Tactic: Apply edge computing

Tactic sort: Awesome Tactic

Type: Architectural Tactic

Tags: edge-computing measured

Title

Apply edge computing

Description

Moving computing resources closer to users decreases the latency. Furthermore, the system can be designed in a way that only processed/aggregated data need to be transported which reduces the amount of data traffic. Transporting less data over the network is expected to reduce the energy consumption.

Participant

Cloud consumer

Related software artifact

Classification software

Context

Edge versus cloud-only

Software feature

Data processing

Tactic intent

Decreasing data traffic, to increase performance and energy-efficiency

Target quality attribute

Performance

Other related quality attributes

Energy-efficiency

Measured impact

According to estimations, the energy consumption of the data processing and ML classification is relatively similar in the edge and cloud-only scenarios. The energy consumption of the data transport, on the other hand, differs several orders of magnitude when comparing the on edge versus cloud-only scenarios. The reduction in data transport is also a main motivation for applying the edge architecture. Therefore, we argue that, in a scenario where large volumes of data need to be processed, applying an edge architecture has a positive effect on the energy consumption of the workload. In this specific case study a difference of 21.242 kWh was identified between the cloud-only and edge scenario, indicating a decrease of the energy consumption of 96% when using the edge scenario.

Source

Master Thesis “Architectural Tactics to Optimize Software for Energy Efficiency in the Public Cloud” by Sophie Vos