What’s Different About Fog Computing Architecture?
As we covered in a recent post, previous approaches to data processing have traditionally been broken down along geographic lines: on-site servers, and off-site data centers, the latter of which has morphed into the monolithic cloud industry we’re used to today. While each offers its own benefits, both on-site and cloud computing are inflexible in their own way: on-site being expensive, difficult to manage and scale, and cloud being geographically distant for connected devices that need near instant data processing.
Modern fog computing infrastructures address those pitfalls by making available a continuum of compute that spans device to cloud. Instead of being tethered to either end, Software Publishers can leverage the speed and bandwidth of local compute for traffic-heavy IoT, for example, and the relative affordability of the cloud for longer-term use cases like machine learning. In between, a spectrum of processing power can be consumed on demand, used for its varying properties of speed, availability, bandwidth, cost.
Where does the compute in “fog computing” come from?
Fog computing provides a ‘best of both worlds’ architecture by filling in the gap between on-site and cloud, but where does that compute come from?
Some fog (or its relative, edge computing,) architectures involve purchasing additional fog or edge nodes that are placed closer to devices that need compute. A distinguishing feature of ActiveAether’s fog computing architecture is that it leverages untapped power on the devices that are already located there.
By installing ActiveAether, any computer owner can act as a “host provider”, renting out idle processing power on the computers in their offices, homes, even under-utilized data centers in the same way Amazon or Microsoft rent out processing power on the computers they own. ActiveAether facilitates payment (through FogCoin) to Host Providers for their contribution when their resources are used, and Host Providers are able to offer competitive pricing because they’re simply making better use of a resource they already own.
What’s different about how services are deployed in the fog?
Resources on the Internet are typically identified with a URL — a static physical endpoint which, for any number of reasons, may be temporarily (or permanently) unavailable. The use of URLs promotes the development of non-portable, tightly-coupled, and platform-specific applications which, in one fell swoop, negates all of the benefits of the Web Services model.
With ActiveAether, software services are instead identified with a URI, abstractly describing a service which at any point in time may have zero or more active endpoints. This relieves the consumer from the tasks of locating and binding to software service endpoints through the introduction of a mechanism which robustly performs these tasks on their behalf.
All Software Publishers need to do is upload their service and its requirements, and ActiveAether will take care of where and when services are deployed. That orchestration functionality is what enables software services to be matched with any number of ideal Host Providers within ActiveAether’s fog platform dynamically, instead of being written to one stagnant location in either an on-site or cloud data center.
Publishing Software to a fog computing platform
In order to provide a software service today, one must be able to write software, have access to an internet-facing server, know how to set up and configure that server, deploy the service, monitor it for liveness and responsiveness, and then manage and maintain the system in perpetuity. This is far too heavy a set of burdens for any one role, and requires far too much domain-specific knowledge for any individual to do everything well.
ActiveAether splits the traditional Service Provider role into two separate roles: Host Provider and Software Publisher. In this way individuals may provide hosting resources and well-configured environments, or may simply write software, but they don’t have to know how to do both well in order to participate in the provision of software services.
Get Started with Fog Computing
Fog computing is already providing much needed flexibility for today’s exciting, compute-heavy trends like IoT and Machine Learning. But enabling Software Publishers to automatically and efficiently deploy web services is a upside in any vertical. The same is true for ActiveAether’s Host Providers, who are earning revenue simply by renting out resources on devices they own anyway. As more of the services we interact with daily are deployed in the fog, we’ll all benefit from a less consolidated internet, that won’t be crippled by one downed Amazon data center.
If you’re a Software Publisher interested in leveraging the fog, get started by signing up with ActiveAether now! If you’re a computer owner, register here as a Host Provider for more information on renting out your processing power in ActiveAether’s global marketplace for compute.