Accelerating the
Industrial Internet of Things

How Cloud Infrastructure Would Need To Address IIoT Hyperscale Demands

Published on 07/28/2016 | Market Sizing

370 3

Ravi Ravichandran

Ravi Ravichandran is the SVP of SaaS/Cloud Services, Customer Success and Operations at GE Digital and Predix Cloud where he leads GE's digital transformation, cloud operations, and infrastructure| devops.



Based on a study conducted by the Market Research Company IDC, the internet of things is estimated to pose a 750 percent data capacity increase. If it happens as per forecast, the entire data industry is likely to face huge implications. Nonetheless, investments in the IIoT platforms in the company’s data centers remains critical to delivering full functionality. With rising data generation and the number of devices connected, it’s probably the right time for businesses to focus on their IIoT requirements both at the Servers and Data Centers.

Security requirements

Being dubbed as the next industrial revolution, analyst’s forecasts have put the expected gains from IIoT at $14 trillion over the next decade. It’s no surprise why corporate entities are in the rush to harness the benefits through new systems, products, and services. One of the key driving benefits is the ability to establish a connection of smart, interlinked sensors and switches that are used to manage a broad range of industrial machines. Once done, the plant can change with the prevailing conditions in real time or predict issues that may affect operations.

While connectivity remains a key driver of the IIoT, one cannot wish away the presence of cyber-attacks that compromise the functionality. Whether it’s accidental or malicious, the fact that the systems connect to corporate networks and the Internet exposes them to risks of attacks that inhibits the full potential of the platforms. In some instances it is even recommended that IIoT cloud infrastructure be isolated from the internal workloads to prevent compromises.

To secure the systems, consider integrating security features to the respective edge devices through a specific software, remote management, monitoring and alerting capabilities. This extends the traditional IT environment all the way into the edge thereby creating a door for a new set of vulnerabilities, incidents to monitor and manage proactively. This extension of corporate IT boundaries is challenging to manage in this hub and spoke model of IIoT. Where there could be thousands of spokes leading into edge devices. We thus now need software that needs to ensure continuous secure code life cycle updates, remote device boot management, communication protocols and authentication and continuous monitoring and management.

Hyperscale requirements

As with increase data storage requirements, users may need to customize their servers to suit a given purpose. The hyperscale servers come with racks exceeding the traditional standard of 42U racks hence accommodating more elements across the servers all the way up to 50U+ racks. For this reason, most of the cloud-based companies such as Facebook, Amazon and Google have embarked on building in house high density racks with enough power to scale up to accommodate the rising demand. The datacenters now need to accommodate high density racks all the way from loading docks to the customer cage as the traditional datacenters cannot even have entry to roll in this racks.

While the hyperscale servers can use a few of the traditional rack switches, Facebook’s Open Computer Initiative calls for new Photonic Connectors and Rooted optical modules. The new network designs and servers save the time taken on troubleshooting lower power losses and allow users to scale up the capacity in small bits.

Infrastructure Scale requirements

Over the last few years, the world has seen a rise in the level of investment in the IOT. IBM’s investment in the internet of things is set to reach $3 billion to run the industrial internet operations. By 2020, IDC estimates the IoT market solutions to reach $7.1 trillion increasing the demand for infrastructural expansions to support the exploding requirements.

While most users anticipate the vast benefits that come with the new IOT platforms, the reality is that the complexity and quantity of the IIoT devices need similar investments in the communications technology and data center facilities. Despite a growing perception that high performing and reliable connection needs massive investments, the future of the current data centers relies on the adoption of smaller and scalable facilities conveniently placed near the customers and workloads at the edge similar in principle to content delivery networks to address latency, performance and just in time actionable predictive analytics. Also, the type of data center required ought to shift from the large centralized facilities to the small edge data center services.

Feature New Record