Artificial intelligence is becoming increasingly popular in today’s working world (especially in everyday office life). The hype surrounding the seemingly almost unlimited possibilities is manifested in a growing number of business models that use AI as an essential element. But one question remains unanswered: Is it best to run my AI applications in the cloud or using edge computing?
First of all, the question arises: What exactly is edge computing and what is it good for?
The advantage of edge computing is its local proximity to the company using it, which keeps latency low for large amounts of data. In addition to AI, typical areas of application are IoT (Internet of Things) and gaming.
One disadvantage is the usually limited computing capacity and scalability compared to cloud solutions. This can make resource-intensive data processing and storage a challenge.
Individual decisions are required
The different operating models therefore bring with them fundamental differences. The scalability of the public cloud enjoys almost cult status, but so does the powerful, customer-oriented data processing within the framework of AI-based process optimization.
Depending on the application and requirements, different aspects may be relevant for the operation of the company’s own AI.
212 companies told us in which areas they see cloud or edge computing as an advantage:
Cloud strengths:
- Security
- Cost
- flexibility and scalability
- stability and reliability
Edge strengths:
- control over your own data
- risk minimization
- Low latency
Cloud security surprises!
The respondents’ assessments of security are particularly astonishing: almost six out of ten companies (57 percent) see cloud computing as an advantage here. The cloud has already become widely established in all its forms, so the contacts know the risks and can assess them better. Edge computing, on the other hand, is a relatively new phenomenon that many IT decision-makers have not yet given much thought to.
Edge Computing is underestimated
Geographical proximity to the user company is a central aspect of any edge approach. The most obvious added value of an edge solution, especially with regard to the operation of artificial intelligence, is therefore low latency. In addition to immense computing power, AI applications require a fast flow of data in order to work efficiently. However, the IT decision-makers surveyed see edge (52 percent) as having only a minimal advantage over the cloud (48 percent).
In addition to speed, opportunities for risk minimization (51 percent) and control over one’s own data (59 percent) speak in favor of edge computing. In this context, the greater security concerns of the contacts are particularly surprising. Not least because at the same time, increasing the level of security is by far the most frequently cited reason for operating an AI model using edge computing (52 percent).
Five Questions for the Perfect Fit
- How extensive is our knowledge of cloud and edge computing?
- How flexible are the procurement models of the edge providers?
- Do we focus on cost efficiency or investment effectiveness?
- How important is low latency to us?
- How sensitive is the data used?
Conclusion
Edge computing is still a largely unexplored field for many IT decision-makers. Obvious added value is partially recognized, but the great potential – for example in terms of data security and the stability of high-performance infrastructure – remains unrecognized.
In order to choose the ideal operating model for your own AI solution, edge models should be considered in addition to public and private cloud. A knowledge advantage regarding edge topics can bring a significant competitive advantage in order to achieve the best price-performance ratio for individual company requirements.
