#virtualmachines

A hypervisor is a key virtualization technology that allows multiple virtual machines (VMs) to run on a single physical server by managing and allocating hardware resources efficiently.

In this guide, you’ll learn what a hypervisor is, how it works, and the difference between Type 1 and Type 2 hypervisors used in cloud computing, data centers, and server environments.

Containers and Virtual Machines (VMs) are powerful technologies used to create isolated computing environments by packaging applications along with their required libraries and system tools. Containers focus on speed and efficiency, enabling faster deployment and better portability. In contrast, Virtual Machines provide stronger isolation by running a complete operating system for each instance, making them ideal for supporting diverse workloads.

In a traditional setup, you’d send requests, wait for approvals, buy hardware, set it up and by the time it’s ready, the project’s already halfway through!

But in the cloud world, that same setup happens in minutes thanks to virtual machine provisioning. With just a few clicks, organizations can create, configure, and deploy servers instantly. It’s like moving from building a car piece by piece to summoning one ready to drive.

Confidentiality VMs running AMD SEV on C4D machines preview

Google Titanium hardware gives the C4D machine series excellent, reliable, and consistent performance with 5th-generation AMD EPYC (Turin) CPUs.

Google Cloud offers Confidential Compute on AMD N2D, C2D, and C3D machines worldwide. The general-purpose C4D machine series' confidential virtual machines (VMs) with AMD Secure Encrypted Virtualisation (AMD SEV) technology are in preview today and will soon be released.

C4D virtual machines use Titanium and fifth-generation AMD EPYC Turin processors. C4D outperforms C3D by 30% on the anticipated SPECrate2017_int_base benchmark, growing performance with fewer resources and maximising expenditures.

C4D supports web, app, and game servers, AI inference, web serving, video streaming, analytics, and relational and in-memory databases.

C4D can execute 38% more Memorystore for Redis operations and 56% more Cloud SQL for MySQL queries than C3D due to its higher core frequency (up to 4.1 GHz) and improved IPC.

With C4D, AMD EPYC Turin may boost web-serving throughput per vCPU by 80% and enhance branch prediction and L3-cache efficiency.

Features of C4D machines

The C4D machine series has these traits:

Titanium and AMD EPYC Turin power it.

Supports 384 virtual CPUs and 3,024 GB DDR5.

Local Titanium SSDs up to 12,000 GiB are supported.

Preconfigured machines with 2–384 virtual central chips are available.

Future bookings, Spot VMs, and on-demand consumption are supported.

Allows conventional network setup with 100 Gbps bandwidth.

Supports Tier 1 VM networking at 200 Gbps.

For HyperDisk volumes only.

Confidential Virtual Machine AMD SEV support Flexible, resource-based committed use discounts

Supports compact and distributed placement policies.

C4D machine series types

Regular, high-cpu, and high-mem C4D virtual machines have predefined configurations from 2 to 384 vCPUs and up to 3,024 GB of memory.

Build your instance with the C4D machine type -lssd option to use Titanium SSD. Selecting this machine type creates a Titanium SSD-partitioned instance of the required size. Different Titanium SSD volumes cannot be joined.

Custom machines are incompatible with C4D.

Supported C4D VM disc types

It supports only NVMe disc interface and Hyperdisk block storage:

Hyperdisk-steady

Hyperdisk Extreme

Certain system types automatically receive local titanium SSD with the -lssd option.

Persistent Disc is incompatible with C4D.

Disc and capacity limits

A virtual machine (VM) can use a variety of hyperdisks, but their total disc capacity (in TiB) cannot exceed:

For systems under 32 vCPUs, all hyperdisks have 257 TiB.

Different computers with 32 vCPUs or 512 TiB for the hyperdisk

Networking C4D machines

Network interfaces for C4D virtual machines must be gVNIC. C4D can handle 100 Gbps for regular networking and 200 Gbps for per-VM Tier_1 networking.

Make sure your operating system image supports the gVNIC driver before switching to C4D or starting instances. Choose an OS image with “Tier_1 Networking” and “200 Gbps network bandwidth” on the Networking features tab of the OS information table for optimum C4D performance. Although the guest OS lists gve driver version as 1.0.0, these images upgrade the gVNIC driver. If your C4D instance runs an operating system with an outdated gVNIC driver, it may have greater latency or less network bandwidth than it should.