In this article the author shares practical experience in designing and implementing protection for a private cloud hosted across two geographically separated sites of a data center (ЦОД).
The material does not contain deep technical details but instead describes the path from an assumed ideal solution to real-world implementation through research, errors, technical constraints, and necessary compromises.
Background and Objectives
The task appeared fairly standard: ensure service resilience, achieve compliance with information security requirements, and preserve normal manageability of the infrastructure.
On paper such projects usually look like combinations of best practices: high availability, segmentation, NGFW, some dynamic routing, and it seems that this should be sufficient.
In practice everything proved more complicated when each element of the architecture began to conflict with the others, especially when several contractor teams participated and each team held its own vision of how the network and security architecture should look.
At one point the discussions started to resemble an engineering construction set: every party offered its own solution, and the task was to assemble from these proposals a working and scalable system.
