Q13Cloud Computing

Virtualization is the absolute foundation of cloud architecture, allowing physical hardware to be mathematically sliced into multiple isolated environments. However, the exact software layer where this slicing occurs defines whether you are running a massive Virtual Machine or a microscopic Container.

1. Hardware-Level Virtualization (Hypervisors)

This is the classic IaaS model (VMware, Xen, KVM).

• Architecture: A heavy Virtual Machine Monitor (Hypervisor) sits directly on the bare-metal hardware. It mathematically traps and emulates all physical CPU instructions and interrupts.
• Execution: Every single Guest VM must boot its own massive, entirely separate Operating System Kernel (e.g., a Windows VM and a Linux VM running on the same physical server).
• Merit: Absolute, mathematically guaranteed isolation and security. A kernel panic in VM1 cannot physically crash VM2.
• Demerit: Massive architectural bloat. Booting takes minutes, and each OS consumes gigabytes of idle RAM just to keep its kernel alive.

2. OS-Level Virtualization (Containers)

This is the modern PaaS/Microservices model (Docker, LXC).

• Architecture: There is absolutely no Hypervisor. There is only ONE single Host Operating System Kernel running on the bare metal.
• Execution: The OS Kernel mathematically isolates processes using cgroups and namespaces. The "Containers" physically share the exact same underlying Linux kernel, but mathematically believe they are isolated systems.
• Merit: Blisteringly fast. Containers boot in milliseconds because there is no kernel to boot. They consume almost zero overhead, allowing a single server to run 1,000 containers instead of 10 VMs.
• Demerit: Weaker security boundary. A catastrophic exploit in the shared Host Kernel will violently compromise every single container.