Dimensioning of mobile data - Non-uniform usage. Dimensioning of voice services (Thanks to Claes Beckman)

Transcription

1 Time Dimensioning of mobile data - Non-uniform usage Dimensioning of voice services (Thanks to Claes Beckman) Jan Markendahl December, 06 Time Time 00:00 0:00 0:00 0:00 04:00 05:00 06:00 07:00 08:00 09:00 0:00 :00 :00 :00 4:00 5:00 6:00 7:00 8:00 9:00 0:00 :00 :00 :00 Call Attempts

2 Today Dimensioning not assuming uniform usage Non uniform usage profile Conditions set by bit rates Erlang type of dimensioning Criticism of simple HW model Capacity of a cell System bandwidth Spectral efficiency Spectral reuse Today Dimensioning not assuming uniform usage Non uniform usage profile Conditions set by bit rates Erlang type of dimensioning Criticism of simple HW model Capacity of a cell System bandwidth Spectral efficiency Spectral reuse 4

3 Capacity dimensioning Mobile broadband Montly demand of MBB spread out - all days of the month - all 4 hours of the day For data NRT data traffic the approach with average data rate per user can be used X GB per user and month -> Y kbps per user Time 5 Capacity dimensioning Mobile broadband Montly demand of MBB spread out - all days of the month - out of 4 hours of the day Time 6

4 Capacity dimensioning Mobile broadband Montly demand of MBB spread out - all days of the month - 8 out of 4 hours of the day Time 7 Capacity dimensioning Mobile broadband Montly demand of MBB spread out - Non-uniform during 4 hours - Peak hours morning and evening Time 8 4

5 Capacity dimensioning Mobile broadband Montly demand of MBB spread out - Non-uniform during 4 hours Office area Time 9 Capacity dimensioning Mobile broadband Montly demand of MBB spread out - Non-uniform during 4 hours Residential area Time 0 5

6 Capacity dimensioning Mobile broadband Units 00 Montly demand of MBB spread out - Non-uniform during 4 hours Example area 50 5,5 Time Today Dimensioning not assuming uniform usage Non uniform usage profile Conditions set by bit rates Erlang type of dimensioning Criticism of simple HW model Capacity of a cell System bandwidth Spectral efficiency Spectral reuse 6

7 Conditions set by bit rates What happens if there are requirements on data rates? Examples: 0 users per cell Mbps per user 0 Mbps per user A short exercise: how many users can be served, - in a cell with capacities as below? Spectral efficiency 5 MHz of Spectrum 0,7 bps/hz,5 Mbps A: 70 users B: 4 users 0 MHz of Spectrum 0 MHz of Spectrum 7,0 Mbps 4 Mbps,0 bps/hz 0 Mbps 0 Mbps 40 Mbps A: 800 users B: 50 users user demand A: [ 5,4 GB/month; 8 hrs/day] => 50 kbps per user B: [,6 GB/month; hrs/day] => 800 kbps per user 4 7

8 Today Dimensioning not assuming uniform usage Non uniform usage profile Conditions set by bit rates Erlang type of dimensioning Criticism of simple HW model Capacity of a cell System bandwidth Spectral efficiency Spectral reuse 5 6 8

9 Cell planning and re-use D D a) K= 7 b) K= 7 Bandwidth for differents systems LTE.4MHz MHz 5MHz 0MHz 5MHz 0MHz Re-use: ( with fractional frequency reuse) UMTS, WCDMA 5MHz carriers 5MHz 0MHz 5MHz 0 MHz Re-use: GSM, N 00 khz carriers Voice: 8 time slots (calls) per carrier Data: 9,6-59, kbps per time slot max downlink 4*59, kbps = 6,8 kbps/carrier Re-use:,, 4, 7, 9 8 9

10 Today Dimensioning not assuming uniform usage Non uniform usage profile Conditions set by bit rates Erlang type of dimensioning Criticism of simple HW model Capacity of a cell System bandwidth Spectral efficiency Spectral reuse 9 Spectral effiency Peak data rate ~0-0 bps per Hz Average data rate ~ - bps per Hz Cell border rate < 0,0 bps per Hz 0 0

11 Spectral effiency and cylinder model Spectral effiency and cylinder model

12 From Ericsson: Capital markets day, May 008 Maximum throughput LTE Spectral efficiency (bps/hz),68,6 5,04 0,08 0,6 4

13 Dimensioning of mobile data - Non-uniform usage Dimensioning of voice services (Thanks to Claes Beckman) Jan Markendahl December, 06 5 Capacity dimensioning The busy hour Capacity that is deployed Blocked traffic 00:00 0:00 0:00 0:00 04:00 05:00 06:00 07:00 08:00 09:00 0:00 :00 :00 :00 4:00 5:00 6:00 7:00 8:00 9:00 0:00 :00 :00 :00 Call Attempts Time 6

14 Traffic An Erlang is a unit of telecommunications traffic measurement. Strictly speaking, an Erlang represents the continuous use of one voice path. In practice, it is used to describe the total traffic volume of one hour. Example, if a group of users make 0 calls in one hour, and each call had an average call duration of 5 minutes, then the number of Erlangs this represents is worked out as follows: Minutes of traffic in the hour = number of calls x duration Minutes of traffic in the hour = 0 x 5 = 50 Hours of traffic in the hour = 50 / 60 =.5 Traffic figure=.5 Erlangs 7 Erlang B Several traffic models exist which share their name with the Erlang unit of traffic. They are formulae which can be used to estimate the number of lines required in a network, or to a central office (PSTN exchange lines). Erlang B is the most commonly used traffic model, and is used to work out how many lines are required if the traffic figure (in Erlangs) during the busiest hour and the number of blocked calls are known. The model assumes that all blocked calls are immediately cleared

15 Excercise - Erlang How many Erlang do YOU produce yourself during your most busy hour? 9 Specified by ETSI Frequency Division Duplex TDMA system Originally at 900MHz, but today also at 800, 800, 900 and some other bands world wide MHz 0 5

16 Basic Frame Structure BCCH TCH TCH TCH 8 time slots (voice channels) /TDMA carrier time slot / cell used for signaling (BCCH) Symbol rate 7 kbps Cell planning and re-use D D a) K= 7 b) K= 6

17 Number of channels to use - Short exercise How many voice channels per site in the cases A. Bandwidth,8 MHz, re-use factor 7, omni antenna B. Bandwidth, MHz, re-use factor, omni antenna Number of channels to use - Short exercise How many voice channels per site in the cases A. Bandwidth,8 MHz, re-use factor 7, omni antenna No carriers:,8/0,= 4, No Carriers per site: 4/7 = => *8 = 6 voice channels (no signalling) B. Bandwidth, MHz, re-use factor, omni antenna No carriers:,/0,= 6, No Carriers per site: 6/ = A. => *8 = 6 voice channels (no signalling) 4 7

18 Network Traffic Load Network Capital requirements are a function of peak demand 0mErlang is the statistical traffic per subscriber in busy hour. Minutes of use is the sum of all traffic, i.e the blue area Service Revenues are a function of average demand 5 Blocking in Cellular systems Typically we plan the voice capacity of our mobile phone systems for % blocking rate (98% availability) 6 8

19 An example Bandwidth, MHz, re-use factor, three sectors Assume that one channel will be used for control No carriers: Total No cannels: 6 Usable No channels: 5 % blocking => 9 Erlang How many 0mE users can be served? 9/(0*0 - ) = 450 How many 50mE users can be served? 9/(50*0 - ) =

20 The examples we did Dec A and B from before, both leading to carriers How many 00 me users can be served for cases K and L Case availability No Ctrl channels No GPRS channels K 90% L 99,9% 4 9 The examples we did Dec A and B from before, both leading to carriers How many 00 me users can be served for cases K and L Case availability No Ctrl channels No GPRS channels K 90% L 99,9% 4 Blocking prob No of voice channels Offered traffic K 0% 6--= 4,47 E L 0,% 6--4=,65 E No users (00mE) K 4 L

21 4