Datasets
Standard Dataset
Data Rate Performance Measurements of NFV Cloud Native Scaling for a Media Application
- Citation Author(s):
- Submitted by:
- Steve Taylor
- Last updated:
- Wed, 06/09/2021 - 06:00
- DOI:
- 10.21227/2e1c-rd87
- Data Format:
- License:
- Categories:
- Keywords:
Abstract
This dataset accompanies a paper that discusses the advantages of a 3GPP-compliant service-based architecture platform that demonstrates the concept of cloud-native service orchestration and routing for a media vertical sector application. Cloud-native service orchestration and routing is a complete end-to-end approach that enables virtualisation and management of multiple layers in the OSI model, which provides considerable flexibility and control to achieve delivery of QoS to users in the face of varying demand, at reasonable cost. The discussion is motivated by a customer-facing trial conducted in Bristol, UK where the platform was deployed as virtual network functions and the vertical service using the cloud-native orchestrator. The architecture of the system is presented, together with an exemplary media application illustrating the benefits of dynamic control of the whole service function. Three scenarios are described: a baseline case, where a single service instance handles requests from multiple clients at different locations, which can become overloaded with a consequent degradation in user experience; a static horizontally scaled-out service scenario, where service instances serving content are placed closest to users on edge hosts with associated performance gains in terms of reduced response time and networking costs, but with increased hosting costs; and a dynamically-managed horizontal scaling case, where storage service instances are enabled automatically based on location-specific demands when needed. It is illustrated that dynamic service instances based on sensing QoS metrics provides an opportunity to achieve a balance between user experience benefits of edge service provisioning and acceptable costs through avoiding waste of cloud resources at the edge.
Please refer to README.md
Documentation
Attachment | Size |
---|---|
Readme | 3.29 KB |
Comments
great
None
great