Accommodating fast-changing requirements
30. January 2024
5 Min read
New DC approaches and architecture
To meet bandwidth demand, migration to 400 G / 800 G continues. Where 40 G (4×10 G) / 100 G (4×25 G) requires eight fibers in parallel pairs (parallel optics principle within Data Center – Tx/Rx), higher speeds require 16 or 32 fibers (8×50 G / 8×100 G). This boosts cable density, and smart migration paths, high-density products, pre-configured solutions, and monitoring/asset management are key to 400G/800G, as also the upcoming panel and cable structure will move from Base 12 to Base 8 application including new cable set up with 16 fiber structure rather than 12f or 24f. DC operators need to utilize valuable space and racks more efficiently as density keeps increasing. Very Small Form Factor (VSFF) Connectors (SN/MDC), ribbon fibers as slim as 82 microns, and high-density connectors and closures are key to this. With thousands of cables running into the DC, fast, easy, splice-free connections are becoming more important. Solutions such as pre-term cabling and new push-pull fiber connector types significantly reduce handling and installation time, guarantee functionality, and increase first-installation quality. Preconfigured cabinets with power, cooling, security, and connectivity offer a neat solution for modular DC approaches.
Digitization is profoundly affecting Data Centers through the availability of new tools and information, and a need to accommodate fast-changing requirements. Decentralization and Hybrid / Multi-cloud strategies are becoming more prevalent, as new applications will use different types of DC elements such as those close to the data source (edge), on-premises DC, and private and public cloud – all in one. AI has a significant effect on DC design and operation, driving Spine-Leaf architecture, which efficiently handles rising data traffic and ensures low-latency connections, but also could become part of managing the increasing complexity of DC, especially in the light of higher efficiency requirements set by adapted norms and governmental guidance. Devices are divided into two layers: spine and leaf. All leaf and spine switches are connected, forming a non-blocking high-bandwidth network.
Equipment is continuously added, moved, or replaced as DCs scale and change, making accurate, real-time visibility into processes and assets difficult. Automation reduces the need for manual intervention, and minimizes errors, particularly in the areas of server provisioning, resource allocation, and performance monitoring. As the scale and complexity of data centers grows, reliance on Automated Management Systems and Data Centre Infrastructure Management solutions increases. Introducing IoT and asset and capacity management functions has made DCIM essential to operational activities. An ‘expert layer’ can pull meaningful KPI-related insights from across DC systems and present these in understandable and actionable ways. Administrators can analyze large datasets to gain insights into trends, usage patterns, and optimization opportunities. DCIM can also support compliance with standards and anticipate potential issues before they result in non-compliance. This could become a professional preparation for regularly required re-certifications. Incorporating AI and AR into data center asset management can lead to enhanced resource utilization and decision-making capabilities. «Digital Twins» can be used to optimize DC design and building.
Newly developed applications demand 5G, low latency, and high-speed connectivity, which enable the growth of edge computing. They require a network of smaller, distributed data centers that can process and store data locally, reducing latency and bandwidth use and becoming a part of the hybrid Data processing from the source of data to the cloud. An IoT-driven surge in data production is pushing data centers to scale up storage capacity and develop more efficient data processing capabilities. To manage and derive insights from this data, DCs are increasingly relying on AI and ML algorithms, automation, and orchestration.
The Data Center Manager will be placed in a challenging position. On the one hand, they will have to comply with ever-increasing application and infrastructure demands on the one hand, while on the other hand, they are being asked – even pressured – to do this more efficiently. A hybrid data center approach would merge facilities very close to data sources with algorithmic and statistical functionalities in the cloud, helping realize this.
DCs are diversifying, scaling up, and requiring new approaches while simultaneously requiring standardization. Smart solutions can facilitate this, without compromising on performance and reliability. However, informed decision-making is a must.
Learn more about R&M’s integrated data center infrastructure solutions: https://www.rdm.com/solutions/data-center/