Abstract
In this paper, we present cloud VoIP scheduling strategies to provide appropriate levels of quality of service to users, and cost to VoIP service providers. This bi-objective problem is reasonable and representative for real installations and applications. We conduct comprehensive simulation on real data of sixty four strategies with dynamic prediction of the load. We show that our prediction rule that consider the number of Virtual Machines (VMs) running in the system improves the efficiency of traditional rate of change algorithm. It provides suitable quality of service and lower cost. Variations of VM startup time delays permit to evaluate the prediction rules under different scenarios and assess the robustness of all strategies. © 2017 IEEE.
Original language | English |
---|---|
Pages | 119-123 |
Number of pages | 5 |
DOIs | |
Publication status | Published - 2017 |
Event | 4th International Conference on Engineering and Telecommunication - Moscow, Russian Federation Duration: 25 Nov 2020 → … |
Conference
Conference | 4th International Conference on Engineering and Telecommunication |
---|---|
Abbreviated title | En and T 2017 |
Country/Territory | Russian Federation |
City | Moscow |
Period | 25/11/20 → … |
Keywords
- bin packing
- call allocation
- cloud voice over IP
- load prediction
- quality of service
- Costs
- Forecasting
- Internet telephony
- Voice/data communication systems
- Bin packing
- Dynamic prediction
- Load predictions
- Prediction rules
- QoS optimization
- Scheduling strategies
- Voice over IP
- Quality of service