Decentralized Data Processing How NASA came in for a landing on the moon on Edge Computing
Even if “Edge Computing” as a modern techno is perceived to be a logical approach in contrast to the Cloud, and Core (data center), it is a principle that has already been applied in the Apollo Mission.
Edge Computing has increasingly become an important part of IT. The CTO of Suse, Dr. Gerald Pfeifer, formulated the role as follows: “the Edge is no longer a niche, to the other take care of, but a third full-fledged modality in addition to the data center, the Core, and the Cloud.” Compared to the Core and Cloud Edge leave, however severe the inherent nail, or restrict, except that it is none of the other two cases (Core or Cloud). Supermarket checkouts fall under the Edge concept, as well as measuring systems, mobile phones, or cars with IT on Board, as well as aircraft.
In retrospect, a lot of “Edge”was
So much was, in fact, looking back, “Edge”, even if there was this concept of the rule of law at the time. Pfeifer estimates that there are Edge Computing for over 50 years. “I would also include the on-Board computer of the first Apollo moon landing, and even before there was the Internet, as such, and where the communication over the radio was, or Point-of-sale environments in the trade,” says the CTO. The difference in the time course characterized by the fact that Edge solutions carved originally mostly “by hand”, and very specific, often quite static. Today, however, standardization or re-use to be return the same way in the technology environment, such as the increasing requirements due to short development cycles.
Data processing at the edge of the network
Ultimately, it is the Edge Computing to distributed data processing at the edge of the network, the results will only be forwarded. The opposite of the Trend that companies increasingly rely on is, not raw data worked to save, because you can’t predict, with the Capture of the data in part yet, what insights can be Data or Predictive Analytics won. The approach depends on the Background of this stress field according to Pfeifer strongly on the particular application and the available infrastructure. You ask yourself questions like: Is personal data, and how it looks to the resulting obligations (keyword: DSGVO/ GDPR)? How many of data, and there is enough bandwidth or the local cache available? The value will be sought, and what is the relationship of this to the more effort and the more costs?
Data capture, Transport and storage
According to the Suse CTO today, it is immensely easy to capture virtually any amount of data to generate. “In addition to the initial processing, Transport and storage, however, remain an issue. I expect that this will remain so for the foreseeable future,“ the Linux professional. The development of Edge Computing in the course of time outlined Pfeifer, that Edge was formerly part of a separate and static, “mostly by thin Connections, connected, quite specialized and separated from other infrastructure to be treated”. In contrast, significantly more dynamic and networking today on the agenda. “Solutions span over the Edge, data center and Cloud. Data are diligently replaced, solutions, agile created and further developed and individual applications are distributed.“
Containers and Kubernetes
It is beneficial, explains the expert, if environments are similar, or even (from the perspective of the developer and operator) are identical or at least compatible. “Modern approaches such as containers as building blocks and Kubernetes as flexible platforms to offer here. A Kubernetes Cluster in my car is going to look differently than the one on my ERP or Webshop runs, but a similar feel.“ Piper compares the difference between a compact car and a tractor-trailer, which also have other characteristics, wherein the function of the pedals of the car radio, or the rules of the road similar but strong or identical. This allows it to develop locally or in the cloud and test in advance, to push functions to the place where most of the data concerns, or to make devices at the Edge, Autonomous.
The interaction of Edge, Core and Cloud
Edge, Core and Cloud are not independent of each other, but in a dynamic exchange and cooperation. On the technical side, this means that issues such as orchestration, high availability, data and system security must cross this Ensemble of Edge, Core and Cloud covered. “I expect that here we see red threads: Kubernetes will also be at the Edge of the De-Facto Standard, not the least specialized, leaner distributions, such as K3s. Linux will continue to replace proprietary stand-alone solutions“, is Pfeifer convinced.