Seven precautions for server hardware upgrade

Hardware upgrade: the server upgrade focuses on four aspects: CPU, memory, storage system, and nic.

CPU: the CPU is the central processor. The English value is centralprocessingunit. CPU is the core component of a computer. It is only as big as a matchbox and as thick as dozens of sheets of paper, but it is the computing core and control core of a computer. The CPU is responsible for reading commands, decoding and executing commands in all operations on the computer. Many servers need to be upgraded. The CPU runs too slowly, resulting in limited application performance.

During the upgrade, you should be clear about the server architecture, CPU interfaces, and the maximum number of CPUs that can be supported. Currently, the widely used PC servers include Intel and AMD architecture servers, as well as some low-end servers using desktop CPU. The CPU architecture used by these servers is different. Therefore, when upgrading the CPU, you must first know which server architecture the enterprise server uses, then, check whether there is space for upgrading the CPU on the server. However, before upgrading the CPU, you must specify the two concepts: Transaction processing speed and concurrent processing performance.

In addition, if the number of customers accessing the server is small, but each customer needs the server to provide a certain type of CPU-dependent Application Service, a high-speed single CPU may be the most useful. However, if many users make access requests to a large volume of data at the same time, when these access requests are opened in an independent process or thread mode, even a low-speed multi-CPU system may be more useful.

Memory: the memory stick is the channel connecting the CPU and other devices, and serves as a buffer and data exchange function.

The maximum memory capacity supported by the server is generally limited. If you want to upgrade the server's memory, you need to first understand the related technical parameters of this server to avoid failure to meet the requirements during the upgrade.

Storage System: the storage system upgrade of servers mainly involves two aspects: Capacity Expansion and RAID Array Construction to achieve double performance or double data security, or even both.

If you are planning to upgrade, check whether the server motherboard meets the relevant requirements. View compatible brand models.

NIC: Nic is also a very important component for servers. The NIC throughput is directly related to the overall performance of the server. If the server bottleneck lies in the NIC, the CPU performance is superior, the larger memory and more hard disk capacity are all virtualized. Therefore, when upgrading the server, you need to pay attention to whether the NIC has become a bottleneck.

Seven precautions for upgrading:

1. Users should fully evaluate their needs and find appropriate technologies.We can neither blindly follow the sales staff, nor simply buy products with advanced technologies. If you choose expensive products, you should choose the right product. What suits us is the best.

2. Users should fully understand the hardware compatibility so that the products they purchased can be well compatible.During the upgrade process, a lot of waste often occurs. For example, although you buy the most advanced CPU, your server does not support it, resulting in unnecessary waste.

3. Pay attention to price traps.We need to fully understand the path of "making money and selling goods". In our country, low-price sales are often the traps that manufacturers send to customers. Some users are very price-sensitive, so they often make a move.

4. If the upgrade is caused by a shortage of resources, such as a slow CPU or a small memory, which limits application performance, we recommend that you expand resources based on the original platform architecture.Such as faster CPU, more memory, larger disk capacity, and more CPU, do not migrate between Unix minicomputers, PC servers, new 64-bit systems, and high-performance servers, or even change the CPU type and operating system.

5. If the reason for the upgrade is that the user's requirements cannot be met due to the original architecture, such as the number of CPU cores and memory size of the PC server, or the storage capacity of the 4cpu UNIX Minicomputer is limited, we recommend that you upgrade the system capability on a similar system.For example, to upgrade a 2-way PC server to a 4-way or 8-way, a 4-way Unix system to a 8-way or 16-way, and change SCSIRAID to a FCSAN.

6. If the reason for upgrading is that the user's needs cannot be met due to the application architecture, you can consider changing the system architecture;If the performance of all applications running on one SMP system is limited, the cluster system can be used to distribute applications to improve scalability. For example, if the database is affected by the I/O bandwidth on the PC server, the transaction processing capability cannot be guaranteed and can be upgraded to a large SMPUnix system.

7. If the reason for your upgrade is cost saving, you can refer to the following approximate formula;Currently, the price ratio of servers is similar to that of high-end PCS: Xeon servers: UNIX servers =; in addition, the cluster system: Large SMP systems =. You can select a platform based on the funding situation.

Summary: as mentioned in the above article, hardware upgrade is one of the following, software-optimized servers also allow old machines to take advantage of the heat (for example, using virtualization technology can increase the utilization of our servers ). At the same time, pay attention to the fact that the current hardware update speed is much higher than the software update speed. When purchasing Server products, we need both flexibility and hardware. Try to think about the problem two years later and make an estimate so that the server can continuously meet application requirements.