There are three major challenges in performance measurement namely: Defining performance measurements and KPIs, Verifying measurement results, Explaining measurement results. In this article i provide an insight into BLOCK ERROR RATE (BLER) MEASUREMENTS.
Block error rate (BLER) is an analysis of transmission errors on the radio interface. Using
formal definitions BLER is a KPI, because it is a formula. It is based on analysis of cyclic
redundancy check (CRC) results for radio link control (RLC) transport blocks and computed
by defining the relation between the numbers of RLC transport blocks with CRC error indication
and the total number of transmitted transport blocks as expressed in Equation below;
BLER is measured separately on the uplink and downlink direction, which is mandatory, because in UTRAN frequency division duplex (FDD) mode uplink and downlink data is transmitted using different frequency bands.
There is no measurement report in UTRAN that contains uplink BLER values although UL BLER is an important criterion for the radio network controller (RNC) to make handover decisions based on uplink transmission quality.
Due to the fact that UL BLER is only computed and used in the RNC internally it also is only available inside RNC software and as a rule is not shown in any performance measurement statistics. Therefore, it is a typical example for performance measurement based on protocol analysis. The UL BLER is especially a very critical parameter to measure user perceived quality of services using RLC transparent mode. UL BLER can also be used to estimate the uplink transmission quality of a call. In this case it is not necessary to differentiate between transport channels that carry user plane and control plane information. Merely count transport blocks and their CRC indicators according to the standard formula given at the beginning of this chapter, no matter to which transport channel single transport blocks belong.
Compared to uplink block error rate Downlink (DL) BLER does not need to be computed by any performance measurement equipment based on transport block counters. Neither it is meaningful to compute it on lub, because transmission errors will appear on Uu – after data has been sent to the cell via Iub in downlink. The job to compute and report DL BLER is assigned to UE using the RRC measurement control message.
The measurement can be reported periodically or if a predefined threshold of CRC errors on downlink transport channels is exceeded. From the performance monitoring point of view the disadvantage of event-triggered reporting is that one can no longer follow a certain measurement over a long period of time. Instead only single peak values (exceptions) are reported while all normal values are hidden.
It can be observed that the number of monitored CRC errors increases especially a short time before and short time after radio link additions in soft handover situations. Most errors, which cannot be compensated using macro-diversity filtering, also occur in these time frames, which are usually not longer than 600 ms (maximum), and on average are approximately 250 ms. Therefore, it makes sense to correlate RRC measurement reports as well as messages used to execute the RRC Active Set Update procedure with BLER measurements, because it helps to optimize settings for measurements that trigger radio link additions. Similar investigation could be done for triggering hard handover to different UTRAN frequencies or to different radio access technologies (RATs).
It can be observed that the number of monitored CRC errors increases especially a short time before and short time after radio link additions in soft handover situations. Most errors, which cannot be compensated using macro-diversity filtering, also occur in these time frames, which are usually not longer than 600 ms (maximum), and on average are approximately
250 ms. Therefore, it makes sense to correlate RRC measurement reports as well as messages used to execute the RRC Active Set Update procedure with BLER measurements, because it helps to optimize settings for measurements that trigger radio link additions. Similar investigation could be done for triggering hard handover to different UTRAN frequencies or to different radio access technologies (RATs).
If UL BLER is high for PS services the receiving RLC entity of the SRNC will order a retransmission of corrupted frames and for retransmissions the full bandwidth of a 64 kbps radio bearer is available without limitation. Hence, high UL BLER cannot be the root cause of low transport channel throughput. It must therefore be assumed that throughput is low because there is no more data to be transmitted in the uplink direction, and due to radio transmission conditions UL BLER always rises if any data is transmitted in the uplink direction.
Another possible correlation can be found when analyzing UL BLER and throughput of PS calls. UL BLER is always high if throughput is low and vice versa.
