Google Cloud is built on world-class technical infrastructure that supports services that are loved and relied on by billions of people across the globe: Google Search, YouTube, Gmail, Google Maps and more. A core tenet at Google Cloud is to leverage Google’s experience building and operating highly available and highly reliable planetary-scale compute, storage and networking systems and data centers. 

Google takes a workload-optimized approach to building its infrastructure, employing a combination of dedicated hardware and software components to meet its workloads’ ever-growing demands. Underpinning this infrastructure is Titanium, a system of purpose-built, custom silicon and multiple tiers of scale-out offloads that together power improvements in the performance, reliability, and security of our customers’ workloads (for example, 25% faster block storage IOPS/instance compared to the other two leading hyperscalers). Unveiled today at Google Cloud Next, you’ll find Titanium technology in many of Google Cloud’s recent infrastructure offerings.

10x demands of tomorrow 

Meeting the growing performance, reliability, and security demands of both legacy and emerging workloads is a constant challenge for cloud infrastructure providers. And now, these demands are multiplying with the heightened adoption of generative AI across almost every industry. Meanwhile, the benefits of Moore’s law have been declining in recent years. We can’t rely on silicon advances alone to meet tomorrow’s needs.

As just one example, this chart shows the exponential computing demands of large language models.

It was clear to us a long time ago that we needed to rethink our infrastructure designs to meet these demands. This is why, for several years, we’ve adopted workload-optimization and intentional design as central principles for our infrastructure platform. We engineer golden paths from silicon to the customer workload, using a combination of purpose-built infrastructure, prescriptive architectures, and an open ecosystem to deliver workload-optimized infrastructure. 

Offloads play a pivotal role

Central to this strategy are offload technologies. Traditionally, the CPU wears many hats: It runs the hypervisor, the virtualization stack to enable your workloads, and manages storage and networking I/O; it’s responsible for security isolation for virtual interfaces and physical hardware, etc. In this model, customer workloads running on the CPU contend for resources with these platform tasks.

Offloads on dedicated hardware perform behind-the-scenes security, networking, and storage functions that were previously performed by the host CPU, allowing the CPU to focus on maximizing performance for customer workloads.

A recent example of an on-host offload or accelerator is the Infrastructure Processing Unit (IPU), a system-on-chip that we co-designed with Intel to enable better security isolation and performance on our 3rd gen compute instances. The IPU enables:

• Predictable and efficient compute
• Programmable packet processing for low latency, 200 Gbps networking with 3x the packets per second compared to our previous-generation compute instances
• In-transit encryption with the PSP protocol

Another important example of Google’s on-host hardware is Titan, a secure, low-power microcontroller that helps ensure that every machine in Google Cloud boots from a trusted state.

But we did not stop there. To meet tomorrow’s demands, we knew we needed to go past the performance that could be achieved using the host’s dedicated offload hardware.

A tiered system of offloads

A key component of Titanium is its modern offload architecture, which combines capabilities whose scale and performance are well-established within Google, as well as new capabilities tailored for cloud use cases. 

Just as modern workloads scale out horizontally in the cloud, with Titanium, we’ve extended the architecture to augment on-host offloads with an additional tier of scale-out offloads that run outside the host. This system of offloads is deployed fleet-wide and dynamically adjusts to changing workload needs to continually deliver the best performance.

Example 1: Block storage
Titanium scale-out offload enables Hyperdisk block storage to deliver stellar I/O performance. Hyperdisk’s Titanium offload on the host IPU works in tandem with the Titanium scale-out offload tier that distributes I/O across Google’s massive cluster-level filesystem, Colossus.

With traditional offload architectures, higher block storage IOPS requires purchasing larger compute instances. For example, you may need to deploy a data-intensive workload on a compute instance with many more vCPUs than the workload needs just to get sufficient storage performance. This tight coupling results in wasted resources and higher costs for customers. Further, even with large instances, storage performance in the cloud may be inadequate relative to what customers are used to with on-prem storage systems.     

With our new block storage, Hyperdisk powered by Titanium, we have decoupled compute-instance size from storage performance. Hyperdisk uses a tier of offloads in our cloud fabric to offload storage I/O from the customer hosts to achieve higher storage performance even with a general-purpose VM.   

In fact, today we are announcing that Titanium-powered C3 VMs with Hyperdisk Extreme now support 500K IOPS per compute instance in preview to meet the needs of the most demanding workloads. This is 25% faster IOPS/instance compared to the other two leading hyperscalers, courtesy of the Titanium system. 

Example 2: Network routing
Virtual network routing is another example of using a second tier of scale-out offloads (“hoverboards”). With Titanium, Google’s Andromeda virtual networking stack on the IPU offload device sends all packets for which it does not have a route to Hoverboard gateways, which have forwarding information for all virtual networks. Hoverboards are standalone software switches that act as default routers for some flows.

Unlike the traditional gateway model, the control plane dynamically detects flows that exceed a specified usage threshold and programs them to be direct host-to-host flows, bypassing the hoverboards allowing hoverboards to focus on the long tail of less frequent flows. Typically, only a small subset of possible VM pairs in a network communicate with one another, so the VMs only have to store and process a small fraction of the usual network configuration on an individual VM host, improving per-server memory utilization and control-plane CPU scalability.

Titanium already powers your workloads

The Titanium journey began years ago with the component technologies described above. Many of our products already benefit from this architecture, and the newest elements of this architecture are now available with our 3rd gen Compute Engine instances such as C3 and the new Hyperdisk block storage. 

Going forward, look for the Titanium architecture to underpin all future generations of our infrastructure offerings, in the process enabling new classes of infrastructure capabilities that move well beyond the confines of a single server.

In the fragmented world of U.S. healthcare, patients often have to wait in line or on hold, navigate multiple patient portals, and fill out numerous request forms—all in pursuit of their own medical history. Healthcare technology startup PicnicHealth is on a mission to put control back with the patient, where it belongs.

PicnicHealth’s growth, from closing successful venture rounds to winning machine learning (ML) competitions, speaks to not only improvements and opportunities in healthcare, but also how startups are leveraging Google Workspace and Google Cloud services to accelerate momentum.

The company does the heavy lifting of collecting records and leverages human-in-the-loop ML to transcribe and validate them with an abstraction team of medical professionals. The records are then structured into a complete medical history that patients can access and share with providers to get better care.

But PicnicHealth helps to improve patient health on more than one front. It allows patients to contribute their data to de-identified medical research, building high-quality, anonymized datasets that researchers and life sciences companies can use to better understand disease progression and treatment in the real world.

Google Workspace has been part of PicnicHealth from day one, helping the founders collaborate and shape the company’s vision using cloud-synced documents and spreadsheets to collaborate and model predictions. “I’ve had a Gmail account since 2006, and in college 100% of people worked out of Google Docs. Workspace continues to be the best choice for online collaboration, and that’s why it’s still the default standard for startups,” said Troy Astorino, CoFounder & CTO of PicnicHealth. “When we started PicnicHealth, my co-founder Noga Leviner was in San Francisco and I was in Southern California, and of course we used Workspace.”

Today, Google Workspace continues to play a central role in the company’s collaboration. “We create design documents in Google Docs, primarily for engineering and product changes, and get really healthy, vibrant discussions through comments,” noted Astorino. “This practice has grown beyond engineering and is used for everything from how the company operates to communication norms.”

Google Workspace offers everything the team needs to collaborate, no matter where employees are. PicnicHealth’s team has spread from San Francisco to being distributed across the country and around the world. Instant collaboration is crucial.

“We work in a complex domain where people need a lot of information to make good decisions,” said Astorino. “Google Workspace allows us to operate in a mode of default transparency, where people can easily get the information they need even if it wasn’t intentionally or directly shared with them. Whether it’s working in Docs or scheduling in Calendar, we can operate much more effectively than we could otherwise.”

By any measure, PicnicHealth’s trajectory is one of record success. The startup is a 2014 alumni of Y Combinator, a program that helped launch household names like Airbnb, DoorDash, and Dropbox. Three years later, the team went on to win the $1 million grand prize at Google’s Machine Learning Competition. And the momentum has continued— PicnicHealth has recently announced a $60 million Series C round, bringing the amount raised to date to over $100 million. With the Series C, PicnicHealth is investing in expanding its reach to more patients across over 30 diseases.

As a healthcare startup, PicnicHealth faced a very particular set of challenges, especially when working with and accessing data. Data fragmentation and interoperability are only some of the challenges of realizing the value of big data in the cloud. The healthcare industry is notoriously difficult to navigate due to sensitive data protection laws and regulations like the Health Insurance Portability and Accountability Act (HIPAA).

PicnicHealth started in the cloud on Amazon Web Services (AWS). However, after migrating over to Kubernetes and facing an expanding list of requirements for HIPAA compliance, the company started to explore alternatives.

“We needed to be HIPAA compliant, which was going to be painful on AWS, and we wanted to get away from managing and operating our own Kubernetes clusters,” recalled Astorino. “We had heard good things about GKE (Google Kubernetes Engine). And particularly valuable for us, — many technical requirements you need for HIPAA compliance are configured by default on Google Cloud.”

PicnicHealth would have had to implement a lot of changes and get specialized instance types to get their existing configuration to work. So, they began experimenting with Google Cloud and discovered a much smoother experience.

“It was a lot easier to manage in terms of product setup and developer experience,” said Astorino. “There is a sane product hierarchy of resources you can access and use through Google Cloud and the relationships between them, from coordinated IAM (identity and access management) to using Google Groups for granting permissions. Overall, it’s cleaner.”

Astorino added that the move has also opened the doors to taking advantage of other services in the Google Cloud ecosystem like Cloud SQL, BigQuery, and Cloud Composer. PicnicHealth also uses Security Command Center because it easily integrates with everything but also helps meet various compliance frameworks’ requirements, providing visibility, near-real-time asset discovery, and security information and event management.

But most importantly, the integrated ecosystem has simplified the work needed for PicnicHealth to create a secure environment for employees to use when working with sensitive medical records while still providing all the tools they need. For example, abstractors not only use Google Workspace but also have Chromebooks because they are easy to manage and secure.

Altogether, Google Cloud helps form a technology stack that has enabled the startup to build a massive labeled dataset containing over 100 million labeled medical data concepts. In turn, it accelerates PicnicHealth’s ability to generate highly-performant AI models and feed other ML pipelines, which has been vital for processing and reviewing data at scale.

To learn more about how Google Workspace and Google Cloud help startups like PicnicHealth accelerate their journey, visit our startups solutions pages for Google Workspace and Google Cloud.