You might have heard about Fog Computing by now if you are a Cloud and Grid Computing geek . Though not everybody is familiar with it, Fog Computing is slowly developing from a concept to something more practical — network-level architecture. Simply put, it is: a technology which enables data processing in smart devices.
In recent years, Cloud Computing has emerged as a one-stop solution for companies looking to scale their computational operations without investing in new infrastructure. It has proved to be more resourceful for businesses looking to cut down on their operational costs by offshoring their data centres to third-party cloud services.
Grid Computing, the actual parent of Cloud Computing, is architecture suited mostly to central authorities like system administrators and integrators. These are the authorities who are familiar with the operations running at system-end; they install and upgrade the servers and applications, and do all the required maintenance using specific software.
Both Cloud and Grid Computing have their specific set of limitations. They are related to hidden costs, network latency and — the most important — security and privacy. For instance: Cloud Computing is mostly criticized for its centralised models, where consumers’ data remain affixed to the whims of monopolized organizations. Furthermore, in its core model, all the applications and processing are centered entirely in the cloud.
Why Fog Computing Makes More Sense
Can we meet the growing demand for server farms as the number of connected smart devices continues to go through the roof? The answer is no. The computing technologies would always need to evolve, just like Grid evolved into Cloud.
Fog Computing could then be called the grandson of Grid Computing, the son of Cloud Computing. The reason why Fog Computing is called Fog Computing, metaphorically, is that Fog is evidently closer to the ground than Cloud is. This means that in the case of Fog Computing, all the data processing takes place at the network level, right in the smart devices.
And indeed, Fog Computing allows smart devices to be — you know — smart. So instead of processing some task using the cloud, one can use all the devices that surround us: from personal computers to smartphones, and from coffee makers to refrigerators.
SONM, a project which also places Fog Computing at the centerpiece of its development, presents an interesting case study in the same regard. The project creatively hybridizes Fog with an open-source PaaS technology named Cocaine. The funny-titled tech has a built-in locator service which helps to monitor and search services within a cluster.
“Given: you use a smartphone (4 cores, 2 GHz, 64 GB storage); half of the resources you own are not used,” the lead developer of SONM explains. “SONM container-app Docker estimates the resources which are not used. You can either provide the app with an access to all idle resources you have at any moment or limit the resources for SONM usage.”
“The devices which are close to yours (this is determined by Locator services built in SONM architecture) can use each other’s idle data and this will be a kind of cloud storage while the devices are knit geographically as well.”
With the said experiment, SONM proves that the buyers of computing power enjoy more cost-efficient solutions than any centralized cloud service can offer. The project plans to take the use cases of Fog Computing further by exploring into scientific projects, site hosting, game servers, neural networks, video rendering and computer graphics technology.
The SONM developers plan to utilize Fog Computing in solo mining, stating that miners “will stop burning [their] kilowatts for PoW mining and start serving calculations for everyone in the network.”
This means you will not have to wait too long to see the benefits of Fog Computing. Time to turn into a geek!
Thanks for a great post, I love the idea of fog computing. Where do you see SONM going? What other companies are looking into the area? I assume alot.