Speakers
Description
The Data Acquisition systems are an integral part of the readout electronics of high energy physics experiments. Conventional data acquisition (DAQ) systems that rely on server polling of distributed clients suffer from high latency, wasted bandwidth, and inefficient resource usage. Polling results in delays that relate to the frequency of the queries and does not scale well as the number of clients increases. As the client count grows, polling becomes inefficient and leads to delays that are proportional to the query interval. This paper introduces a proactive, client-focused, event-driven architecture that overcomes these shortcomings.
Texas Instruments TM4C1294 microcontrollers are used as client nodes within the proposed system. It detects events and promptly transmits data packets to a central server. The server, built using Python and LabVIEW, is constantly monitoring to process incoming data in real time.
In comparison to a polling baseline of 100 ms, performance assessments clearly demonstrate advantages.
The event-driven approach achieves a 94.7% reduction in average latency (from $53.2 ,\text{ms}$ to $2.8 ,\text{ms}$), a 96.21% reduction in jitter (from $29 ,\text{ms}$ to $1.1 ,\text{ms}$), and a 80% decrease in server CPU usage (from $5%$ to $1%$), with comparable memory consumption.
These results confirm that the client-push model enables faster, scalable, and resource-efficient DAQ systems, meeting the stringent requirements of modern distributed and industrial applications.
| Position | Student |
|---|---|
| Affiliation | Sardar Patel Institute of Technology |
| Country | India |