A process is the basic unit for the operating system to allocate resources (CPU time, memory), and a thread is the basic unit for scheduling execution. A thread must belong to a process, and a process can contain multiple threads.
Nginx and redis are common multi-process models, while Tomcat and memcached are multi-threaded models.
- Multi-process data sharing is complex and requires communication mechanisms such as pipes, signals, message queues, shared memory, sockets, etc., but data synchronization is simple, while multi-threads share process data segments, data sharing is simple, but synchronization is complex.
- Multi-process occupies a lot of memory, complicated to create, destroy and switch, and the CPU utilization rate is low, while multi-thread takes up less memory, and the creation, destruction, switching is simple, and the speed is fast.
- Multi-process programming and debugging is simple, and the processes will not affect each other. It is suitable for multi-core and multi-machine distribution. It is simple to expand to multiple machines, while multi-thread programming and debugging are complicated. One thread hangs up and may affect the entire process. It is suitable for multi-core distributed .
If you need to create and destroy frequently, need a lot of calculations, switch frequently, and use threads first when multi-core distribution is required, it is suitable for selecting processes when you need to be more stable and secure, and support multi-machine distribution.
| Contrast dimension |
multi-Progress |
Multithreading |
Summarize |
| Data sharing and synchronization |
Data sharing is complex and requires IPC; data is separated and synchronization is simple |
Because of the shared process data, data sharing is simple, but it is also because of this reason that synchronization is complex |
Each has its own advantages |
| memory, CPU |
Occupies a lot of memory, switching is complicated, and CPU utilization is low |
Low memory usage, easy switching, and high CPU utilization |
thread dominance |
| Create, destroy, switch |
Creating and destroying, switching is complicated, and the speed is slow |
Creating and destroying, switching is simple and fast |
thread dominance |
| programming, debugging |
Simple programming and easy debugging |
Complex programming, complex debugging |
process dominance |
| reliability |
Processes do not affect each other |
If a thread hangs, it will cause the whole process to hang |
process dominance |
| distributed |
It is suitable for multi-core and multi-machine distribution; if one machine is not enough, it is relatively simple to expand to multiple machines |
Suitable for multi-core distributed |
process dominance |