With the rise and trial application of the Internet, blockchain technology and IPFS, Server lease It has increasingly become the first choice of many enterprises and operators. Choosing servers whose performance configuration does not meet the standard will only bring many operational problems to the enterprise; However, in the absence of actual demand assessment, it will undoubtedly bring cost waste to choose servers with strong performance and high price rashly; Therefore, if you cannot choose a server that is easily called a computing bottleneck to save money, or if you do not fully consider data redundancy, normal business operations will be affected. At this point, enterprise users need to decide which server to rent from different perspectives to find a better balance between meeting technical requirements, business development and cost control. So to sum up, we can summarize the following three questions to get answers:
What applications does the server run?
How many users need to be supported?
How much space is needed to store data?
How important is my business? As for performance, we think we need to evaluate the number of users to decide.
First question: What applications does the server run?
This is the first thing to consider. Here, you should determine the performance, capacity and reliability requirements of the server according to the application type, that is, the purpose of the server. We will discuss it according to the common infrastructure of front-end server applications, server data and servers
stay Web front end: Under normal circumstances, we think that most network front ends of the server have low requirements for the server, such as static network server, dynamic network server, image server, etc. Because under the existing technical framework, we have many solutions to the performance expansion and reliability problems of front-end servers, such as LVS, Nginx reverse proxy, hardware load balancing (F5, A10, Radware), etc.
Application server: It is responsible for computing and function implementation, so we need to select a fast server for the application server based on Web architecture. In addition, application servers may require a lot of memory, especially Ruby, Python, and Java servers based on Windows infrastructure. This type of server requires at least one Xeon configuration. As for reliability, if only one application server is used in your architecture, it must be reliable enough, and RAID is definitely an option that cannot be ignored. However, if there are two or more server applications and a load balancing mechanism with redundancy is designed, we can customize them as needed.
In addition to being an application server in the network architecture, Special applications: It is also used to process streaming video coding, server virtualization, media server, etc. Or as a game server (logic, map, chat), the CPU and memory requirements are also high. At least consider the single channel Xeon server with high storage reliability requirements for server virtualization,
Public services: We refer to servers such as mail server, file server, DNS server and domain controller server. Generally, we will deploy two DNS servers for mutual backup, and the domain controller server will also have a backup server (dedicated or non dedicated), so there is no need to be strict about reliability. As for the mail server, it needs at least enough hardware reliability and capacity, mainly responsible for mail data. Because many users have no habit of saving and archiving mail data, when they reinstall the system, they always rely on downloading data on the server. With regard to performance issues, it is necessary to evaluate the specific number of users to make a final decision.
Database: The last application we discussed is also the most demanding and important server. Whether you use MySQL, SQLServer or Oracle, in general, we think it needs fast CPU, large memory and stable and reliable hardware. A single Xeon CPU/4GB RAM/Rad1 is absolutely an entry-level configuration. As for the exact configuration, we need to discuss the business requirements again before making a decision.
Second question: How many users do enterprises need to support to rent servers?
If an enterprise wants to rent a house on a server, how many users do it need to support the selected server? We know that when we choose servers and rent houses, servers must provide certain services, and users of these services must also be considered. You need to evaluate several specific questions:
1. How many registered users?
How many users will access online at the same time under normal circumstances? What is the peak of daily synchronized Internet access?
These problems have a crucial impact on our decision to use what kind of CPU and how much memory we use. It is recommended that your technicians and business departments sit together to discuss these issues, and finally convert these numbers into
Some more specific technical figures, such as the number of concurrent connections (in many cases, the number of users is not the same as the number of connections). At the same time, you also need to make a prediction and plan for future user growth as accurately as possible. Your server needs to support more and more users.
2. How much space is needed to store data?
How much space is needed to store data? We need to calculate this problem from two perspectives. One is what types of data are available, including the space occupied by the operating system itself, the space required to install applications, the data generated by applications, databases, log files, mail data, etc. If it is a Web2.0 website, you also need to calculate the storage space of each user;
Another point of view is from the time axis. These data are growing every day. You should at least make an accurate calculation of the data growth in the next year (suggested 2 to 3 years for Micronet), which may still require your software developers and business personnel to provide sufficient information. Finally, you still need to multiply the calculated numerical results by a factor of about 1.5 to facilitate various data backup and file transfer operations during maintenance.
Reference: When an enterprise chooses to rent a server, the server parameters mainly include the following:
(1、CPU: CPU represents the computing power of the server. If the website traffic is large, if the website traffic is large and there are many dynamic pages, it is recommended to select more than 2 cores of CPU.
(2) Memory: Memory is a transit station for data and an important factor in determining the speed of website opening. The larger the memory, the larger the available cache, and the faster the opening speed. Choose the appropriate space according to the size of your website. If it is a general personal blog or an enterprise display website, you can choose a smaller space. If it is a mall, the news category should have a larger space.
(3) Hard disk: The hard disk is a place to store data. The size of the hard disk should be determined according to the data size of the website, and the remaining space should be taken into account when selecting. In addition, the I/O reading speed of the hard disk directly determines the speed of file reading. The reading speed of the server hard disk is much faster than that of other websites, which is generally sufficient.
(4) Bandwidth: Select the appropriate bandwidth according to the nature of the website, and look at your website type, structure, traffic and other indicators or estimates. Generally, there are not many pictures of new websites, but they have also been processed. 2M is enough. As long as the bandwidth is real, not much, enough, and appropriate; Or, select 2M first, and then add if it is not enough, which is more stable. If it is an ordinary website. For example, text websites, photo websites, forums, etc., occupy very little bandwidth. 5000 visits may also take up less than 2M bandwidth. For video, download and other websites. Maybe hundreds of people need more than 10M of bandwidth to be online at the same time.
(5) Line: The choice of lines is sometimes very difficult for us. If we were targeting local users, we could have chosen a single line server, but the advantage of a dual line server is generally dual IP, and for a city, broadband is different from Telecom, Unicom, and Mobile. Take Shenzhen as an example. There are telecom, China Unicom and mobile users. Therefore, single line or double line can not be satisfied at this time. Choosing BGP multi line is a better way to achieve user experience. (BGP multiple lines One IP multiple lines choose their own optimal path access)
(6) Operating system: The choice of operating system is related to personal familiarity and the specific situation of the website. You can choose an operating system if you know more about it. In addition, Windows system supports asp programs well, but it takes up more memory. Linux system is more friendly to php programs, saving more memory, and some programs may only support a certain operating system.
Finally, after the enterprise decides to choose the server lease, it should decide according to its own business needs and the number of users when choosing the server lease. I hope this article will be helpful to you!!!
