Cigref publishes the report “Edge Computing and Post-Cloud Perspectives: Opportunities and Implementations”, the result of the reflections of its working group led by Emmanuel GAUDIN, CIO of the Lagardère Group, with the support of technical expert Kim Nguyen, enterprise architect at Enedis.
This report – a summary of the findings of Cigref’s “Post-cloud and edge computing perspectives” working group – is intended for all employees who wish to gain a better understanding of this subject in order to be able to identify possible implementations of edge computing as applied to processes, services or customer offers.
Edge Computing and Post-Cloud perspectives: opportunities and implementations
Today, large companies and public administrations are transforming themselves in order to exploit the value of data from various sources – sensors, complex machines/systems, human experts, environmental, etc. – gathered within the ecosystem created by their suppliers, customers, vendors and partners. At present, this business-generated data is mainly processed in the cloud. However, the exponential growth of such data raises the question of where it is to be stored and processed. For the cloud has several limitations: physical (latency, bandwidth, etc.), legal (data ownership and liability) and economic (transfer costs, storage, processing costs). Hence the emergence of edge computing, which is developing to meet these challenges.
Edge computing, a response to multiple challenges
Edge computing enables data to be processed as close as possible to its source, either directly by the object itself that produces and captures the data (the car, telephone, etc.), or nearby in small local datacenters; in other words, on the periphery, or “edge”, of the network. Edge computing, although not always directly linked to cloud computing, offers options that are complementary to it. Indeed, some edge computing solutions are completely independent of the cloud. The “post-cloud” era does not mean the decline of the cloud; far from it! But in this era, there will be an even faster growth in data stored and processed as close as possible to their source. Although the cloud continues to be the heart of the network, we already know that not all data will pass through it.
Provide storage, analytical processing, AI and automation capabilities close to the data source
Edge computing is an architecture that provides storage and processing capacity at the edge with analytical features, decentralised intelligence and automation. This architecture makes it possible to take into account the issues of data sensitivity, cyber threat management, real-time management for very low latencies and local data management, and provides a response to issues of bandwidth or network connection, resilience and local decision-making management. However, edge computing projects will face several challenges in extracting value from various data sources, regardless of whether or not these are cross-referenced with one other: recovering data, managing it interoperably, maintaining network connectivity, ensuring optimal security, deploying on a large scale, managing complex systems and reducing the carbon footprint of digital technologies.
Implementing an edge computing project: decision criteria
The participants in Cigref’s “Post-cloud and edge computing perspectives” working group have listed the following criteria to be considered when assessing whether the best type of architecture for a particular service or use case is indeed an edge computing solution:
- benefit of using data locally versus in the cloud;
- latency time required;
- bandwidth required for connectivity;
- transfer costs;
- computing and processing capacity;
- safety constraints;
- data sensitivity;
- carbon accounting / carbon footprint.
Key milestones to consider in order to maximise the value of edge computing
Next, the participants recommended several steps for deriving maximum value from the implementation of the edge computing project, from a technical, economic and human point of view. These include:
- searching for and identifying the service or added value provided by the product;
- building robust business cases;
- determining key competencies for deriving value;
- supporting change related to edge projects and managing the impact of such projects on the organisation and its employees through the implementation of strong governance. The importance of data availability and quality requires all teams to adopt a “technology and data” mentality.