Comparison between service steering and random combing between different carriers
1. Comparison between Service Steering and
Random Combing between different carriers
1.Overivew:
Multi carrier solotion is one of the most common solotions
that all network might be using now in order to enhance
their capacity, RF conditions, and user perception after all
In order to use multi carrier solution we have to choose
one of two access techniques either Random combing or
service steering both are commonly used in many
networks that already have multi carrier condition, but
which one is better? And how to choose between them in
the next points we will go deeply in order to understand
which one we can use
2.What is service steering and how it work?
Service steering is the technique in which we define
special service to different carriers for example we define
F1 (first carrier) for voice and R99 services and define F2
(second carrier) for the HSPA services, the benefit of this
one is to force all the data traffic on F2 that will make the
fist carrier more free with much better radio conditions so
it will enhance the accessibility of voice services as there
will be much improvement in f1 radio conditions
2. Also for the data services, all users can use all the power
allowed on F2 which will enhance the throughput ( to
certain point)
3.What is the bad effect of service steering?
Service steering will work perfectly fine as long as the
Data traffic is not high on the network, when the data
traffic increase on the network you can easily see that
there will be congestions on the data layer even that you
will have many free resources that can be used, which is
not our target for sure to ensure the maximum usage of the
current available resources. Other problems will be
demonstrated later
Also the service steering situation will lead to lower RAB
Assignment Success rate, mainly due to the DRD usage
also the time of call establishment will be much higher
compared with random combing
4.Now what is the Random combing?
Random combing is the opposite of service steering in this
part we will not force the traffic on a certain carrier even
more we will make it like free competition between
different carriers, so all services will be available on all
carriers in this case we will have the maximum utilization
of the nodeB resources making sure that the users can get
the maximum benefit available under condition that we
can balance the traffic between these different carriers
3. 5.Does Random combing have any problems?
Sure it have, random combing to give the required effect
the network should have continuous coverage in both
carriers, since we are going to enable free combination
between different carriers we also should have enough
license to make both carriers have the same services
exactly no advantages can be giving to any one of them,
any advantage will lead directly to non balanced situation
of traffic
6.So which one we should use?
If the network doesn’t have high data traffic and doesn’t
have enough second carrier coverage so we don’t prefer
the usage of the random camping service steering will be
much better
But if we have high data traffic and continuous second
carrier coverage, and want to enhance the RAB Setup
success rate, enhance the access delay time, enhance the
HSPA throughputs we have to use Random camping,
In the coming parts we will show with technical proof why
we may have:
• Long delay for RAB access
• Bad utilization of the resources
4. • Many RAB failures leading to low RAB Success
rate
7.Service steering bad effects technical
analysis:
Long connection setup time:
By comparing the RAB setup time of DRD with
None-DRD from the test data, we found the average
setup time with DRD is 200ms longer than none-
DRD.
Below figures from the layer three messages and
signalling trace to show the time delay exactly
RAB setup with DRD
5. Rab setup without DRD
This can confirm with no doubt that the time
required to initiate a call using DRD is 200 ms longer
than the time required for non DRD case under the
same conditions
lower RAB setup success rate
According to the process of DRD, every time when
the DRD happens, the NODEB and UE will start the
radio link synchronization again. The one more time
RL synchronization will increase the chance of RAB
setup failures.
Basing on service steering strategy, (F2 carry HSPA
traffic), the load of F1 and F2 is different, it will
increase the chance of DRD failures
6. Unbalanced resource utilization
As we mentioned before, in case of high traffic the
usage of the resources will be much different
between both carriers, clearly we will have more
utilization on the second carrier than the first carrier
7. specially in case of power below trends to show the
different in TCP usage between service steering two
carrier network
Also RTWP is showing big difference in the point of
UL interference, since F2 s having more traffic so
more UL interference will be found on F2
8. 8.Now how we can implement the Random
camping in order to avoid the above
drawbacks?
In order to enable Random camping below parameters
should be tuned as mentioned
Camping Policy
F1,F2:Random Camping; Command reference as the
following:
ADD UINTERFREQNCELL:CELLID=F1,
NCELLID=F2, IDLEQOFFSET2SN=0,
BLINDHOFLAG=TRUE, SIB11IND=TRUE;
ADD UINTERFREQNCELL:CELLID=F2,
NCELLID=F1, IDLEQOFFSET2SN=0,
BLINDHOFLAG=TRUE, SIB11IND=TRUE;
//BLINDHOFLAG: is on in order to allow LDR to be uses
on the same sector cells only
// IDLEQOFFSET2SN is set to zero in order to allow the
free idle mood camping which is valid only when SIB11
Indicator is set as TRUE.
9. Access Policy
• R99,HSPA service random camping on F1 or F2, and
launch access in the camping carrier. There is no DRD in
HSPA service between two carriers.
Detailed parameters is as following:
Disable ServiceDiffDrdSwitch of F1/F2 by the command
MOD UCELLDRD;
Disable LdbDRDSwitchDCH of F1/F2 by the command
MOD UCELLDRD;
10. Disable LdbDRDSwitchHSDPA of F1/F2 by the
command MOD UCELLDRD
Connection Policy
• Adopt LDR: Detailed parameters is as following:
9.What is the improvement we can see after
implementation?
In our case since we had big improvement in case of real
field we had significant enhancement which can be proved
by the below KPI trends
Below to show that we managed to balance the same
HSDPA user number on both carriers
11. Average throughput on F1 increased with almost
same throughput on F2 so this mean that overall
throughput on the Area totally increased as well
Total HSDPA data volume also didn’t change much,
but the traffic is passed from F2 to F1 leading to
more data balance and resource utilization
12. PS RAB Setup success rate improved on F2 and on
the area overall as in below trend
13. The RRC Call drop rate improved a lot as we
decreased the DRD attempts
PS Call drop rate also improved a lot
14. 10. Conclusion and summary:
In the parts we covered we can get better
understanding in the below main points:
What is service steering and what is random
camping?
When we may prefer one of them on the other?
What is the main drawbacks of service steering and
how we proved them?
How to implement the random camping?
What enhancement you may expect when you use
random camping