According to a report by the World Economic Forum, digital transformation is a 100 trillion USD opportunity in GDP growth. This opportunity isn’t just in the standard technological industries, either—it also affects everything from healthcare, retail, and media to travel, transportation, and logistics.
Naturally, with such an immense opportunity comes challenges. In addition to the obvious technical execution and commercial viability, any digital transformation must also solve for latency, risk, complexity, capacity, and sustainability.
So, how do we design for the future today? This article explores what being data- and AI-ready really means.
The Data Economy
Before we can fully understand how infrastructure needs to evolve, we first need to understand why it needs to evolve. That “why” comes down to one word: data.
Beginning in the 1980s, when all aspects of society faced increased digitalization, our economy shifted from being physical-powered to digital-powered. As this digitalization continued and reached new heights over the next few decades, we’ve now arrived at the next phase of our economy’s transformation: being data-powered. The rise of high-density data created from new applications, services, and technologies is generating not only immediate economic growth, but also explosive growth in data, which creates a continuous data creation lifestyle filled with new interactions and shifting data flows. Our new data economy stems from this abundance of perpetually compounding data and all the fast-growing technologies (like mobile networks, IoT, AI, and ML) that both support and produce it.
As a result of this shift, the physical locations at the centers of economic trade are now also centers of enterprise data generation, management, and exchange. The challenge these facilities face is that the magnitude of data growth from AI and related technologies is forcing rapid—and continual—changes to where and how data is created, processed, stored, and exchanged. In other words, our current systems aren’t AI-ready. For example, while many businesses have already undertaken cloud migration for their data systems, new points of data generation (like the rapid growth of enterprise AI adoption) along with security and compliance requirements have intensified data creation and processing at the edge. However, as data creation and utilization continues to evolve, data attracts more and more applications and services, which becomes a force multiplier that creates Data Gravity (a data gravitational pull that attracts applications and services and, subsequently, more data creation)—something current IT architectures don’t take into account. Being AI-ready means being data-ready, and our current systems need to catch up.
A New Architecture
Rapid data creation and escalating needs for storage and processing capacity put tremendous strain on legacy servers and applications. As data volumes experience rapid growth, it’s no longer feasible to support existing application flows in a performant manner. The cost of transport and increasing latency associated with larger data and AI-intensive flows and centralized data storage make legacy architectures unsustainable.
In order to create digital infrastructure that can support Data Gravity and thus be both data- and AI-ready, we need a new architecture that can overcome the challenges associated with data exchange and platform growth. Enterprises also need to be able to serve customers, partners, and employees across all channels, business functions, and Points of Presence (POPs). That means building a system that is distributed, data-centric, and hybrid to invert traffic flow, leverage interconnection, and bring clouds and users to the data, to integrate private and public data sources. Deploying centers of data exchange at local POPs where enterprises do global business could remove Data Gravity barriers to accommodate data and AI-intensive workflows that vary by participant, application, information, and location-specific needs.
But what does that look like in practice?
AI-Ready Digital Infrastructure Foundations
Today’s business and technology leaders require a business platform that operates ubiquitously and on-demand, augmented by real-time intelligence. To solve digital transformation challenges in our burgeoning data economy requires digital infrastructure foundations that support four strategic IT priorities: network, security, compute, and data.
Network
Network traffic is shifting, and there’s increasing pressure for enterprises to optimize and manage network infrastructure accordingly. According to Cisco’s AI Readiness Index, 79 percent of companies have challenges with network latency in supporting AI workloads. Digital Realty’s recent Global Data Insights survey came to a similar conclusion: 77 percent of IT leaders identified latency-specific performance requirements.
Legacy networking for critical infrastructure is costly and inflexible. Enterprises require a new network infrastructure architecture that reduces latency, localizes access to data, standardizes network deployments, and provides any-to-any interconnection. To enable data- and AI-ready network architecture, enterprises need to create local ingress and egress points in colocation to:
- Consolidate and localize traffic
- Segment and tier traffic
- Interconnect network, cloud, and service providers
- Deploy, interconnect, and host SDN edge
Security
Securing, controlling, and protecting infrastructure in today’s data-powered economy is not only difficult to manage, but also increasingly imperative: According to Cybersecurity Ventures’ cybercrime report, the global cost of cybercrime will reach 10.5 trillion USD annually by 2025.
Enterprises require a new security infrastructure architecture that allows them to secure controls and infrastructure in a distributed manner and unlock seamless global security. To enable data- and AI-ready security architecture, enterprises need to operationalize distributed security access points in colocation by:
- Implementing ingress / egress control points
- Hosting IT and security stacks at these points
- Connecting directly to Software-as-a-Service security and data services
- Enhancing security posture while reducing vulnerability points
Compute
Legacy IT inhibits modern high-performance computing and is capital-intensive and hardware-oriented. Based on the data from Digital Realty’s Data Gravity IndexTM 2.0, the majority of enterprise data is created and utilized outside the public cloud. Hybrid and multi-cloud deployments decrease infrastructure costs and IT labor, but they also can introduce workload performance and scalability issues.
Enterprises require a new secure Hybrid IT architecture with compliant data governance practices that enable them to transact, process, and action insights seamlessly. To enable data- and AI-ready compute architecture, enterprises need to host applications and optimize for workloads and scale through a distributed architecture designed to:
- Host analytics adjacent to network and data aggregation points
- Accelerate compute-intensive workloads with GPU / DPU
- Enable microservices with virtualization and containers
- Scale processing with clustering and high-performance interconnects
Data
With the explosive cost associated with data transit and reduced performance due to increased traffic, latency, and backhaul, effective data integration and management is more important for enterprise success than ever. According to Cisco’s AI Readiness Index, 81 percent of respondents admitted that their data still exists in silos across their organizations.
Enterprises require a new data infrastructure architecture that localizes data aggregation, staging, analytics, streaming, and management at global PoPs. To enable data- and AI-ready data management, enterprises need to localize data sets for secure data exchange through a distributed architecture designed to:
- Host data adjacent to network and analytics aggregation points
- Optimize data exchange between users, things, networks, and clouds
- Enable real-time intelligence across distributed workflows locally and globally
- Create secure B2B data exchange to unlock new business opportunities
A Proven Approach to a Data- and AI-Ready Architecture
While these architectures won’t be implemented overnight, they are already being adopted successfully around the globe. For instance, at Digital Realty, we built our patented PDx® methodology around these principles. Our PlatformDIGITAL® solutions create centers of data exchange that, when deployed on a single global data center platform, enable enterprises to colocate and interconnect digital infrastructure foundations in network, security, compute, and data. We also recently released our Asia-based Apollo AI platform, which allows us to enhance energy efficiency across our global portfolio. Data- and AI-ready architectures are here, and we need them yesterday in order to keep up with the data economy. The only question is: Are you ready?
