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.

Responding to multiple, simultaneous disruptive forces has become a daily routine for most demand planners. To effectively forecast demand, they need to be able to predict the unpredictable while accounting for diverse and sometimes competing factors, including:

• Labor and materials shortages
• Global health crises
• Shifting cross-border restrictions
• Unprecedented weather impacts
• A deepening focus on sustainability
• Rising inflation

Innovators are looking to improve demand forecast accuracy by incorporating advanced capabilities for AI and data analytics, which also speed up demand planning. According to a McKinsey survey of dozens of supply chain executives, 90% expect to overhaul planning IT within the next five years, and 80% expect to or already use AI and machine learning in planning.

Google Cloud and SAP have partnered to help customers navigate these challenges and supply chain disruptions starting with the upstream demand planning process, focusing on improving forecast accuracy and speed through integrated, engineered solutions. The partnership is enabling demand planners who use SAP IBP for Supply Chain in conjunction with Google Cloud services to access a growing repository of third-party contextual data for their forecasting, as well use an AI-driven methodology that streamlines workflows and improves forecast accuracy. Let’s take a closer look at these capabilities.

Unify data from SAP software with unique Google data signals

When it comes to demand forecasting and planning, the more high-quality and relevant contextual data you use, the better, because it helps you understand the influencing factors of your product sales to sense trends and react to disruptions or capitalize on market opportunities more timely and accurately.

The expanded Google Cloud and SAP partnership helps customers who use SAP® Integrated Business Planning for Supply Chain (SAP IBP for Supply Chain) bring public and commercial data sets that Google Cloud offers into their own instances of SAP IBP and include them in their demand planning models in SAP IBP. So, in addition to sales history, promotions, stakeholder inputs and customer data that are typically in SAP IBP, a demand planner can incorporate their advertising performance, online search, consumer trends, community health data, and many more data signals from Google Cloud when working through demand scenarios.

More data enables more robust and accurate planning, so Google continues to build an ecosystem of data providers and grow the number of available data sets on Google Cloud. Some current providers include the U.S. Census Bureau, the National Oceanic and Atmospheric Administration, and Google Earth, and partnerships are underway with Crux, Climate Engine, Craft, and Dun & Bradstreet to help companies identify and mitigate risk and build resilient supply chains.

Augmenting demand planning with additional external causal factor data is a starting point to drive more accurate forecasting. For example, knowing what regional events may be happening, or the weather patterns that may impact sales of your products, allows you to react faster to these changes by making sure adequate supply is being provided. The result is a more accurate overall plan that reduces resource waste and out-of-stock events. Planners can respond with more accurate and granular daily predictions about sales, pricing, sourcing, production, inventory, logistics, marketing, advertising, and more based on the expanded data.

Get more accurate forecasts with Google AI inside

Extending the already expansive algorithm selection available in SAP IBP, the release of version 2205 allows SAP IBP customers to access Google Cloud’s supply chain forecasting engine, which is built on Vertex AI — Google Cloud’s AI-as-a-platform offering — from within SAP IBP as part of their forecasting process.

The benefit of using an AI-driven engine for demand forecasting is that it meaningfully improves forecast accuracy. Most demand forecasting today is done through a manually set, rules-based model versus an AI-driven model that is smarter and gets better at predicting demand as it works.

Take the fastest path from data to value with streamlined workflows

Vertex AI can include relevant contextual data sets for demand planning, and the results can be shown in SAP IBP for planners to incorporate when building their workflows.

In addition to more accurate forecasts, planners can work faster and more efficiently as they build potential scenarios, meaning they can do more simulations than they do now so that a wider range of disruptions can be modeled. Customers of SAP IBP don’t have to do any of the heavy lifting. They just have to share their data from SAP IBP with Google, then access the process workflow capabilities to set up automated workflows that use the combined data. Google makes the data available so that planners can use it as they’re setting up their workflows in Vertex AI.

Users of the Google Supply Chain twin and SAP IBP can combine the rich planning data from IBP with additional SAP data and other Google data sources to provide better supply chain visibility. The Google Supply Chain twin is a real-time digital representation of your supply chain based on sales history, open customer orders, past and future promotions, pricing and competitor insights, consumer history signals, external data signals and Google data.

Leverage Google data signals with SAP IBP for more accurate forecasts

It’s not difficult to access these new capabilities, and the benefits are more accurate near-term forecasts and more return on your investments in SAP IBP and Google Cloud. If you happen to be at the Gartner Supply Chain Symposium from June 6-8th in Orlando, Florida, stop by our booth to say hello. Or, get started now

When moving to the cloud, many organizations concentrate their focus on the change in technology, and overlook an area just as complex: cultural change. 

At Google, we’ve spent years nurturing our culture and workforce to best operate in the cloud, and the Google Cloud Professional Services team leverages the lessons we’ve learned for the benefit of enterprise customers embarking on their own cloud journeys. 

While it can be tempting to believe in a universally ‘correct’ strategy for change management, there is no one-size-fits-all answer. Every organization will have its own unique considerations. But with that said, there are some core strategies we’ve found to be relevant and useful across a broad range of businesses. 

1. Define your purpose for moving to Cloud
While pockets of cloud use and experimentation can evolve independently and in parallel across an organization, it’s important to make some deliberate decisions before starting a larger migration. At this stage, we recommend having a detailed answer to two key questions to ensure a successful cloud migration:

• Where do you want to go? (Or “What’s your cloud vision?”) 
• How do you plan to get there?

Start by having a conversation with leaders and those who will be key to the journey about how far you want to push your cloud vision. This alignment ensures everyone is on the same page—and will provide greater direction, allowing more deliberate action. 

2. Find the change path which is right for you 
Whether a ‘lift and shift’ approach to the cloud is right for you, or a more transformative approach with a lot of re-architecting—the most important thing is to find the flavor of change which is appropriate to your context and level of ambition.This will both shape your key migration activities, but also the level of impact to be managed within your organization.  

There are many ways to embark on a change journey for cloud migration (which one can find in the chart below). It is important to deeply understand the needs of your business and its people and determine what strategy makes the most sense.

3. Learn from best practices
Based on the lessons we’ve learned along our own journey, and the work we’ve done with customers, there are a number of recommendations we can share that can make a cloud migration more successful. We go into these in more detail in our new whitepaper, but below you can find the ones we think are most relevant: 

• Share the vision—and measure, measure, measure. Once you’ve crystallised your cloud vision with leadership and key stakeholders, share that vision widely. Set success goals and communicate them to hold yourself accountable.  
• Be clear about the capabilities you will need in the future—and where you’ll get them. For example, if your vision is to become a cloud-first, data and AI-led organization, ensuring you have the right data science skills and machine learning capabilities in your organization to achieve that vision becomes a critical step—be they home-grown or bought-in.
• Find the right balance between capabilities that should be under central control, and capabilities that should be decentralized, or agile. For example, should machine learning be something that sits centrally, or should it be spread across your organization? For every business, the solution will be a little different, and there’s no “one true answer.” There’ll be lots of different opinions about this, so the sooner the conversation starts, the better.  
• Start thinking about the needed tech and non-tech skills now, and how you’ll fill the gaps. Building the tech skills will take time, and not everyone will feel comfortable with the future picture of collaboration, innovation, and agility. 

To help businesses navigate their own cloud journeys, Google Cloud Professional Services has released a new whitepaper that can help guide organizations. “Managing Change in the Cloud” is closely aligned with the Google Cloud Adoption Framework and is a practical guide for organizations looking to maintain momentum in their cloud adoption. You can download the whitepaper here.