I am beginning to hear a lot of sound from the technology press warning that cloud computing is at risk from emerging edge computing technology. As somebody who works on both sides of the community, I need to clean a few things up.
Indeed some processing is shifting to the edge of the network, as it is logical. For example, take a vehicle management system that includes two sides: the edge computing side and the cloud/central computing facet. Both have various functions.
The edge computing side that is in the vehicle needs to respond immediately to changing data in and about the vehicle, such as an impeding crash or weather-related dangers. It does not make sense to send that info all of the way into a central cloud server, where the decision is made to apply the brakes, then back to the vehicle. By then you’ll have hit on the semi.
However, edge devices are typically much lower powered, with limited storage and compute capabilities. Deep learning processing and predictive analytics to determine the best strategy to automobile maintenance based on petabytes of historic data is best done on back-end cloud-hosted servers. See how that works?
The edge computing market will continue to grow. A report on the topic, sponsored by software provider AlefEdge, pegs the size of this edge-computing market at more than $4 billion by 2030. At the exact same time the cloud computing market is going to be 10 times that, and you’ll come across the growth of both markets more or less proportional.
Edge computing needs cloud computing, and the other way round. Really, public cloud computing suppliers will benefit from the use of edge-based systems, providing small cloud service replicantssmaller edge-based version of cloud services.
These edge-based cloud replicants are likely to be the path of least resistance to begin an edge computing-based project. Plus, they will come with built-in native cloud security, governance, and management.
Moreover, there’s the public cloud private clouds, that have become de facto cloud-connected edge-based systems too. Each one of the major providers have them today, and they are indeed edge apparatus, in the truest sense of this term. Cloud replicants are going to be a subset of native cloud features that live in data centers.
A few conclusions can be drawn from this: Edge computing is an old architectural concept that’s found new legs recently thinking about the price of microcomputing (a Raspberry Pi costs less than a genuine raspberry pie), therefore we are able to build architectures which are more responsive and resilient to outages.
Central higher-powered computing services existing in the cloud are still of enormous value, both for the scalable computing and storage infrastructure, and also for the requirement to centralize the storage and analysis of information.
Mic dropped.
