“Bypass” refers to a mechanism that allows two networks to be physically connected directly without passing through the network security equipment system under specific trigger states (power failure or system crash). Therefore, with BYPASS, when the network security equipment malfunctions, the networks connected to this device can still be connected to each other. Of course, at this time, this network device will no longer process the packets in the network.

Bypass in medicine means “a replacement tube used in catheterization surgery to allow blood or other body fluids to bypass a blocked or diseased organ”.

For example, a gastric bypass (gastric replacement tube). "Bypass" also means "ignore" or "pay no attention to". In this article, unless otherwise specified, "BYPASS" all refers to one of the modules of bandwidth management equipment, meaning "bypass unit". It refers to when network equipment fails (including both normal and abnormal failure modes. Normal failure, such as manual switching and other methods. Abnormal failure has a wide meaning, such as hardware failure, the number of connections exceeding the threshold, etc. There are great differences among different products. This article will not describe it further). The BYPASS module has two application modes: internal and external.

Whether internal or external BYPASS is to avoid network interruptions caused by single points of failure. Once these situations occur, BYPASS will be activated immediately and ensure the unimpeded flow of the network. The design requirements for network interruptions in some networks are very strict, even calculated in * minutes / year.

The advantage of the built-in BYPASS is the small size of the device, which makes it easy to deploy. The advantage of external BYPASS is a separate device to serve as a BYPASS module. Generally speaking, the external BYPASS module is required to be designed as a passive mode of operation, that is, the external BYPASS does not require a power supply to work properly, which is initially intended to better ensure stability and reliability.

We know that any device that requires power support will definitely be damaged, it's just a matter of time. Looking at these electrical products around us or the equipment used on the network, almost without exception, they all have this potential hazard.

According to the above diagram, you will find that in the external BYPASS mode, it is a passive BYPASS module that is truly connected to the network. And the passive BYPASS itself provides two sets of interfaces. One set is connected to the network, and the other set is connected to the host. The host is the one that needs power supply to work.

Let's imagine that when the hosts in both internal and external forms are damaged respectively, the network is unimpeded. At this stage, there is no difference between internal and external yet. But next, we must replace the damaged equipment. At this time, the problem arises. For internal equipment, since the BYPASS module is solidified inside the equipment, when we replace the equipment, we must interrupt the network again. But for external equipment, it is not necessary because the host is not connected to the network. Even if the equipment is replaced, there is no need to interrupt the network. This is also the biggest difference between the two modes.

Generally speaking, enterprise-level network interruption recovery requirements are relatively loose and can accept the internal BYPASS form, but it is not absolute. It should be flexibly selected according to the user's design goals.