Internet of Things providing an edge for micro data centers

New data capture and analytics techniques thrown up by the Internet of Things (IoT) are having a profound effect upon the way back office information technology (IT) shop engineering is developing and the data center is going through rapid changes.

By Adrian Bridgwater August 20, 2016

The growth of the Internet of Things (IoT) has helped continue to drive the need for data center construction and development. The streams of data produced by the proliferating variety of devices has to reside somewhere and most typically it has to reside "in the cloud," as the expression goes. That cloud is the cloud data center and the cloud data center houses the server racks that hold the IoT.

Advantages of "edge" computing

Cloud data centers do a great job if looking after IoT data, but one of their limitations is that they are not completely virtual i.e. they have to physically exist somewhere on planet Earth and this it turns out is their Achilles heel. Not just for reasons of compliance and governance, very often we want data to reside ‘on the edge’ i.e., close to its original source. This is precisely why we talk about so-called Edge Computing.

Wouldn’t it be great if the data center could be brought closer to the sensors, devices, and networks that are producing IoT data?

It’s a small world after all

Micro data centers are compact prefabricated pieces of hardware and management software intelligence designed to live in a single unit. In terms of form and function, micro data centers boast many of the components found inside a full blown data center but scaled down and in some areas rationalized for size.

Inside the micro data centers ‘box’ we will find processing power, memory, input/output (I/O) intelligence, cooling systems, uninterruptible power supply (UPS) hardware, security software, and telecommunications power.

Manufacturers producing micro data centers include Huawei, Zellabox, Dell, Canovate Group and SGI (formerly Silicon Graphics).

Dale Green, Digital Realty’s marketing director, said all IoT initiatives require a tailored data center strategy that balances current needs with future growth and potential applications. "As the Internet of Things localizes data streams down to the smallest wearable device, many organizations are growing increasingly concerned with latency and the emerging need to maintain data centers within physical proximity to their customers. In situations such as this, data and computing infrastructure will need to reside in close proximity to users and the devices, while still being able to connect directly to trading partners and the digital supply chain," he said.

Benefits of Web-scale techniques

Paul Phillips, regional senior director for Nutanix, thinks that the explosion of consumer devices requiring cloud based services presents an insurmountable challenge to the traditional three-tier architecture that the data center has relied upon for the past two decades. "Google, Facebook and the other new wave of cloud based companies recognized this early on and chose a different path with what is often referred to as web-scale technology," Phillips said.

Phillips said that the ability to scale in small increments, seamlessly, and with 100% uptime can only be achieved using the same Web-scale techniques these companies pioneered. "However, this technology needs to be consumable, easy to manage and simple to implement if it is to be pushed to the edge, and yet intelligent enough to be managed centrally as a single platform right out of the box," he said. 

Adrian Bridgwater is editor at Internet of Business, which is hosting the Internet of Manufacturing Conference November 1-2, 2016, in Chicago. This article originally appeared here. Internet of Business is a CFE Media content partner. Edited by Chris Vavra, CFE Media, cvavra@cfemedia.com.

ONLINE extra

See additional stories about the Internet of Things (IoT) linked below.

Original content can be found at www.internetofbusiness.co.uk.