Unicom network management protocol block diagram

Because the protocol framework is the same / but the transmission methods involved are different, the service framework of Google RPC C++ cannot be used, and as the basic protocol, it belongs to the protocol of the application layer, only handwritten code, the real southbound network management

p>

Because all manufacturers use the same set of protocols, but the actual implementation of internal logic is different, so in equipment maintenance and network management maintenance, the problems involved are the most. If a unified framework logic is used, the problem of network management protocolOccurs, also reduces the response

The protocol itself is not difficult, the protocol is very perfect, all the situations have been considered, just need to be implemented

Domestic operators do not know whether they will use NetConfig as a network management solution for repeater base station BBU in the future

The rest is to analyze timers, exception handling, alarm handling, and design considerations for compatibility

The polling method is used as a timer (there are too many ways to implement poll_timeout)
Alarm processing uses the bitmap algorithm (all kinds of alarms will be easily shielded, and the performance will be improved)
Requirements for the entire network management, If it is a BBU, base station, RRU, AU, EU, the business models involved are few and single, and the development difficulty is lower
If it is a real monitoring and management center, the development difficulty is much greater, and the off-the-shelf is generally selected.The high concurrency framework is not handwritten by yourself, and the business models involved are diverse

I make a low-concurrency monitoring and management center, which is used for equipment debugging, testing, management, problem collection, and can also practice conventional algorithms, (hash map sort stack list queue ...) less business models, developmentIt is also easier to set up, but it is time-consuming. As a second personal project, it is suitable (x86_64 platform, with less restrictions. The first arm platform is actually for the convenience of being compatible with all hardware interfaces, management, and the underlying processing is ok, applicationThere are obvious shortcomings in the processing, and there are traces of over-design, not clear, the framework model is out, but it has to be given, and probably, the C++11 standard syntax, in the application layer software, should be dig deep)

Characteristics of the linux kernel, object-oriented, conventional algorithms, the process things are already familiar, hard-core things, gradually deepening from the basics, conventional algorithms -> core algorithms
linux kernel driver subsystem + network subThe system is my goal
Driver framework + data structure + chip specification + conventional algorithm = the core of embedded software development

Students of pure application software focus on upper-layer applications, so they have high requirements for algorithms, but they are actually to optimize and deal with complex business models
Students of embedded development are the link between the top and the bottom, the focus of work,It can be hardware-oriented or software-oriented, but it is a low-level development, because there is no complex business model, so the algorithm requirements are not high.
With the in-depth study, it is found that the bottom layer of the system and the designed business model are very complex, and no logic is redundant., so it's also the stage of learning the conventional algorithm + core algorithm in the kernel