High concurrency and distributed _ Distributed computing

When it comes to high concurrency, a lot of people think about distribution, so what's the difference?

Concurrency and distribution are completely different concepts.

Concurrency is reflected in the amount of time, such as video online live site, there may be tens of thousands of people need to access the server at the same time, this is concurrency.

Distribution is to distribute tasks to different points, and distributed computing is the most common distributed computation. Through a kind of distributed programming method, the CPU, memory and so on on different systems are computed, and the results are collected into the control center for processing. For example, the most famous is distributed computing weather climate, such as aerospace information. The synchronous accelerator in Shanghai needs the distributed system, and the distributed computing method is used to compute the model.

For the specific domain of the Internet, the high concurrent solution does not necessarily need to pass through the distributed system, can through DNS polling, to achieve the burden of server load sharing. All servers are the same configuration, access to the same content. But for some database processing system, need to return the result quickly, also can distribute the computation task to the different system through the distributed system, speed up the computation speed. But for the internet, high concurrency can also be solved in a distributed way, such as the concept of cloud, cloud storage needs to be distributed through the system, can achieve high concurrency of storage. This needs to depend on demand and tolerance.