This is the Fremont district with a sign that says welcome to the center of the universe, It also … [+]
Data is distributed. By its very 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 wider orchestrated data entity woven between different aspects of an IT stack is fundamental to how modern technology works.
Cloud computing clouds are also distributed. A proportion of our cloud estate will always be located within the four walls of a business within its own datacenter and this will typically start with instances of private cloud. Clearly there are now major connection pipes to the public cloud and the technology layer that cloud services provider hyperscalers offer. Plus of course there’s the in-between world of hybrid cloud as well. But significantly for this discussion, there is also a burgeoning portion of cloud and data residing on the computing edge i.e. that space we defined as the Internet of Things (IoT), outside of traditional datacenters where it is sometimes only occaisonally connected to traditional cloud services.
This diversified variegation of information, compute, storage and analytics allows us to now suggest that the distributed cloud is cementing itself as a normality. So, how do we explain this technology?
What Is Distributed Cloud?
We might 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 one single part of the hyperscaler’s datacenter – and some of those elements of cloud may be further distributed when they are delivered in hybrid cloud format via an organization’s own on-premises datacenter.
In fact, we call this type of cloud service not necessarily distributed, primarily because it shares the same centalized infrastructure layer.
A distributed cloud on the other hand is more markedly characterized by the fact that the cloud control plane (the centrally managed cloud orchestration interface) is separated from the data storage layer itself. As explained here previously, a cloud computing control plane makes decisions about distributing, provisioning, scheduling, maintaining and operating cloud workloads. When this cloud management function is separated from cloud data storage, that storage is able to reside in multiple locations… and that’s good news if we want to create more flexible interconnected applications. It’s also good news if we move computing resources around faster and it provides a boost in terms of its ability to facilitate resilience 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 datacenters (including air-gapped environments) where even the control plane and operations can be distributed.
Designed to meet the specific needs of customers around digital sovereignty, latency or regulatory requirements and with AI and data-intensive workloads in mind, Google Cloud says that GDC enables organizations to take a pivotal step towards accelerating AI adoption into the business operations layer. Through the use of GDC, organizations can meet local requirements for cloud environments, accelerate AI adoption and foster an open and value-driven approach.
Functions Of Distributed Cloud
Google Cloud positions its distributed cloud technology as a means of creating smarter networks. With GDC, customers will get an environment to run sensitive network data and AI workloads that some country regulators may require to be local or on-premises. From capacity planning to root cause analysis, there is automation of reporting, classification and analysis here.
Companies like Orange (a GDC customer) will be able to run generative AI models on-premises in an environment that is integrated into similar Vertex AI services on Google Cloud. According to the Google Cloud team, Orange’s operations and customer service teams benefit from those AI models by getting the answers they need faster, while customers will experience a faster time-to-resolution and improved quality of service.
Beyond the contact center, Orange has used AI and Google Cloud technology to deliver personalized recommendations for relevant phones, plans and services – all functions that the company says are aligned to improving customer lifetime value. GDC also allows generative AI-based speech recognition to occur in each Orange country, bringing these AI technologies even to countries without a Google Cloud region.
What Distributed Does Dextrously
As Enrico Signoretti, VP of product and partnerships at Cubbit writes on TechRadar, “One of the distributed cloud’s paramount benefits is the unprecedented degree of control it offers. Indeed, the distributed model eradicates the common issue of vendor lock-in, while also allowing organizations to precisely dictate the geographical perimeter where their data resides. This could mean having parts of your data securely stored in France, Italy, Germany, or literally any place you want, offering unprecedented levels of redundancy while complying with data localization requirements. Beyond data sovereignty, distributed cloud storage facilitates comprehensive independence over all facets of data management, ensuring that organizations can comply with evolving regional, European and global regulations without relinquishing control to third-party providers and hyperscalers.”
While this might sound like a European issue (given the example countries noted above) and with legal requirements differing on balance more that any state-by-state federal ruling across the USA, the action to separate and define in this way still has complete relevance for the North American market, especially where firms have an international presence. Regardless, data governance and digital sovereignty will remain increasingly pressing issues in all countries.
“Google Distributed Cloud is delivering the capability customers need to run AI anywhere, keeping their data local and addressing latency, reliability, regulatory, or sovereignty needs,” said Sachin Gupta, vice president and general manager of the infrastructure and solutions group at Google Cloud. “From helping Orange manage data and AI needs across 26 countries to simplifying connected operations across tens of thousands of locations for McDonald’s and providing fully air-gapped cloud solutions to CSIT in Singapore, GDC extends Google’s leading AI, data and security capabilities to the customers and partners’ datacenters and edge.”
As cloud computing continues to evolve, it is inevitably finessing services that align to more specific use cases, it is morphing to shape its services to suit the scale of particular needs and functions more closely… and it is extending itself to apply more precision-engineered management controls to every particular customer requirement. All of which generally explains what is happening here with distributed cloud services.
Cloud and data are distributed, let’s share it out.