Abstract
- An end-to-end testbed is developed by integrating ETSI-compliant MEC and OpenAirInterface (OAI) infrastructure, that triggers integration of multi-standardization support in a single architecture.
- An open-source robust 5G testbed, which is easily deployable, portable and hardware independent
- Demonstration of a real-time applications i.e a video streaming application, addressing object detection scenarios has been demonstrated to illustrate its operability
- Support for network slicing has also been incorporated into the testbed.
Introduction
- Multi-access edge computing (MEC) must be leveraged to meet low-latency and high-reliability goals. Subscribers can receive value-added services by supporting computation and storage near users.
- The MEC architecture proposed by ETSI caters to key requirements for the next generation of applications.
- Also, the OAI kinds of initiatives are increasingly becoming the vehicle for 5G deployments, allowing network customization and real-time monitoring.
- Seeking a flexible architecture for wireless networks, this paper puts forth a novel 5G MEC-OAI testbed adhering to the existing standards.
- The 5G MEC-OAI testbed is easily deployable, portable and hardware independent that can be used in network slicing scenario.
- A video streaming application has been demonstrated on this architecture, addressing object detection scenarios.
5G MEC-OAI Testbed Setup
MEC System Setup:-
The architecture is based on ETSI MEC standards[1]. The components deployed on the virtual machine are as follows i.e. UALCM, MECH, MECO.
5G OAI Testbed Setup:-
The OAI platform provisions complete implementation of the cellular networks that are 3GPP standard compliant [2]. It is also beneficial for testing real-time scenarios.
References
[1] [3GPP, “Architecture for enabling Edge Applications,” 3rd Generation Partnership Project (3GPP), Technical Specification (TS) 23.558, 05 2022, version 17.3.0. [Online]. Available: https://www.etsi.org/deliver/etsits/123500123599/123558/17.03.0060/ts123558v170300p.pdf
[2] M. Etsi, “Multi-access edge computing (mec); framework and reference architecture,” ETSI GS MEC, vol. 3, p. V2, 2019.
[3] open air interface, “5g software alliance for democratising wireless innovation,” 2014.
Acknowledgement
This work is a part of a Ministry of electronics and Information Technology (MeitY) funded project, Next Generation Wireless Research and Standardization on 5G and Beyond.
Conclusion
- Proposed an open-source 5G MEC-OAI testbed, with the MEC being ETSI-compliant.
- A discussion of the existing standards was followed by the experimental setup.
- The objective is to support real-time, low-latency applications in a reliable manner via an open-source testbed that is easily deployable and hardware-independent.
- An application has been deployed on the testbed, addressing object detection use cases. The edge capabilities can be enhanced with the integration of network-slicing technology.
Bharti School of Telecommunication Technology & Management
Industrial Significance
Patents, collaborations
Technology Readiness Level:
Work in progress
A 5G OpenAirInterface (OAI) Testbed with MEC:
Deployment, Application Testing and Slicing Support
Priyal Thakkar, Shashvat Sanadhya, Pimmy Gandotra, Brejesh Lall*
Figure 1: 5G MEC-OAI testbed deployment setup.�Integration with OAI 5G Core
The setup in Figure 1, illustrates a single UPF (OAI Minimal setup), which is used to execute the integration between the ETSI MEC system and OAI 5G core [3]. �
Result
- The proposed testbed is appropriate for numerous applications that require high processing power, intensive storage, and data offloading strategies. Two computer intensive applications i.e. person detection and vehicle detection (Figure 2) have been deployed.
- UE sends live feed video to the MEC host server for object detection.
- The time required to send frames to the MEC host is reduced by 96%.
- This framework can be used for a variety of latency-compute applications.
Figure 2: Object detection algorithm deployed on MEC Host.
Communication Technologies