temperature range is 61-75 degrees Fahrenheit/16-24 degrees Celsius, 40%-55% humidity), you also need to consider the fire extinguishing, air flow and power distribution and other factors. A company I used to work to ensure that the data center is as spotless as possible and that no cardboard boxes can be stored in the room. This is because the cardboard particles may enter the airflow, potentially polluting the server, because the distribution mechanism is to send cold air to the front of the rack. This example may be extreme, but the importance of this is illustrated.

2. Safety

Needless to say (but anyway I would have said), physical security is the foundation of a reliable datacenter. Keep your system in a tight open shutdown state, allowing only those authorized to enter, in conjunction with the necessary access to only the data that is transmitted to the server, the application, and the network. It is certain that the most valuable asset of any enterprise is in the data center. Three-stream thieves will steal http://www.aliyun.com/zixun/aggregation/9600.html "> laptops or Mobile phones." Professional thieves will gaze at the data center. The door lock can be overcome, so I suggest using an alarm at the same time. Of course, alarms can also fail, so think about your next step: Lock the server rack, too? To provide standby power for your security system? It depends on your security needs, but remember, "Security is a journey, not an end." ”

3. Accountability

As a system administrator, I can prove that most it people are professional and trustworthy. However, I do not deny that there is a need to have a certain accountability system in the data centre to track its behaviour. The data center should record login information through access (and I recommend that these logs be managed by a department outside the IT department, such as the security Department, or that a copy of this information is kept in the hands of the IT director and Vice President). Visitors should sign in or log out and remain under supervision at all times. A review of networks, applications, and documentation resources should be initiated. Finally, each system should have a certain owner, whether it be server, router, Datacenter cooler or alarm system.

4. Strategy

Each process involved in the data center should have a set of policies behind it to ensure environmental maintenance and management. You need the policies that the system accesses and uses (for example, only the database administrator can fully control the SQL Server). You should also have data retention strategies-how long do you keep backups? Are you leaving the station to save these backups? What if the backup content expires? The same applies to installing new systems, checking obsolete equipment or services, and dismantling old devices-for example, erasing a server hard drive, donating or recycling hardware.

5. Redundancy

The first car I owned was a blue Ford Pinto. At that time it cost my parents 400 dollars, the oil cost is 1 dollars per gallon, I drive it everywhere. It has a spare tyre that often comes in handy. I say this not for nostalgia, but for a point of view: Even my vintage car has some redundancy. Your data center should be more sophisticated, more expensive and critical, so you need more than just a "spare tire" to keep it healthy. Everything you need to keep your business running should be kept at least two copies, whether it's a mail server, an ISP, a data fiber link, or a VoIP phone system virtual machine. Three or more will not hurt in many cases!

Not only are redundant components important, but the process of testing and ensuring their stability is also important-for example, regular fault drills, new methods, and so on.

6, monitoring

Monitoring the uptime and monitoring conditions of all systems will bring significant positive value, but this is only the beginning. You also need to monitor how much bandwidth is used, as well as power, storage, physical rack space, and anything other data centers provide as a "commodity."

There are also free tools like Nagios to monitor some basic details, and more complex solutions such as Dranetz measuring power. The process also includes a power outage or low threshold alarm--to make sure your alarm is insured, this can be independent of the data center (for example, if your e-mail server is on a VMware ESX host and the host fails, then another system should monitor this situation, Be able to give timely notice).

7, scalability

So today your company needs 25 servers for a range of tasks including virtualization, redundancy, file servers, e-mail, databases, and analytics. So what do you need next month, next year, or next ten years? Make sure you have an appropriate scale of data center, enough capacity to increase power, network, Physical space and storage. If your data center needs are growing--if your company is profitable--and that's certainly the case--start planning today.

Planning for scalability is not the only thing, this is an ongoing process. Wise companies will actively track and report this idea. I have seen a number of references in these reports, which speak of the need to fill the gaps in the scalability of critical areas as quickly as possible.

8. Change Management

You might say that change management should be grouped under "strategy", but I would say it's both a strategy and an idea. Appropriate change management guidelines will ensure that your data center, which has not yet planned, discussed, and agreed to provide failure measures or Plan B, will not have any problems. Whether you introduce new systems or eliminate old systems, the lifecycle of all your data center components should be consistent with the prospect of change management.

9, decorate

All the IT professionals I know are not time-critical. The deployment of a new system may result in some details being cut off due to painful deadlines – details that always seem to include something that makes the environment neat and beautiful.

Successful system implementation does not only mean access and opening, but also integrates devices into the data center through standardized and supported methods. Your server racks should be clean and orderly (the production system is in a rack and the test system is in another rack). The cable length is moderate and follows the wiring guide instead of being randomly laid.

10. Document Record

The last point is appropriate, useful, and timely documentation-if you don't follow the procedure strictly, you may be ignoring this in the implementation process. Just listing a bunch of diagrams about switch deployment and server access is not enough, and your change management guidelines require that documentation should be relevant and that all the details involved should be available.

It may sound a bit perverted, but I've always adhered to the "car hit" rule. If I get hit by a car tomorrow, you don't have to worry about whether my work document or personal document has been updated, because every week I make sure that all changes and adjustments are recorded accordingly. It's no exaggeration to say that if I decided to change my job, I wouldn't spend two weeks on the system.

Full details

The advantage of these ideas is that they are completely unrelated to hardware or software. Whether your datacenter contains servers running Linux, Windows, or other operating systems, or just a collection of network switches and mainframes, hopefully this will be useful for you and your business.

Combine all of these together, imagine your IT environment as a wheel, data center is the hub, these 10 elements are the surrounding "tires."