This is the Fremont district with a sign that says welcome to the center of the universe, too … [+]
Data is distributed. By nature, we work with enterprise-level data across different documents, applications, databases, and deeper systems. The fact that we can distribute any single piece of information into a larger orchestrated entity of data that interweaves between different aspects of an IT stack is fundamental to how modern technology works.
Cloud computing clouds are also distributed. A percentage of our cloud assets will always reside within the four walls of a business within its own data center, and that usually starts with private cloud instances. It is clear that there are now significant pipelines to connect to the public cloud and the level of technology offered by hyperscale cloud service providers. Plus, of course, there’s the in-between world of hybrid cloud. But importantly for this discussion, there is also a growing segment of cloud and data that resides at the computing edge, that is, that space we have defined as the Internet of Things (IoT), outside of traditional data centers where it is sometimes only occasionally connected to traditional cloud services.
This diverse variety of information, compute, storage and analytics now allows us to suggest that the distributed cloud is becoming established as the norm. So how do we explain this technology?
What is Distributed Cloud?
We could be forgiven for thinking that all cloud is distributed. After all, cloud services are generally created by drawing power from more than one single blade server in a single part of the hyperscaler’s data center – and some of these cloud components may be further distributed when delivered in hybrid cloud form through an organization’s own data center indoor data.
In fact, we call this kind of cloud service not necessarily distributed, mainly because it shares the same central infrastructure layer.
A distributed cloud on the other hand is more strongly characterized by the fact that the cloud control layer (the central managed cloud orchestration interface) is separated from the data storage layer itself. As explained here previously, a cloud computing control layer makes decisions about the distribution, provisioning, scheduling, maintenance, and operation of cloud workloads. When this cloud management function is separated from cloud data storage, this storage can be in multiple locations… and that’s good news if we want to build more flexible interconnected applications. It’s also good news if we’re moving computing resources faster, and it provides a boost in its ability to facilitate resiliency and redundancy.
Google Cloud recognizes the benefits of distributed cloud. The company’s Google Distributed Cloud (GDC) is a fully managed hardware and software service that brings the power of Google’s AI services to the computing edge and data centers (including air-gapped environments) where even the control layer can be distributed and functions.
Designed to meet specific customer needs around digital dominance, latency or regulatory requirements, and with artificial intelligence and data-intensive workloads in mind, Google Cloud says GDC enables organizations to make a critical step towards accelerating AI adoption at the business operations level. Through the use of GDC, organizations can meet local requirements for cloud environments, accelerate the adoption of artificial intelligence, and promote an open and value-based approach.
Distributed Cloud Operations
Google Cloud positions distributed cloud technology as a means of building smarter networks. With GDC, customers will gain an environment to run sensitive network data and AI workloads that may be required by some country regulators to be local or on-premises. From capacity planning to root cause analysis, reporting, classification and analysis automation is here.
Companies like Orange (a GDC customer) will be able to run on-premise AI production models in an environment integrated with similar Vertex AI services on Google Cloud. According to the Google Cloud team, Orange’s operations and customer service teams benefit from these AI models by getting the answers they need faster, while customers will experience faster time to resolution and improved service quality.
Beyond the contact center, Orange has used AI and Google Cloud technology to provide personalized recommendations for relevant phones, plans and services – all features the company says are aligned with improving customer lifetime value. GDC also enables the creation of AI-based speech recognition in each Orange country, bringing these AI technologies even to countries without a Google Cloud region.
Everything distributed does skillfully
As Enrico SignorettiVP of Product and Partnerships at Cubbit writes on TechRadar, “One of the key benefits of the distributed cloud is the unprecedented degree of control it offers. Indeed, the distributed model eliminates the common issue of vendor lock-in, while also allowing organizations to precisely dictate the geographic perimeter where their data resides. This could mean having parts of your data safely stored in France, Italy, Germany or literally anywhere you want, offering unprecedented levels of redundancy while complying with data localization requirements. Beyond data sovereignty, distributed cloud storage facilitates complete independence in all aspects of data management, ensuring organizations can comply with evolving regional, European and global regulations without ceding control to third-party providers and over-scaling.
While this may sound like a European issue (given the example of the countries noted above) and with legal requirements that differ more than any federal state in the US, the act of separating and defining in this way is still entirely relevant to North American market, especially where companies have an international presence. Regardless, data governance and digital sovereignty will remain increasingly pressing issues in all countries.
“Google Distributed Cloud provides the ability customers need to run AI anywhere, keeping their data local and addressing latency, reliability, regulatory or sovereignty needs,” he said. Sahin Gupta, vice president and general manager of the infrastructure and solutions group at Google Cloud. “From helping Orange manage its data and AI needs in 26 countries to streamlining connected operations across tens of thousands of locations for McDonald’s and providing full-gap cloud solutions to CSIT in Singapore, GDC is expanding its leading AI capabilities , data and security of Google to customers and partners’ data centers and edge.”
As cloud computing continues to evolve, it inevitably improves services that align with more specific use cases, morphs to shape its services to better fit the scale of specific needs and operations… and expands to implement more precise engineering management controls for each specific customer requirement. All of this generally explains what’s going on here with distributed cloud services.
Cloud and data are distributed, let’s share it.
