What’s next in the evolution of the Edge Compute Continuum?
The internet of things world and the solutions it constantly breeds are increasingly becoming a part of our everyday lives. Our home thermostat can be programmed remotely and learns from the temperature patterns in the house. The fridge makes sure we know if we’re running out of food. And our blinds are cued to open every morning to the tune of Bohemian Rhapsody. On full volume.
In order to make these IoT interactions possible, new coding strategies and enterprise iot solutions have emerged. MQTT leverages simple messaging for low-bandwidth networks. API-first turns product development on its head by putting the API first, before the product. Intelligent Edge computing takes each edge from being a data collector and aggregator to functioning like a remote brain.
So what’s next for the internet of things world? It’s a concept called the Edge Compute Continuum, and no it is not a concept from Star Trek. In fact, we expect it to become the new norm for all things IoT and IoT ai.
The Edge Compute Continuum represents many layers of computer infrastructure that can be leveraged as part of a whole IoT application. This leverage provides a hierarchy-based fastest path to every device.
How does the Edge Compute Continuum work?
New concepts in the internet of things world build on the backs of previous technologies. The Edge Compute Continuum is simply the IoT application of a concept called ubiquitous computing. Ubiquitous computing can be anywhere and everywhere, like a past fling you can’t seem to avoid, and it is device-, location-, and format-agnostic.
To understand the Edge Compute Continuum, consider the path of any online interaction you might have, even one as simple as visiting a webpage. In the process of accessing the webpage, you may only perceive your interaction with a server. But, there are A LOT of hops along the way between your browser and the destination content.
When looking at a static website, you may never communicate with a hosting server, but instead, a cached version stored in a CDN. All the devices that enable those hops - the routing gear, the cell phone towers, the satellites - all represent computing opportunities for enterprise IoT solutions.
Take Netflix Open Connect CDN - consumers use it to watch a wide variety of movies and tv shows. However, it takes a long time to pull all that video content from a central cloud to homes around the world. Instead of taking the long route by retrieving from the cloud, Netflix Open Connect’s hardware uses local ISPs to keep large caches of content. Granted, Netflix is a different use case than most IoT applications, but it is similar in that IoT will use tremendous amounts of bandwidth at high speeds.
How will the Edge Compute Continuum manifest in 2018?
Edge computing options are currently solo operations, meaning that edges don’t talk to each other. But with the forward momentum of the Edge Compute Continuum in the internet of things world, edge computing offerings capable of chaining together will start to enter the market. Instead of costly one-off solutions, we’ll see middleware capable of making the edge computing connection process easy and transparent for developers. Ubiquitous computing will be made available as standard development architecture.
What does that really mean? Let’s say your car runs an IoT traffic application, and it wants to know the status of traffic on I-35 near Austin. The car might first ask the wifi router in your parking garage. If the edge computing platform running in the wifi router has recently answered this question, it may respond quickly to your car. If not, it may request this information from a cell tower where another edge sits with a broader set of constituents.
The cell tower may be able to answer the traffic request if it has recent information, or it could ask a local ISP data center. This ISP data center is the target of local traffic data, so it knows the status as data on vehicle movement is already sent to it. Instead of sending this request (and all that data) up to the cloud, the ISP can run the cloud logic locally. This means that functions previously performed only in the cloud can now run close to the data source.
The ISP data center can later push the data up to the cloud for consumption by other cities or keep it local to reduce computing power waste. And the driver never has to know how their question was answered. The car’s IoT application is free to run the optimal path to get the answer - fast.
You can see how interconnected the internet of things world really is.
Are we talking about edge or fog computing?
While the scenario described above does sound a lot like fog computing, the fundamental element of a fog is unstructured and unorganized communication. While a communication model where everyone speaks to everyone is great for a social network, it fails when it comes to enforcing truth and guaranteeing task completion.
Consequently, the Edge Compute Continuum remains a different animal - forcing a communication hierarchy, guaranteeing a source of trust, and ensuring that every device that sends or requests data gets an accurate, responsive channel of communication.
This year, we will continue to see IoT applications in everyday products, homes, and businesses. As more silicon providers make devices cheaper to manufacture and distribute, the demand to better utilize IoT infrastructure will rise. A coherent offering from IoT middleware vendors making the edge available across the continuum will be the theme of the internet of things world in 2018.