All geek questions in one place

How to achieve high availability?

My boss wants to have a system that takes into account the catastrophic event of the continent. He wants to have two servers in the USA and two servers in Asia (1 entry server and 1 production server on each continent).

  • If an earthquake breaks the connection between the two continents, both must work alone. When the connection is restored, they should synchronize each other in normal mode.
  • An external cloud system is not allowed because it is not sure.
  • The system must consider scalability, which means adding new servers should be easy to configure.
  • Servers must be load balanced.
  • The connection between the servers should be very secure (encrypted and sent over SSL, although SSL will take care of encryption).
  • The system should allow one and only one user to log in with one account. (be careful with the delay between the continent and the two users sharing the account can log into both registration servers at the same time).

Please, help. I am already at the end of my mind. Thank you in advance.

+3
source share
4 answers

I believe that these requirements (if properly analyzed) are essentially incompatible, since they cannot work according to the CAP theorem.

, , . , , :

  • "", "" DC, ( ). .
  • ( , , , ).

, . , , . CAP.

, , , , , , , .

, , - . , , . - , , - . , , .

+6

, . , , , , , , . , ; , , , , . ?

.

+4

( Rackspace), .

+1

, .

- SLA, , , .

, , Google, Yahoo, Microsoft Amazon ( ), - , .

0

Source: https://habr.com/ru/post/1737279/

More articles:


All Articles