I work in the field of network security, and currently my main job is to provide cloud DDoS protection services to foreign governments and enterprises.
We use a self-built cloud solution, setting up our cleaning equipment in multiple nodes around the world and establishing internal network connectivity to achieve load balancing across multiple nodes. This allows us to handle large volumes of traffic, exceeding Tb-level, for cleaning.
Perhaps all companies that started with hardware face some challenges when they want to establish cloud services. Initially, the hardware design may not be sufficient for cloud construction, or there may be a lack of corresponding hardware maintenance systems when running on the cloud. While individual devices are good products, there are some shortcomings when it comes to achieving cluster deployment and management.
Our company's products are no exception. The company specializes in DDoS protection, and our flagship equipment is the DDoS protection system. However, when I joined the company and started working on maintaining our self-built cloud, I discovered these issues. I have never been involved in software development in a meaningful way, only having some experience with lightweight code development. For agile development, our hardware-based company is still somewhat slow.
The domestic network in China can be seen as a large-scale LAN, with three major operators each responsible for their own networks. There are not many problems within the country. However, on the international level, it is more open and free, and each operator has its own unique problems. These problems converge on our network, creating a collection of issues. But for each problem, there is always a solution. From finding a solution to actually deploying it online, it takes at least three quarters...
I am not against the development processes set by the company. It's just that for this type of cloud service, we are novices and need to learn more. We need to shorten the development process as much as possible to solve customer problems in a short time.
But then again, a shorter development process means more resources need to be invested. If we are only solving a problem for one customer, the company may reduce investment because it does not correspond to the income. This creates a vicious cycle. Demand is the primary productivity, and without demand, there is no need for investment. This is currently the operating strategy of many companies.
Many companies have emerged from the process of constantly stumbling and struggling. The products presented to customers have gone through many iterations and already meet the needs of the majority. As for the few exceptional requirements, the company will discuss whether it is necessary to invest further in development. It's all about weighing the risks.
I am just a low-level employee, with less than two years of work experience, and still lacking in experience. I may not have a comprehensive understanding of the problems I see, and I still need some time to learn. I don't expect to learn a lot from just one company, but I strive to make progress every day.