1 / 14

Expected IFLB RAB Distribution under Ideal Conditions

Expected IFLB RAB Distribution under Ideal Conditions. Introduction. IFLB is used to control the RAB distribution between LTE frequencies Current practice uses a basic model to predict the distribution That model does not consider lbThreshold , the minimum load need to trigger offloading

dasan
Télécharger la présentation

Expected IFLB RAB Distribution under Ideal Conditions

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Expected IFLB RAB Distribution under Ideal Conditions

  2. Introduction • IFLB is used to control the RAB distribution between LTE frequencies • Current practice uses a basic model to predict the distribution • That model does not consider lbThreshold, the minimum load need to trigger offloading • This over site is significant at typical loading levels and has lead to the erroneous belief that the feature is not working • In this presentation we will • Calculate the expected distribution ratio using the current method • Expand the method to include lbThreshold • Compare both methods with field results

  3. Parameter settings 1 1 qciSubscriptionQuanta / ∑qciSubscriptionQuantaCell 2 cellSubscriptionCapacity x lbThreshold / qciSubscriptionQuanta

  4. lbThreshold Impact at Low load • Consider two cells with an IFLB relationship enabled • RAB’s are added one at a time to Cell A • Cell A can have 10 UE’s before off loading to Cell B • This means for less than 10 UE’s the RAB distribution will be 100% on Cell A • Clearly this is very far from the expected distribution

  5. Modeling lbThreshold Impact • Our model must include the 100% distribution at low load and progress towards the Expected Distribution • We do this by considering two extreme situations • Only adding RAB’s to Cell A • Only adding RAB’s to Cell B • To facilitate the formation of a mathematical model we make the following assumption: • When the number of RAB’s is greater than the threshold, RAB’s will be distributed proportional to the Expected Distribution

  6. Only adding to Cell A • By only adding to Cell A, Cell B must have 0 RAB when below the lbThreshold (10 RAB) • Then for every RAB on Cell A, there must be one on Cell B as the Expected Distribution is 50% • Formula to describe this is: • RB = MAX(0,RA x QA / QB - TB) RX Number of RAB on Cell X QXQCISubscriptionQanta for Cell X TX Threshold for activation for Cell X [RAB]

  7. Only adding to Cell B • By only adding to Cell B, Cell B must have 10 RAB before the first appears on Cell A (10 RAB) • Then for every RAB on Cell A, there must be one on Cell B as the Expected Distribution is 50% • Formula to describe this is: • RB = MAX(0,RA x QA / QB + TB) RX Number of RAB on Cell X QXQCISubscriptionQanta for Cell X TX Threshold for activation for Cell X [RAB]

  8. Plotting Results • Formula for distribution of RAB on Cell A is straight forward • RA / (RA + RB) • Plotting shows we do tend to the Expected Distribution while also modeling the predicted low load behavior

  9. Parameter settings 2Field Settings 1 qciSubscriptionQuanta / ∑qciSubscriptionQuantaCell 2 cellSubscriptionCapacity x lbThreshold / qciSubscriptionQuanta

  10. Field Results • Results were collected from an IFLB field trial • Average RRC Connections was used as a proxy for the number of RAB • Cells had up to 8 IFLB neighbors • The median value for RAB Distribution was 43.3% on Frequency A • The design was for 58.3% • So let’s examine factor in the number of RAB’s…

  11. Field Results • The plot below shows that the number of RRC connections significantly impacts the distribution of RAB’s. • The distribution of RAB’s for cells with greater than 40 RCC connection’s is close to the Expected Distribution • The majority of Freq A cells are below the 40 RRC Connection point

  12. Field Results vs Model • The model reasonably predicts the spread distribution with low loading • It suggests that the “Add to Cell B Only” phenomenon is dominate in the network • This indicates that Freq B is attracting initial attaches • Idle mode users having a preference for Freq B is one possible explanation

  13. Conclusion • We demonstrated that the impact from the load balancing threshold is significant at low loading • We developed a model to more accurately predict low load distribution • We demonstrated the expected load balancing distribution will fall in a range of values • That median of RAB distributions is not an accurate indication of IFLB design efficacy • That field results are likely being impacted by • Many cells with a low number of connections • Freq B attracting initial connections

  14. THNAK YOU

More Related