Home > Net > Ask a question of architecture, single sign-on (SSO subsystem related user information
Ask a question of architecture, single sign-on (SSO subsystem related user information
Time:09-30
consult everybody a great god, and I'm SSO system saved user information, the subsystem business only preserved the user ID, batch relate to user information?
Such as subsystem. A, I have A order list, as follows:
when managers to the background check order, how to put the user information (such as name, age) and order information together?
My idea is to write an interface SSO system, used to batch to obtain user information, as long as the incoming to obtain a user ID set,
For example, 8111, 17, 10... This way, when the system load A first page 10 data, then the background request interface is the 10
ID, and then get to user details after returned to the subsystem. A, in A system integrating returned to the front page,
In so doing, the feeling is very LOW, performance is not good, please everyone a great god what is the solution?
Or subsystems in a structure, not only save user ID? Please comment, thank you very much
CodePudding user response:
Can two database user table synchronization
CodePudding user response:
Create a Token mechanism, single point correction
CodePudding user response:
Across the library? If the MSSQL, try to add links to database server
CodePudding user response:
When 1. Users of the system in accordance with the "consistent hash" hash points inventory 2. Set up a system of intermediate processing flow, such as the spark of batch, make some statements need redundant indexes, to spark group of the department, in batches to hash cluster management database consistency check
Ps: the consistency of the hash, you baidu, I simply say that an institution assumes that the hash is 2001, for example, database records this is 2001 institutions of information, you can say you have a server tube 1 w institution, then obviously the first server is over 2001, you go to the first server queries, this process is based on the consistency of the hash in the cluster is automatic, don't need you tube,
About offline flow Stream batch you can baidu, I won't explain
CodePudding user response:
Actually work reports and polymerization are under offline "secretly", this is a few years ago the fire of a "Hadoop core (distribution + step by step, secretly underneath the N machine data, step by step, to aggregate, words is not a.net garbage, but because of China's net, the preacher like dealing with nouns, don't like serious to do the practical work, they think this kind of work in furtive polymerization under can't reflect the feeling of the tall, too low)