All geek questions in one place

Future client-server code verification?

We have a web client server product. It is expected that the client will be used by more than 1M users (a well-known company is going to use it).

Our server is configured in the cloud. One of the main design issues is how to make the entire program a proof in the future. Say:

  • The cloud provider drops, then automatically moves to backup in another cloud.
  • Move to another server, etc.

The options we still thought were:

  • DNS: Launches the DNS name server in the cloud.
  • Directory Server The directory server also lives in the cloud.
  • Let our server return future movements and future URLs, etc. to the client - in which the client is specifically designed to handle these scenarios.

Since this should be a common problem, is this the best solution for the same? Since our company is very small, we consider at least a technically and financially expensive solution (for example, option 3, etc.)?

Can someone point some pointers to the same thing?

TO

+3
source share
3 answers

I would choose the directory server option. He is the most flexible and gives you maximum control over what is happening in this situtaion.

To avoid the directory itself becoming the only point of failure, I would have three or four of them working in different places with different providers. Ask the client application to randomly select one of the directoy URLs at startup and work its way up until it finds one that works.

, , , - , , . p >

0

Re. DNS: , , ( ).

IP-, . IP , , .

0

, 100% , , : , ?

, DNS .

, , : , , .. , DNS .

, DNS- , . ( TTL), ( , , - ).

0

Source: https://habr.com/ru/post/1742735/

More articles:


All Articles