GCP – Now GA: Compute Engine C3 bare-metal and X4 instances

Enterprise workloads present unique challenges when migrating to a public cloud. Whether it’s a 32TB in-memory database, a private cloud built on a commercial hypervisor, or an application that can’t run in a virtual machine, these business-critical workloads are a high-priority for technology leaders to migrate to the cloud. 

To help ease and accelerate migrating these workloads, we’re excited to announce that Compute Engine’s X4 and C3 bare-metal machine types are now generally available. These machine types address unique compute needs within the general-purpose and memory-optimized families:

The new X4 instance series includes three new instance types to address extra-large in-memory databases such as SAP HANA.

Three new C3 bare-metal shapes cater to a variety of applications, such as commercial and custom hypervisors, that require direct access to CPU and memory resources. 

Underpinning X4 and C3 is Titanium, Google Cloud’s system of purpose-built custom silicon and multiple tiers of scale-out offloads. By freeing up the CPU, Titanium provides performance, reliability, and security improvements for a variety of workloads. In fact, Titanium is the foundation to all third-generation and newer Compute Engine machine types.

“Intel and Google recognize the unique needs of enterprises migrating to the cloud. Through our collaboration on X4 and C3 Bare-metal instances powered by 4th Gen Intel Xeon processors, we’re providing businesses with unique instances and performance enhancements along with the flexibility and power they need to seamlessly transition even their most complex workloads, unlocking the full potential of the cloud.” – Suzi Jewett, General Manager – Intel Xeon Products, Intel Corporation

Now GA: X4 for extra-large memory-intensive workloads 

Powered by 4th Generation Intel Xeon Scalable processors, X4 offers industry-leading compute and block storage performance, with up to 1,920 vCPUs and 32TB of DDR5 memory. X4 instances are ideal for extra-large, memory-intensive workloads such as relational in-memory databases and graph in-memory databases, and are offered in three configurations:

Uncompromising performance and availability for all memory sizes

As an application’s in-memory database grows, database administrators and architects need to be able to scale their workloads without sacrificing performance. X4 offers a 17:1 Memory / Core ratio for all three of its instance shapes to help ensure performance scales with size In fact, X4 has achieved the highest SAPS benchmark certification results of any hyperscaler for Intel servers. This instance shape’s design is one reason why X4 is the only 32TB hyperscaler instance certified by SAP and the only hyperscaler instance certified for 24TB and 32TB OLTP and OLAP with SAP ‘Standard Sizing’ — no other cloud providers hold these certifications. 

With support for Hyperdisk Extreme, X4 supports up to 10,000 MBps of throughput and 400K IOPS. In the unfortunate event of a node failure or restart, this block storage performance enables sub-one-hour memory rehydration time, meeting recovery time objectives (RTOs) for databases as large as 32TB. In addition, X4 is backed by Compute Engine’s Memory-optimized 99.95% single-instance SLA. 

“In the past few years, our SAP HANA systems have seen significant data growth with an increasing need for higher performance. With the 24TB X4 machines and Hyperdisk storage, we have been able to raise the ceiling for our future data growth and are also looking to see improvements in our performance. Added to this, Google’s X4 machines are cloud native, giving us opportunities to automate system management and operations.” – Shawn Lund, US Chief Technology Officer, Deloitte

A maintenance experience built for business-critical workloads

When designing the X4 maintenance experience, we consulted with enterprise customers to learn what they need to run their business-critical workloads in the cloud. We came back with three key takeaways: 

Maintenance events should be rare.

Customers want control over maintenance execution.

Customers need ample upfront notice of upcoming maintenance events.

We designed the X4 instances’ maintenance experience with these takeaways as guiding principles. For one, by leveraging Titanium offloads, we achieved a drastic reduction in the number of infrastructure maintenance events that we initiate. And when a maintenance event is necessary, we give you a minimum of 60 days notice, and let you trigger that maintenance event at any time within that period. 

Together, X4’s performance and reliability characteristics have impressed some of our most demanding customers and partners.

“SAP Enterprise Cloud Services has been evaluating and exploring the integration of X4 and C3 Metal systems for RISE with SAP, and has been impressed with its reliability, performance, and ease of management.” – Lalit Patil, CTO for RISE with SAP, Enterprise Cloud Services, SAP SE

Now GA: C3 bare-metal machine types

Bare-metal machine types are ideal for customers with workloads that require direct access to the machine’s CPU and memory resources, e.g., hypervisors or container platforms; aren’t supported in virtual machines; or have special performance monitoring or licensing requirements. 

In particular, customers in financial services, as well as SaaS/PaaS security, dev/test and private cloud platform sectors, often need bare-metal machine types to meet their performance goals.

“We’re excited for the launch of Google Cloud’s new bare metal machine types. Nutanix looks forward to collaborating with Google to help organizations capitalize on the benefits of cloud while achieving a universal hybrid multicloud operating model for their legacy and AI-enabled modern apps.” – Tarkan Maner, Chief Commercial Officer, Nutanix

C3 bare-metal instances are based on the 4th Generation Intel Xeon Scalable processors — the same raw server platform that powers C3 virtual machines (VMs). On Google Cloud, these instances exclusively provide access to Intel integrated CPU features such as Intel Quick Assist (QAT), Intel Data Streaming Accelerator (DSA), and Intel In-Memory Analytics Accelerator (IAA). They offer many of the same attributes as the largest C3 VMs, for example the same Hyperdisk storage options, and the same standard and Tier 1 bandwidth. Finally, C3 bare-metal instances have the same controlled maintenance experience as the rest of the C3 machine family. These instances are offered in three configurations: 

A continued focus workload-optimized infrastructure 

X4 and C3 bare-metal instance types are the culmination of several years of investment in the Compute Engine architecture, with the Titanium offload system at its core. Unlike the similarly named Bare Metal Solution, X4 and C3 bare-metal machine types are not limited to specific workloads or use cases; they are fully integrated, native Compute Engine instances that run on Google Cloud’s server fleet in Google Cloud data centers, and can be consumed just like any other Compute Engine VM. You can expect more bare metal shapes in future machine families, as we continue to support your infrastructure modernization efforts. 

To get access to X4 instances, please contact your account team or submit an inquiry to your Google Cloud Sales Specialist. 

C3 bare-metal machine types are currently available in us-central1, us-east5, us-east1, europe-west1, europe-west4 and will be coming to additional regions with C3 VMs. Like other members of the C3 machine family, you can purchase them on-demand, with committed use discounts (CUDs) or Flex CUDs. For more, see the C3’s predefined compute resource pricing.