Like many buzzwords, Fog computing sounds a little woolly and nondescript. But make no mistake, the term is here to stay.
Fog computing is defined as “extending cloud computing to the edge of an enterprise’s network”, and facilitating “the operation of compute, storage and networking services between end devices and cloud computing data centers.”
The term was first coined by Cisco to denote cloud computing being closer to the physical place where the data is generated and acted upon.
With fog, previously known as distributed computing, data is collected by devices at the edge of the network and then sent to the datacenter/cloud later. By doing this, the data does not have to be sent to the cloud immediately (reducing stress on the system), and it can also result in a better application performance from the local IoT device in use.
So, we know what Fog computing is but what is the benefit? Well, most benefits are actually around the deficiencies of cloud computing.
Traditional cloud computing architectures – and that process of moving all data from the network edge to the datacenter for processing— adds latency (lag). There are bandwidth issues (if lots of IoT sensors are located in one place, like an offshore oil rig for example, traffic can outstrip bandwidth capacity), as well as concerns of data integrity, reliability and security.
There are industry regulations to adhere to and privacy concerns around the use of certain types of data. Cloud servers also only communicate with IP, and not over the countless other protocols used by IoT devices.
Subsequently, the best place to analyze most real-time data (before being sent to the ‘big’ cloud or datacenter) is locally, near the devices that produce that produce and act on that data. This is fog computing.
This analysis takes place on ‘fog nodes’ – local devices. Any devices with computing, storage and networking connectivity can be a fog ‘node’. They could be switches, routers, video surveillance cameras or industrial controllers.
IDC estimates that the amount of data analyzed on devices that are physically close to the Internet of Things is approaching 40 percent.
Fog computing could be used to lock doors, change equipment settings or apply brakes on the train. There are so many more potential use cases.
There are now industry efforts to push this concept further. Fog computing has its own working group, the OpenFog Consortium, and recently, Cisco, Intel, Microsoft, ARM, Dell, and Microsoft backed a new consortium called the OpenFog initiative.
It’s not a silver bullet – fog could be cheaper, but is also in danger of being a sprawling and expensive mess if it goes wrong. We’ll have to see if these tech giants can drive it to success.