How IoT and Edge Networking Are Shaping Data Centers
Want to learn more on how IoT and edge computing will change data centers? Click here to learn more about edge computing and the future of data centers.
Join the DZone community and get the full member experience.Join For Free
Everyone has been talking about the Internet of Things (IoT) and devices that have become smart enough to send data collected from their surroundings to a cloud service for further processing. The driving thought behind the innovation is that these devices would lead to more efficient processing and the ability to catch problems as they occur. Because of IoT, data centers now have to revamp the way they think about an operation. Given the rise of the IoT, the whole paradigm of data centers may require a shift going forward in order to deal with this added increase in load from the IoT devices. Intel estimates that, by 2020, we may be seeing as many as 200 billion connected devices. With this much data to manage, the traditional paradigm of the data center will be forced to adapt.
Development of the Data Center
In the past, and largely at present, data centers have been a central location from which data flows downstream to connected devices. Data centers, in this way, serve as points for our devices to connect to and to use the bandwidth there to retrieve anything we want that's on the Internet. With the advent of the IoT, these centers are now serving as two-way highways since data is also being sent to these locations from the IoT-enabled devices. This requires a lot more upstream bandwidth in order to deal with the enormous incoming flow of data — far more than what our devices upload to the Internet. To this end, the data centers will now have to adapt to the changing paradigm by providing a faster upward communication stream to cope with the massive data flows from the IoT devices.
Is Edge the Solution We're Looking For?
Edge computing, according to General Electric, refers to the infrastructure that resides close to the existing data architecture and is tasked with ensuring that data collection and analytics are processed at the source of the data, usually by leveraging devices that are not always connected to the network as a gateway. In simple terms, what Edge computing does is it limits the amount of overhead processing the network requires in order to get things done. In a case where a company is using a cloud platform to do their IoT data processing, this is very important since not all of the data being transmitted is going to be relevant to the decision-making process. Furthermore, companies that utilize cloud platforms for processing are charged based on how much data is processed by the platform, making it in their best interest to limit the amount of data being sent for processing. In fact, in properly functioning IoT devices, there should be no anomalous data that causes decisions to be made. Edge deals with this problem by sifting out the true but irrelevant data before it's sent for processing, lowering the costs of processing significantly.
Practical Implementation of the Edge Case Scenario
PC Magazine notes that edge computing speeds things up immensely, especially if the connection to the overall wider network isn't so good. It does this by providing a repository for data that is both close to the user and able to handle responses from the user at a faster rate. In the case of an application that connects to something remotely using M2M connectivity, like for example a mine or an oil platform, this can be a major consideration because of the remoteness of the user location and the poor network coverage in those areas. By combining computing and analytics at the edge of the connection location, it creates a more stable user experience as well as provides a more solid use of the edge network and cuts down on the amount of data that needs to be processed. If anything, we can think about these locations as batch-processing centers, to borrow an analogy from the earlier days of computing. The amalgamated data can then be shunted in payloads to the processing center when enough data has been collected from the user.
Internet of Things and the Data Center of the Future
Data centers have long been a mainstay of computing architecture, and one of the myths surrounding the implementation of the IoT was that it would mean the demise of the data center. This can be both true and false. The data center, as we have traditionally known it, will have to change in order to provide the requisite architecture to the Internet of Things, but at the same time, they won't be phased out, because they are still a very necessary part of how data is passed around the Internet. While the paradigm is likely to shift away from a receiver-centric model into one that handles edge computing, those data centers are still likely to be a major constituent of the future of computing, working hand in hand with the IoT to provide a better standard of living for all.
Opinions expressed by DZone contributors are their own.