计算机科学
微服务
云计算
可扩展性
软件部署
服务(商务)
计算机网络
分布式计算
服务发现
万维网
软件工程
操作系统
Web服务
经济
经济
作者
Saidulu Aldas,Andrew Babakian
出处
期刊:IEEE Access
[Institute of Electrical and Electronics Engineers]
日期:2023-01-01
卷期号:11: 132286-132295
标识
DOI:10.1109/access.2023.3335994
摘要
The 3GPP has defined the 5G network infrastructure differently from earlier generations, utilizing a Service-Based Architecture (SBA) to accommodate scalability and ensure cost-effectiveness in a cloud and microservices-friendly design. Each 5G Network Function (NF) operates as a service, seamlessly scaling to support an on-demand load and latency requirements. However, this introduces significant complexity to the infrastructure due to an influx of messages across 5G services, making request handling and resiliency challenging. While the 3GPP has provided guidelines and recommendations to address this complexity, it is left to the vendors to develop a working model and design choices within the cloud-native framework. This paper explores the intricate challenges that 5G encounters within the current cloud-native deployment landscapes, with a focus on service-to-service communication, service discovery, and network programmability. It highlights the essential role of Service Mesh in fulfilling 3GPP’s vision, proposing enhancements tailored for 5G deployment scenarios. In addition, this work provides architectural guidance for enhancing Service Mesh frameworks to support 5G Core Signalling, extending the Envoy proxy to meet 5G’s unique semantic requirements. Finally, this paper sheds light on the future requirements for developing optimal Service Mesh infrastructure that minimizes performance implications while maintaining seamless scalability for cloud-native environments.
科研通智能强力驱动
Strongly Powered by AbleSci AI