Mobile traffic forecasting for maximizing 5G network slicing resource utilization
read more
Citations
Network Slicing and Softwarization: A Survey on Principles, Enabling Technologies, and Solutions
Channel State Information Prediction for 5G Wireless Communications: A Deep Learning Approach
AI-Driven Zero Touch Network and Service Management in 5G and Beyond: Challenges and Research Directions
DeepCog: Cognitive Network Management in Sliced 5G Networks with Deep Learning
A Comprehensive Survey of Enabling and Emerging Technologies for Social Distancing - Part I: Fundamentals and Enabling Technologies
References
VLSI module placement based on rectangle-packing by the sequence-pair
From network sharing to multi-tenancy: The 5G network slice broker
NVS: a substrate for virtualizing wireless resources in cellular networks
Radio access network virtualization for future mobile carrier networks
SLAW: self-similar least-action human walk
Related Papers (5)
Frequently Asked Questions (16)
Q2. What future works have the authors mentioned in the paper "Mobile traffic forecasting for maximizing 5g network slicing resource utilization" ?
The forecasting solution designed builds on HoltWinters theory to predict future traffic levels per network slice, considering different service classes, and is effective in maximizing the number of requests that can be admitted by the an admission control decision engine. Their main findings can be summarized as follows: i ) Holt-Winters theory can be effectively applied to network slicing traffic forecasting both for regular and irregular slice requests, ii ) elastic traffic network slice requests help in increasing the maximum achievable system utilization, iii ) the forecasting benefits increase as the number of network slice requests and system capacity increases, and iv ) low SLA violation risk levels result in very significant system utilization gains.
Q3. Why does the network slicer outperform the legacy scheme?
As soon as the network becomes congested, i.e., some network slice requests must be rejected, the utilizationof their proposal outperforms the legacy scheme (GF 0) due to a wider distribution of network slice request values.
Q4. What is the key objective of the novel network slice traffic scheduler?
The key objective of their novel network slice traffic scheduler is to minimize consumed resources while guaranteeing the traffic SLAs within a network slice.
Q5. What is the main idea behind the 5G network slice broker?
To support a signaling-based slice allocation, certain 3GPP interfaces need to be enhanced (Type 5 and Itf-N) to enable the instantiation and configuration of network slices, indicating the time duration, the required resource amount, and additional requirements, such as, e.g., the slice SLA.
Q6. What is the complexity of the Network Slices Packer?
Given that the knapsack problem solution can be solved with O(n log n) computational time, the complexity of the Network Slices Packer is dominated by O(n3 log n).
Q7. What is the function used to monitor the traffic levels of the network slice?
The list of granted slice requests is sent to the Slice Scheduling module, which allocates network slice physical resources and monitors (with a penalty history function) the served traffic levels and potential SLA violations.
Q8. What is the key aspect for an efficient network slice admission control mechanism?
A key aspect for an efficient network slice admission control mechanism is to accurately predict the tenants’ trafficevolution in the near future.
Q9. What is the only information used for the forecasting?
When no forecasting solution is applied (w/o forecasting) or during the training period (for adjusting the forecasting algorithm parameters), the only information used are the SLA requests.
Q10. Why do the authors prefer best-effort slices to be accepted?
The authors observe that the total number of admitted slices increases with the number of best-effort slice requests, showing that best-effort slice requests are preferred due to the higher flexibility.
Q11. What is the effect of the forecasting algorithm?
The effectiveness highly depends on the accuracy of the forecasting algorithm: the more accurate, the more aggressive the authors can be in leasing available resources while keeping a small probability of violating slice SLAs.
Q12. What constraint is included in the Flexibile Geometric Twodimensional knapsack problem?
In addition to the constraints included in their probelm formulation, the Flexibile Geometric Twodimensional knapsack problem also includes an additional constraint on weight capacities.
Q13. What is the effect of the forecasting-aware Network slicer?
when irregular shape patterns are considered, the time complexity of the Forecasting-aware Network slicer further increases.
Q14. What is the way to predict traffic levels?
Due to the penalties imposed by traffic SLAs, the authors focus only on the upper bound of the prediction interval as it provides the “worst-case” of a forecasted traffic level.
Q15. How can the simulated annealing algorithm change the permutations?
With this representation, the simulated annealing algorithm can easily change the permutations by checking at every step kk whether the new locations are (i) feasible and (ii) provide a greater objective function value, i.e., ΔF = Fkk+1(x)−Fkk(x) >
Q16. How can the forecasting process be more accurate?
the forecasting process can easily categorize the traffic requests based on the associated service requirements, thereby performing a prediction separately per slice.