The role of the cloud and fog computing in the emerging Internet of Things (IoT) and how they should interplay with each other has been a frequent topic of debate within the IT community. We’ve discussed fog computing on Interconnections and the strategic role it can play in the success of the IoT. However, not every IoT use case is tailor-made for fog computing.
Fog computing operates between end devices and clouds, and it’s all about distributing some computing, storage and analysis out at the cloud’s edge, close to the “things” that are constantly creating and processing mountains of data for analysis and action. Certain functions, such as the Industrial Internet, are perfect candidates for fog computing because big data analytics can track various sensors, detect anomalies and predict and address budding operational or machine failures or performance issues. For example, a jet engine failure can be predicted and avoided well in advance, if maintenance and flight data is collected and analyzed by interconnected IoT systems, databases and analytics to prognosticate the failure of a specific part and its impact on the workings of the engine.
But in some life and death healthcare scenarios, high-latency delays in the roundtrip of data between local devices and remote cloud-based IoT systems can create cases where vital patient information, which can impact treatment options, cannot get into the hands of doctors fast enough. Critical and sensitive patient data needs to be processed, analyzed and protected locally, rather than placed, somewhere in the “fog.”
Cloud and fog computing have their place in the IoT
Cloud and fog computing are still essential for the success of IoT, as most existing organizations and corporate data centers simply cannot offer the computing and storage scale required to deploy a successful IoT predictive analytics solution that can cope with massive amounts of sensor data. Cloud services such as Amazon, Microsoft, IBM, Google and Oracle are a few providers that can deliver the humongous, on-demand scale required and the ready-to-run IoT stacks and PaaS architectures organizations need but cannot build up themselves.
In addition, some “cloud-like” hardware and software resources can be put at the network edge, where data can be ingested, processed, analyzed and acted on immediately, to supplement fog computing and lower latency, thus reducing the risk of delays. As part of such efforts, major players such as Cisco, Intel, Microsoft and Princeton University and ARM have formed the OpenFog Consortium. The group’s goal is to develop standards, APIs and reference architectures for successful fog computing scenarios, including gateways for ingesting, managing and analyzing data from local sensor networks.
Leveraging an interconnection-first strategy with the IoT, cloud and fog
At Equinix, we see a complementary solution Â¾ the marriage of IoT, the cloud and fog with an Interconnection Oriented Architecture™ (IOA™) strategy deployed on Platform Equinix™. With an interconnection-first strategy, IT infrastructures can be extended via Equinix Performance Hubs in distributed Equinix colocation data centers in major metropolitan centers. When necessary, organizations can deploy a thin client side node to collect, process and forward data securely to the nearest cloud service or Performance Hub.
Direct and secure connectivity to multiple cloud and fog services can be provisioned using software-defined, virtualized connections via the Equinix Cloud Exchange. A proximate, multi-cloud infrastructure delivered by the Cloud Exchange allows organizations to split workloads among cloud services to maximize efficiency and minimize costs. In addition, high-speed, direct connections eliminate the latency problem and, at times, the need for specialized gateways.
Directly and securely interconnecting local IoT data repositories via the Equinix Data Hub allows the lightning-fast ingestion, integration and analysis of local IoT data. Organizations can gain valuable operational and customer insights from the huge amounts of IoT data being created at the edge of their corporate networks without backhauling the data over the public internet to centralized data centers, and risking delays and security breaches.
As Frost and Sullivan points out in its report, The Fog Rolls In: Network Architectures for IoT and Edge Computing, Equinix has developed a set of open deployment reference architectures for these purposes, with tools such as templates, patterns, and architectural blueprints, as well as cookbooks and deployment guides. To facilitate seamless end-to-end IoT data migration, we are also working with cloud partners, such as Microsoft Azure, on platform-specific blueprints.
Fog computing will no doubt play an important role in successful IoT-cloud-fog deployments, but organizations should also consider the benefits of an IOA strategy when diving into this new frontier.
Read “The Fog Rolls In: Network Architectures for IoT and Edge Computing” to learn how interconnection can help you create high-performance and secure cloud and fog infrastructures for the IoT.