Micro-environment services, my business logic is no longer in a single process, but rather dispersed in a lot of progress. Orders, logistics, inventory, price. Each is a tomcat process.
Every process, every tomcat has its own entry point. Then led me to prevent a large attack surface than the original more. Then the risk will be higher.
Performance issues, the original business logic are all in the same process there. So I need information security are also in there. After such a request comes in, I want to experience what the user identity, authority. I have been inside this process can be completed.
In the micro-architecture of the service, I need information about the likely security inside me this process is not, for example, to access an order of service time, the information, the user's information rights in this process which, the tomcat which is out of reach, I may need a remote call. I call the security center, certification center to obtain information relevant. So not a request, whether it is external to come, or should do security check between services. So do check depend on this remote links may result in delays caused service performance problems. Especially for the performance of extremely sensitive service. The service itself may Originally, I responded to a few milliseconds. I now I have to do a remote call to verify the safety, added a few milliseconds. Increase the performance of a few milliseconds for extreme service, his response time has doubled. This time, this problem is also micro-service performance challenges.
Between security services, the original time need only consider external incoming request is not safe, come in the future, from the logistics transfer orders, transfer orders from inventory, this is in my old tomcat inside, without any consideration of safety The problem.
But in the micro-service scene, when I went to call the logistics from order, in fact, I need across the network. Out of my process into another process. This time I will ensure that this communication is secure. This is also the challenge of micro service
requests across multiple micro-services, it is difficult to track. For a service observability, it is a very important indicator. Observability log includes three log,
Distributed inside each process will record their own separate log, the log is dispersed to remember. Order your own diary, inventory your own diary. At this time I need a mechanism to string together all the logs.
The log aggregation up. The most typical is the flow control of the entire service of blank, rather than a single one control service.
A request to enter time spent on each service
container of immutable is a very important principle
multi-language architecture, each service can use their own micro appropriate language to create.
End