Emerging data center workloads drive new infrastructure demands

Last updated:August 2017

Editor's note

Emerging data center workloads -- ranging from big data to machine learning -- offer a number of benefits to the enterprise. It's important to understand, however, how to properly evolve or, in some cases, create an infrastructure that can handle those workloads' demands.

The internet of things (IoT), big data, artificial intelligence (AI), machine learning and other data-intensive workloads can accentuate the limitations of the storage, network and compute capabilities of any infrastructure. Before you implement them, understand potential challenges and how to avoid them. Monitor possible areas of concern, prepare your architecture and take these steps to more seamlessly deploy these emerging data center workloads in your organization.

1Ease into a big data implementation

Big data implementations enable enterprises to gain new insights into customer demands, potential security threats and more. However, don't make an immediate jump into the big data realm. First, take inventory of your current resources, including processing power and memory, to determine if they're capable of accommodating these workloads. Provision resources based on your needs, and add elements into your architecture in installments. Conduct additional research to ensure that when big data workload surges occur, your enterprise can handle the added stress without bottlenecks or shutdowns.

2Data center teams prepare for AI, machine learning

Data center workloads for AI and machine learning continue to emerge, as enterprises recognize the value they bring. But before deployment, data center teams must evaluate their existing architectures to ensure these compute-intensive workloads don't become overwhelming. For example, more advanced machine learning algorithms, as well as deep learning frameworks, can present scalability challenges. To address them, IT teams need to evolve their storage and networking architectures and also prioritize automation.