Phenomenon of load occurrence

&http://www.aliyun.com/zixun/aggregation/37954.html ">NBSP;

As you apply your application, you will find that while you are carefully "cooking" your application, the difficulty will increase as the load increases. Of course you can use horizontal or vertical expansion, you can also better programming, so that your system can support a larger load. The reduces the number of servers added, as well as improving the reliability and responsiveness of the entire application, This will save you costs . At the same time, this should be the volunteering of excellent engineers.

A large number of objects

Once the number of objects is too large, we all face scaling problems. Obviously, with the increase in the number of objects, resources that can be used for various types of objects will become increasingly stretched.

Failure to recover can cause infinite flow of events

In the case of large network failure, there will be no time to do system recovery, the system is always under the burden.

A large number of high-priority jobs

For example, the redirection of a route is a high-priority activity. If there is a large number of route redirects that can neither be unloaded nor degraded, resources will continue to be consumed to support these high-priority tasks.

Data Flow Increase

As the volume of data increases, the system load will increase. As the request source increases, the system load increases.

function spread (Feature creep)

Vulnerabilities will appear in the system as more than expected features are added.

A sharp increase in client

More clients mean more resources are consumed. More threads are created to drive events. More memory is occupied on the client request queue, more network bandwidth is used for communication, and each client data needs special maintenance.