- The customer QoE is all about instantaneous speed usually customer dont care on tonnage but the churn is directly related to instantaneous speeds. So during dimensioning how tonnage has precedence over instantaneous speeds?
Speed and tonnage are, in principle, unrelated. I doubt that any customer QoE will be affected by the instantaneous speed. The customer may complain about the speed as this is the only parameter that he knows about. The customer correlates that the network is slow so the speed is low and should be improved. In reality, without knowing he is actually referring to tonnage.
Let’s address the speed issue again, through two examples:
Example 1
Let’s assume a voice connection that uses a vocoder that has a data rate of 12.5 Kbps. This means that a 250 bits packet has to be delivered every 20 ms. LTE performs a bundling and delivers this packet in 4 TTI. The TBS associated with it is TBS=328 bit (Index=6, QPSK), which results in a MAC of 352 bits.
Considering a 10 MHz LTE channel, this packet will be delivered at average rate of 120 bit per RB (Resource Block), meaning an output data rate (speed) of 1.25 Mbit/s. This is the instantaneous speed of delivery of the vocoder packet. If the user was getting this packet at half or even a one hundredth of this speed, he will not notice.
The tonnage continues to be 12.5 kbps though.
Example 2
Consider an Ethernet connection that several routers connected to it. The router has several users connected them.
When a user wants to send a package it will listen to see if there is anyone transmitting. If yes, he has to wait and do a whole procedure to access again. This is called oversubscription and will be discussed in the next webinar.
When he succeeds, he gets the full bandwidth of the Ethernet channel, so his speed is very high. If he has a tester and he is measuring the speed whit which the packet is being transmitted, he can measure 100 Mbit/s.
Let’s assume that each time that the user wants to transmit a 1 kbit packet he has to wait 10 ms to get a free channel. The maximum tonnage he can achieve is 100 kbps.
- Interference consideration is well taken into account during link budget planning and kept at minumum. With increased number of users and random application behavior the user experience varies a lot. How Tonnage can help it?
Link budget does not affect interference. The only way to avoid interference is not to use the same resource in proximity. This means that only part of the resources can be used at each cell. This reduces the tonnage that can be carried, but not the instantaneous speed. This will be covered in much more depth in the next webinar.
- Can you please give an example of GBR with a certain data tonnage?
Voice carrying RAB is considered GBR (QCI=1). A typical GBR for voice is 12.5 kbps.
- In your last table how you go from typical usage to tonnage per hour? For example, what should I do to go from 51MB/hr to 26 MB/month from example on slide 19?
The last column states 26 MB/month/per unit. The unit is specified in the question and it is 1 minute.
(51 MB/hr/60 minute)* 30 days= 25.5 MB/month
- When planning a network, two types of planning are considered ( service Coverage / RF sensitivity coverage) , the second is more considered for prelaunch, the first is considered when dealing with particular service heavily loaded spots. is it true ?
Yes it is true. HOWEVER... This is the commonly used methodology, start the design by looking at RF sensitivity (pre-launch) and after launch consider service coverage. This is NOT the methodology that we advise though because when you are doing and RF sensitivity analysis only you might seriously overlook interference issues that might cause your network to have to be completely redesigned within just aa few months from deployment. Ideally, even for pre-launch you should consider WHERE your users are and their tonnage.
- Is Data Speed calculated from the Data tonnage or vice versa?
Data speed and data tonnage are independent parameters. Data speed defines how fast a packet of data is delivered, while data tonnage defines how much data is sent over a larger period of time.
With regards to the QCI, what category would be given to a smartphone using a speed test app? This may well affect the results of such speed tests.
Yes, the CCI varies from one protocol to another. Speed test are different and each uses a different protocol. Measured speed will differ between protocols.
- If I'm dimensioning Wi-Fi networks; what traffic parameters should I consider? Is oversubscription a factor to consider?
You should always consider tonnage to dimension a network. Unconstrained user traffic should be independent from oversubscription. Oversubscription will affect user traffic, due to the adaptive property of some protocols.
- What is the difference between packet error loss rate and bit error rate in an LTE network?
BER measures the remaining error at the PHY level after applying FEC. BLER is the error rate left after correcting the remaining errors using HARQ. PER is the error left after correcting the remaining errors using ARQ at RRC level. PER is the remaining error that will go to the Internet.
- Would the average number of user that are considered in a network based on the average number of UE's a service provider have sold. Also for green field operators, how do they determine what is an appropriate number of users they should dimension their network for?
Yes, the number of users can be calculated from the number of terminal sold of each type. In Greenfield deployments this should be estimated by the marketing team.
- So, 4G rely on data tonnage rather than data speed
Yes, 4G and 3G should be analyzed considering traffic volume (tonnage). Data speed is a qualifier
|
