Redefining Data Centers at the Edge

By: Abhishek Nagar | Co-written by: Shivakumar Vaidyanathan

Bigstock

Walled-off, centralized, physical data centers still serve a purpose, but they are no longer the star of the show. As the Internet of Things (IoT) grows and the 5G networks expand, data no longer sits at the center. It now lives at the edge. In our connected world, it’s imperative that we adapt data center architecture to support the current climate of data demand. Businesses must redefine their concept of data centers to match this new reality.

 

 

Cloud Load, AI Adoption and IoT

What has created this new reality? With the introduction of intelligent applications, the need for fast computing has increased.

By 2019, all the people and things connected to the internet will generate 507.5 zettabyes of data. This increases the compute load on the cloud.1 That load creates challenges, such as latency and bandwidth demand, which is compounded by the physical distance between connected devices and data centers.

Artificial intelligence (AI) adoption is yet another challenge. Global spending on AI and cognitive systems is projected to be $57.6 billion in 2021, with spending expected to achieve a compound annual growth rate of 50.1 percent in the 2016–2021 forecast period.2

Internet of Things (IoT) devices continue to proliferate, with Gartner estimating 20.4 billion IoT devices in use by 2020.3 That figure doesn’t include smartphones, tablets or computers. With that many devices, organizations face the challenge of collecting and storing large volumes of data, as well as the need to process and analyze it.

The spread of 5G networks is expected to significantly increase data transferred to data centers, augmenting the need for intermediary servers. The 5G network will likely be the underlying fabric that will support the ecosystem of internet-connected devices.

Infrastructure for the Connected World

Today’s data center needs to keep up with the pace of cloud-based business and big data. At present, roughly 10 percent of enterprise-generated data is created and processed outside of a traditional data center or cloud. Moving compute resources closer to the point of data generation allows for more efficient data processing because processing power is moved closer to the source.

Rather than being centralized, this new concept will support a multilayered and distributed data infrastructure. In an edge infrastructure, the objective is to reduce latency and unnecessary traffic, and to establish a hub that connects peers and thins out computational loads. With an edge data center, compute power becomes localized and more widely distributed. The edge computing paradigm places applications, data transfer, storage and services at the network edge. The result is reduced transmission costs, greater privacy, lower latency and a better user experience. By moving closer to the data source, the business gains faster insight and can generate better, more informed business decisions.

Getting to Know the Edge

Edge computing moves information processing and content collection and delivery closer to the sources of that information. The goals of edge computing are to reduce latency and unnecessary traffic, as well as establish a hub for interconnection between interested peers. This creates data thinning of complex media types or computational loads.

These new edge data centers will offer high availability and adjust for decentralized control and management. The critical factors businesses will look for are scalability, reliability, efficiency and resiliency, all while ensuring data remains secure. Ease of configuration will be a crucial asset for edge data centers. Enterprises will find value in preconfigured or standardized packages that allow rapid deployment and simplified relocation. This ability will speed up and simplify the plan, design and build process for these centers.

To make the creation of edge data centers easier and more convenient, businesses will seek solutions that offer flexible installation options that don’t take up much room or require extensive site work. Organizations will look for the right fit for their business requirements, with a range of choices for power density and rack options. Edge data centers will also need to be easy to manage and monitor, making it easier for current IT staff to manage their operation. Remote monitoring will be a key feature sought by enterprises. The use of automation will ease monitoring and control of the entire physical infrastructure.

Data center architecture and infrastructure will keep pace with how technology and data creation change. As it does, we will redefine our notions of what a data center is and how we will implement that change to redefine our businesses.

IBM Business Resiliency Data Center Services can help you create a robust data center architecture that meets the demands of an always-on, always-connected world. Contact us to schedule a consultation.

1. http://www.businessinsider.com/edge-computing-in-the-iot-forecasts-key-benefits-and-top-industries-adopting-an-analytics-model-that-improves-processing-and-cuts-costs-2016-7

2. https://www.idc.com/getdoc.jsp?containerId=prUS43095417

3. https://www.gartner.com/newsroom/id/3598917

Topics: , , ,

About The Author

Abhishek Nagar

IBM Global GTS Product Marketing Manager, Business Resiliency Services

Abhishek Nagar has over nine years of B2B marketing experience across both products and services in software industry. In his current role, Abhishek leads global product marketing role for Data Center Services, Business Continuity Consulting and Backup as a service for the IBM Services business. He has been awarded and recognized as a strong and... Read More