A data center is a network of specific equipment that is globally coordinated to transmit, accelerate, display, calculate, and store data and information on the Internet network infrastructure.
Introduction
The definition given by Wikipedia is "a data center is a complete set of complex facilities. It not only includes computer systems and other supporting equipment (such as communication and storage systems), but also includes redundant data communication connections and environmental control equipment. , Monitoring equipment and various safety devices". In its book "The Datacenter as a Computer", Google interprets the data center as a "multi-functional building that can accommodate multiple servers and communication devices. These devices are placed together because they have the same pairing Environmental requirements and physical security requirements, and such placement is easy to maintain", and "not just a collection of servers."
When using internet access, you only need a PC with Internet access plus data center software. The router in the article refers to the data center network topology when the general company goes online. It is connected to the Internet through the router. At this time, some settings must be made on the router. It can also be realized by directly connecting the broadband access line to the PC. No ISP is required to activate the service. However, choose the appropriate dynamic domain name software according to the obtained external network IP address and method.
Regarding the address of both ends of the gre, the address of the radius server, and the port address of the enterprise router are automatically obtained when the wireless device dials. There is no need to set up automatic negotiation between the wireless device and the carrier's communication device.
If the data center needs to adopt wireless mode, it can also be realized by H7920 mobile router. But pay attention to bandwidth.
When the dedicated line provided by the operator is used for access, the RJ25 interface is generally used at the user end, and the data center does not require any hardware. A PC is sufficient. There is no need to open any services. But generally when using dedicated line access, APN or VPDN will be used to build the internal private network, so that a fixed IP address can be assigned for easy management.
The specific network structure of the folding is as follows
The wireless DDN system is divided into two parts: a monitoring point and a data center. The monitoring point uses GPRS DTU, which can provide RS-232, RS485, and Ethernet interfaces. The data uses broadband ADSL or dedicated line to access the Internet. The specific access methods are as follows:
a) Monitoring point access method: The monitoring point is connected to the GPRS DTU transmission module through RS-232, RS485 or Ethernet interface, and then DTU related parameters are set. Each GPRS DTU transmission module can be loaded with a China Mobile data SIM card. . b) Data center access mode: Data center access is mainly divided into two categories:
1) Adopt Internet access. The data center adopts broadband ADSL access method, which has the advantages of large bandwidth and economical cost. The disadvantage is that the security is poor and the delay is slightly longer than that of dedicated line access. Use LAN shared Internet access, which is basically similar to ADSL access, but pay attention to port mapping when accessing. Use telephone line dial-up Internet access. In this way, the access bandwidth is relatively narrow. Therefore, it is only suitable for networking with a small number of points and a small amount of data. The network topology is as follows:
2) The mobile company’s access to the data center uses wireless network operators (referring to China Mobile) to provide channel access. In practice, this method has the following types of connections: The mobile company’s computer room dedicated line access method is used, which has a large bandwidth ( Generally 2M), which is very good for delay and security, but it is more expensive to rent this dedicated line. It is especially suitable for users who need high safety factors such as banks and POS network. It is also recommended to use this method when there are a lot of sub-stations. GPRS wireless access method is adopted, which connects a GPRS MODEM in the data center and dial-up Internet access through the PC in the data center. The advantage is that the networking is fast and the cost is economical. The disadvantage is that the bandwidth is narrow (GPRS uplink 10Kbps, downlink 40Kbps), and the delay is large. This method is used more in testing. If you plan to use it in the actual project, please note that the SIM card needs to apply for APN.
Key elements
With the emergence of new technologies such as cloud computing and virtualization, the data center has evolved into a completely different environment. However, any data center needs certain key elements to ensure smooth operation. Whether your data center is only as big as a locker, or as big as an airport, or even a data center that Google is rumored to be building on a floating barge, these elements will apply:
1. Environmental control
A standardized, predictable environment is the foundation of any high-quality data center. It's not just about cooling and maintaining proper humidity (Wikipedia recommends a temperature range of 61-75 degrees Fahrenheit/16-24 degrees Celsius, and a humidity of 40%-55%). You must also consider factors such as fire protection, airflow, and power distribution. I have been in contact with a company that is very serious about ensuring that their data center is kept as pure as possible, and no cartons can be stored in that room. This is due to the distribution mechanism that draws cold air to the front rack, allowing cardboard particles to enter the airflow, which may contaminate the server. This may be a bit extreme but illustrates the importance of this concept.
2. Security
Needless to say, physical security is the foundation of a reliable data center. Keep your system properly, and allow only authorized personnel to enter, and hold the permit card to access the server, applications and data through the network as necessary. It is safe to say that the most valuable assets of any company (except people, of course) exist in the data center. The target of third-rate thieves is laptops or personal mobile phones. Professional thieves target the data center. The door lock can be opened, so it is recommended to use an alarm instead. Of course, alarms may also be useless, so think about the next method: lock the server rack? Prepare a backup power supply for your security system? Hire security? It depends on your security needs, but remember, "Security is a long-term process."
3. Accountability
It should be said that most IT personnel are professional and trustworthy. However, this does not negate the need for accountability in data centers to track human-machine interactions. The data center should record the details of certificate access (and these records should be kept by departments other than IT, such as the security department, or a copy should be kept in the hands of the IT director and the vice president at the same time). Visitors should register when entering and leaving, and are kept under surveillance. The audit of network/application/file resources should be started. Last but not least, there should be a certain person in charge of every system, whether it is a server, router, data center cooling device, or alarm system.
4. Strategy
Behind every process in the data center, there should be a strategy to help maintain and manage the environment. You need system access and usage policies (for example, only the database administrator can fully control the database). You should have a data retention strategy-how long should backups be stored? Do you want to keep them outside the factory? If so, when will it expire? The same concept applies to the installation of new systems, inspection of obsolete equipment/services, and removal of old equipment—for example, clearing server hard drives, donating or recycling hardware.
5. Redundancy
Just as an old car needs a spare tire, your data center may be newer, more expensive, and very important, so you need more than just a spare tire to ensure it keeps healthy operation. You need at least two copies of everything the business depends on, whether it is a mail server, an ISP, a data fiber link, or a VOIP voice phone system. Three or more backups will not hurt in many cases!
In addition to redundant components, the process of testing and ensuring the normal operation of the system is also important, such as regular failover training and research on new methods.
6. Monitoring
Monitoring the uptime and health of the system has forward-looking value, but this is just the beginning. You also need to monitor how much bandwidth, energy, storage, physical storage space is used, and any other "commodities" provided by the data center.
Free tools such as Nagios can perform basic monitoring, and Dranetz can complete more complex solutions such as power measurement. Alarms caused by operational interruptions or low thresholds are part of the monitoring. Make sure to place a failsafe for your alarms so that they can be independent of the data center (for example, if the email server on the VMWare ESX host fails, another system needs Able to monitor and be able to issue notifications).
7. Scalability
Does your company now need 25 servers to complete a series of tasks including virtualization, redundancy, file services, email, database, and analysis? What will you need next month, next year, or next decade? You must ensure that your data center has sufficient scalability to increase power, network, physical space, and storage.
Planning for scalability is not a static thing, it is a continuous process. Smart companies actively track and report on it. These reports can point out the next areas where scalability needs to be met, such as lack of physical storage space.
8. Change management
You might think that change management is part of the "strategic policy". However, I think this is both a strategy and a philosophy. Appropriate change management guidelines can ensure that there are no unplanned events in the data center. Whether it's launching a new system or revoking an old system, the life cycle of all elements in the data center must be consistent with the change management plan.
9. Be methodical
Every IT professional feels that time is pressing. Fear of missing the deadline may lead to some "shortcut" behaviors. Once so, it is often difficult to ensure that the environment is good and tidy.
The realization of a successful system does not only mean installing it and then opening it, it also includes the integration of equipment in the data center in a standardized and technically supportable method. Your server rack should be clean and logical (the production system is on one shelf and the test system is on another shelf). Your cable should be of moderate length and operate according to the cable operation guide, rather than being folded randomly.
10. Documentation
The last point is that you need to have appropriate, useful and timely records. If you do not follow strict procedures, it is easy to cause problems in the implementation. It is not enough to piece together the switch layout and the location of the server plugs into a diagram. Your change management guidelines should include keeping relevant documents and making them available to all relevant personnel as the details are added.
This is not alarmist talk, but guided by the principle of "car accidents". Assuming I am hit by a car tomorrow, other people at least don't have to worry about whether my work or personal files are up to date, because I spend time every week to ensure that all changes and adjustments are recorded accordingly. What's even more exaggerated is that if I decide to change jobs, I don't have to spend two more weeks frantically recording everything I do to the system.
