In this era, many companies will choose to use cloud server, briefly say the choice of cloud server, I hope to bring you reference.

In general, the three parameters of cloud server are price, core number and memory. Some cloud service providers also provide a number of technical white papers for reference, which most buyers rarely read.

The core number

There is a difference between the number of cores and the number of physical cpus. In the early days, physical cpus had only one core, so the use effect can be regarded as the same. But multicore cpus are not uncommon these days.

  • A dual coreCPUThe ability to run two processes/threads at the same time.
  • Simultaneous multithreading and hyper-threading /HT are both technologies that increase the number of threads a single core can execute at the same time.

Check the number of physical cpus

cat /proc/cpuinfo| grep "physical id"| sort| uniq| wc -l
Copy the code

Check the number of cores per physical CPU:

cat /proc/cpuinfo | grep "cpu cores" | uniq
Copy the code

Check the total number of logical cpus.

cat /proc/cpuinfo| grep "processor"| wc -l
Copy the code

View the CPU model:

cat /proc/cpuinfo | grep name | cut -f2 -d: | uniq -c
Copy the code

General return

1  Intel(R) Xeon(R) CPU E5-2680 v4 @ 2.40GHz
Copy the code

Check whether the CPU is 64-bit:

The CPU op – mode (s)

The above information is displayed at the same time

lscpu
Copy the code

Return roughly as follows

Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                1
On-line CPU(s) list:   0
Thread(s) per core:    1
Core(s) per socket:    1
Socket(s):             1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 79
Model name:            Intel(R) Xeon(R) CPU E5-2680 v4 @ 2.40GHz
Stepping:              1
CPU MHz:               2394.448
BogoMIPS:              4788.89
Hypervisor vendor:     KVM
Virtualization type:   full
L1d cache:             32K
L1i cache:             32K
L2 cache:              256K
L3 cache:              35840K
NUMA node0 CPU(s):     0
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon rep_good nopl xtopology pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe  popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm rdseed adx smap xsaveopt aratCopy the code

memory

In the case of Web services, according to most synchronous threading models, each connection needs to consume the corresponding memory, the larger the memory, the greater the number of concurrent access, it is a basic common sense, of course, this is not true, a server internal resource consumption and competition make it impossible to increase the number of concurrent access indefinitely.

Suggestions for purchasing a cloud server

For Spring Boot, which is popular in JVM architectures, the ratio of servers purchased is generally recommended to be 1:2 or 1:4. It can make full use of both core computing performance and memory to achieve a balance of computing, memory and price.

For vert. x, Quarkus and other asynchronous frameworks, the memory requirements are relatively low. You can choose 1:1 and 1:2 servers to make full use of CPU performance and reduce memory waste.

The most important thing is to buy on demand

Talk about cloud server from actual business is playing hooligan. The most important thing to look at is how resource-intensive the application you deploy on the server is, whether it’s high IO or computationally heavy.

During a moderate stress test, you can observe the memory and CPU usage to know whether the current system load is insufficient memory, insufficient CPU computing capability, or a BUG.

Most cloud platforms can be configured to upgrade and will always find their own cloud server configuration.

Current cloud server usage

After purchasing the server, the cloud server is actually under low load most of the time, which is actually a waste. Although many cloud platforms offer flexible scaling, it is not a perfect solution.

In recent years, emerging Serverless, cloud native, cloud functions and other attitudes to provide new solutions, worthy of attention.

Network Packet Receiving and Sending (PPS)

The PPS refers to the number of packet packets sent per second, which is related to the CPU.

Indicator – main frequency

CPU Computing capability.

memory

Applications need running memory.

The core number

It relates to the parallelism capability of the system.

review

The performance of cloud servers with the same specifications of different service providers is also different, but the impact is relatively small.

As for why the common specifications on the market are 1:1, 1:2, 1:4, in fact, there is a certain relationship with the characteristics of the application of daily deployment, relatively speaking, it can play the performance of most machines.

In this era, many companies will choose to use cloud server, briefly say the choice of cloud server, I hope to bring you reference.

In general, the three parameters of cloud server are price, core number and memory. Some cloud service providers also provide a number of technical white papers for reference, which most buyers rarely read.

The core number

There is a difference between the number of cores and the number of physical cpus. In the early days, physical cpus had only one core, so the use effect can be regarded as the same. But multicore cpus are not uncommon these days.

  • A dual coreCPUThe ability to run two processes/threads at the same time.
  • Simultaneous multithreading and hyper-threading /HT are both technologies that increase the number of threads a single core can execute at the same time.

Check the number of physical cpus

cat /proc/cpuinfo| grep "physical id"| sort| uniq| wc -l
Copy the code

Check the number of cores per physical CPU:

cat /proc/cpuinfo | grep "cpu cores" | uniq
Copy the code

Check the total number of logical cpus.

cat /proc/cpuinfo| grep "processor"| wc -l
Copy the code

View the CPU model:

cat /proc/cpuinfo | grep name | cut -f2 -d: | uniq -c
Copy the code

General return

1  Intel(R) Xeon(R) CPU E5-2680 v4 @ 2.40GHz
Copy the code

Check whether the CPU is 64-bit:

The CPU op – mode (s)

The above information is displayed at the same time

lscpu
Copy the code

Return roughly as follows

Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                1
On-line CPU(s) list:   0
Thread(s) per core:    1
Core(s) per socket:    1
Socket(s):             1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 79
Model name:            Intel(R) Xeon(R) CPU E5-2680 v4 @ 2.40GHz
Stepping:              1
CPU MHz:               2394.448
BogoMIPS:              4788.89
Hypervisor vendor:     KVM
Virtualization type:   full
L1d cache:             32K
L1i cache:             32K
L2 cache:              256K
L3 cache:              35840K
NUMA node0 CPU(s):     0
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon rep_good nopl xtopology pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe  popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm rdseed adx smap xsaveopt aratCopy the code

memory

In the case of Web services, according to most synchronous threading models, each connection needs to consume the corresponding memory, the larger the memory, the greater the number of concurrent access, it is a basic common sense, of course, this is not true, a server internal resource consumption and competition make it impossible to increase the number of concurrent access indefinitely.

Suggestions for purchasing a cloud server

For Spring Boot, which is popular in JVM architectures, the ratio of servers purchased is generally recommended to be 1:2 or 1:4. It can make full use of both core computing performance and memory to achieve a balance of computing, memory and price.

For vert. x, Quarkus and other asynchronous frameworks, the memory requirements are relatively low. You can choose 1:1 and 1:2 servers to make full use of CPU performance and reduce memory waste.

The most important thing is to buy on demand

Talk about cloud server from actual business is playing hooligan. The most important thing to look at is how resource-intensive the application you deploy on the server is, whether it’s high IO or computationally heavy.

During a moderate stress test, you can observe the memory and CPU usage to know whether the current system load is insufficient memory, insufficient CPU computing capability, or a BUG.

Most cloud platforms can be configured to upgrade and will always find their own cloud server configuration.

Current cloud server usage

After purchasing the server, the cloud server is actually under low load most of the time, which is actually a waste. Although many cloud platforms offer flexible scaling, it is not a perfect solution.

In recent years, emerging Serverless, cloud native, cloud functions and other attitudes to provide new solutions, worthy of attention.

Network Packet Receiving and Sending (PPS)

The PPS refers to the number of packet packets sent per second, which is related to the CPU.

Indicator – main frequency

CPU Computing capability.

memory

Applications need running memory.

The core number

It relates to the parallelism capability of the system.

review

The performance of cloud servers with the same specifications of different service providers is also different, but the impact is relatively small.

As for why the common specifications on the market are 1:1, 1:2, 1:4, in fact, there is a certain relationship with the characteristics of the application of daily deployment, relatively speaking, it can play the performance of most machines.