Nature proved early on that in complex systems, modular designs are the ones that survive and thrive. An important contributor to this success is the critical reliability advantage of fault tolerance, in which a modular system can shift operation from failed modules to healthy ones while repairs are made. In data centers, modular design has already taken root in new fault-tolerant architectures for servers and storage systems. As data centers continue to evolve and borrow from nature's blueprints, Network-Critical Physical Infrastructure (NCPI) must also evolve to support new strategies for survival, recovery, and growth.

Modularity is an established technique for organizing and simplifying a complex system. Nonetheless, in man-made systems on the brink of the evolutionary transition from monolithic to modular design, there can be skepticism and slow starts until modularity settles in and begins to deliver its time-tested benefits.

The mission of this paper is to illustrate, through case studies, how modularity not only delivers its more obvious and easily understood benefits but also its most subtle, least understood, and profound reliability benefit: fault tolerance.
The inherent fault tolerance of modular design provides a powerful new defense against failure, introducing into complex systems a strategy for reliability that is not only adequate, but superior.