as shown in the figure, the current situation is:
Note: these three systems can be regarded as micro-services connected through the HTTP interface.
requirement: no matter external callers call our business system, account system, or internal business system call account system, they must verify the system through the interface.
current situation:
if you go 1, that is, the external caller invokes the account system after being verified by the interface.
if you go to 2p3, that is, the external caller invokes the business system after being verified by the interface.
if you go 4pm 5, the performance is the internal business system calling the account system.
that"s all right.
but if the external caller leaves, that is, the external caller invokes the business system after passing the interface verification, and then functionally requires the business system to call the account system, then the design needs to do interface verification twice, which obviously doesn"t seem appropriate.
problem: avoid performing two interface verifications when external callers call two systems at the same time. How should the architecture be designed?