By Kayahara Masayuki, General Manager, Service Provider Solutions Department, NEC; and Tamer Bdran, Senior Vice President, Business Unit Head Telecom Services, Digital Government & MD of UAE Office, NEC
AI has rapidly emerged as a transformative force, driving digital transformation across every industry. At the core of AI operations are its workloads, which reside in data centers and drastically alter their dynamics. Here are some key AI data center trends we are currently experiencing:
- Moving Back to Private Cloud: In recent years, the flexibility of public cloud services has made them increasingly popular. However, as AI applications demand more computational resources and handle sensitive data, businesses are re-evaluating their cloud strategies. Many companies are now opting for private clouds, thanks to their cost savings and enhanced security features compared to public cloud services.
- Edge and Distributed Data Centers: These days, more distributed data centers are being established across diverse geographical locations. This trend is being driven by several factors, including cost savings due to reduced power and cooling expenses, as well as the need to address local demands in each region with improved latency. Distributed data centers also play a vital role in disaster recovery.
- Sharp Rise in Data and Transmission: In the AI era, the value of data is soaring to new heights, leading to a skyrocketing increase in its volume and transmission. AI workloads depend on the synchronization of a vast array of compute resources, resulting in an exponential rise in data center traffic. Additionally, the proliferation of distributed data centers has led to substantial surges in the volume and complexity of data exchanged between them.
- AI, Intent-Based, Data Center Operations: As data continues to grow, the demand for data center resources rises correspondingly. Consequently, while the significance of the infrastructure becomes ever more critical, its complexity also escalates. To address this, there is a significant push to use AI to analyze the data and simplify operations by modernizing existing automation tools.
- GPU-as-a-Service: GPU-as-a-Service (GPaaS) is a game-changer for the telecom Infrastructure-as-a-Service (IaaS) business. By offering on-demand access to GPU servers hosted in data centers, operators can provide new services that enable businesses to use their own large language models (LLMs) to drive innovations.
Emerging Challenges for AI Data Center Network for Telecoms
Despite the many opportunities ushered in by AI, the evolving dynamics associated with AI have made it imperative to reconsider how data center networks should be constructed. Here are some of the key challenges operators are currently experiencing in the AI era:
- Modernization towards AI Networking: To synchronize AI workloads across distributed compute resources in real-time, the network needs to be modernized with characteristics such as ultra-low latency, lossless transmission, and high throughput, all while maintaining operational integrity. Additionally, as data centers become increasingly distributed, it is essential to transform not only intra-data center capabilities but also data center interconnects for future-oriented AI networks.
- ROI Improvement for Private Cloud Networking: Due to the shift towards private networking, the workloads in the private cloud are increasing, along with the network demands. To accommodate this growth, it is essential to reassess the robustness, scalability, security, and optimal architecture of the network in the private cloud to ensure a better return on investment (ROI).
- Unifying AI/Automation Framework for Efficient Data Center Operation: AI has the high potential to revolutionize data center operations. However, operators often already utilize various automation tools, and without a proper strategy, the holistic operational flow cannot be optimized, thereby limiting the benefits of AI. Therefore, revising AI and automation strategies is essential to fully capitalize on the advantages of AI adoption.
- Review of Single-Vendor Dependent Approach: Due to geopolitical and business situations, the risk of single-vendor dependency is becoming more evident. Political uncertainty and business mergers and acquisitions (M&As) drastically affect the supply chain. It is therefore important for each operator to diversify their risks by adopting multi-vendor approaches.
- Climate Impact Mitigation for Sustainability: The surge in compute resources driven by AI substantially increases power consumption, presenting new challenges for ESG considerations. To achieve their sustainability goals, operators must implement a variety of measures, such as utilizing renewable energy sources, revising cooling strategies, and optimizing power usage.
NEC’s Global Network Transformation Services for AI Data Center Networks
NEC’s Network Transformation Services manage multi-vendor and multi-domain technologies, ensuring comprehensive lifecycle management, whilst also guaranteeing interoperability and security among diverse systems. Through our holistic portfolio, known as NEC Enterprise Network for Datacenter, we deliver unique value to our customers, including:
- Global Engineering Expertise in Data Center Networking for Optimal TCO: Drawing on our extensive experience—achieved through networking with Tier-1 operators and enterprises in over 150 countries—we collaborate with leading global vendors and our R&D to provide best-of-breed AI networking solutions in any kind of cloud environment. Our approach ensures that the end-to-end IP/optical networks within, and between, distributed data centers are thoroughly optimized for AI workloads, offering secure, ultra-low latency; lossless, high-throughput transmission; and superior TCO for each operator.
- Enabling Unified AI and Automation Workflow for Operational Simplicity: By leveraging our expertise in networking and AI/automation, our Network Automation Service revamps existing workflows and tools, providing a comprehensive automation strategy and seamless integration. This holistic approach empowers operators to optimize their entire workflow rather than adopting a fragmented, piece-by-piece method.
- Promoting Multi-Vendor Approaches to Mitigate Risks Posed by Single-Vendor Dependency: We empower operators to unlock their network by providing services for multi-vendor solutions. Through our services, we ensure seamless interoperability among different vendors and guarantee that the team becomes fully proficient in designing and operating with new vendor solutions.
- NEC as Your Sustainable Partner: Earning us the No. 2 spot on TIME Magazine's 2024 list of the 'World's Most Sustainable Companies,' we are recognized for our eco-friendly practices and environmental stewardship. Committed to promoting a green society, we assist operators in achieving their ESG goals by offering sustainable services such as power optimization and hardware life extension.
With NEC, your data center network will advance to the next level through AI-driven digital transformation bolstered by enhanced networking.
Highlights from NEC Headquarters: What's Ahead for Telecom and Advanced Technology
