Channel Element Hakkında
Channel Element Hakkında
Channel Element Hakkında
In a RAN, CE resources are managed by both the RNC and NodeB. The NodeB
reports its CE capacity to the RNC. The RNC determines whether to admit a new
service based on the number of CEs that need to be consumed and controls CE
resources during CE congestion. This ensures the proper use of CE resources. The
NodeB dynamically manages CE resources and rapidly adjusts the number of CEs
that can be consumed based on the actual service rate. This increases CE resource
usage.
DL used capacity:
o Sample 1:
100 AMR + 20 x 384 kbits + basic consumption (16 + 32) = 100 + (20 x 16) +
48 = 468 CE
o Sample 2:
o Sample 3:
UL used capacity:
o Sample 1:
o Sample 2:
o Sample 3:
M5001C0 MAX_AVAIL_CE = 92
M5001C1 MIN_AVAIL_CE = 92
M5001C2 AVE_AVAIL_CE = 92
Example 2:
CE Availability:
The change of licensed capacity is made in the middle of the measurement period; 60 CEs are
added.
Counter values:
Functional overview
For Flexi Multiradio WCDMA BTS Rel.2 and upcoming HW versions a new baseband
concept has been introduced.
The new architecture in the baseband removes the traditional channel element (CE)
based capacity allocation for HSPA. Dimensioning of HSPA is based on different HSPA
processing sets (PS) instead. The new software design aims to maximize HSPA
capacity and provide a future proof basis for the coming new HSPA+ features.
The new software design aims to maximise HSPA capacity and
provide a future proof basis for new HSPA+ feature design. Different tasks for baseband
processing have been allocated so that underlying support for HSPA+ peak rates as well
as other 3GPP HSPA+ features is as efficient as possible. This is visible in radically higher peak rate
capabilities and as completely new dimensioning rules for HSPA.
Channel elements are no longer used for HSPA at all, and instead dimensioning is
based on different HSPA processing sets. The new architecture allows up to 1 Gbps
total throughput per BTS with two system modules. This value is describing the
baseband capabilities, and is not possible to achieve over the air interface yet at least
with 3GPP Rel11 features and 12 cells.
HSPA capacity increase has been designed so that AMR capacity, still basing on
channel elements, is not decreased. To the contrary there is some additional channel
element capacity now available for AMR calls. For details of the new baseband software
arechitecture capabilities for both HSPA as well as for AMR calls.
The HSPA processing sets are capacity units, defined by total BTS throughput and
number of users in BTS in the following way:
• for HSDPA
– HSDPA BTS Processing Set 1: supports maximum number of 32 users and
overall BTS throughput up to 7.2Mbps
– HSDPA BTS Processing Set 2: supports maximum number of 72 users and
overall BTS throughput up to 21Mbps
– HSDPA BTS Processing Set 3: supports maximum number of 72 users and
overall BTS throughput up to 84Mbps
• for HSUPA
– HSUPA BTS Processing Set: supports maximum number of 24 users and
overall BTS throughput up to 5.8Mbps
The table above shows the maximum throughput capacity of one HSDPA or HSUPA
scheduler. Whether it is possible to reach this, depends on the number of availablesubunits in the
related system module. The subunit capacity is mapped to throughput capacity (in Mbit/s) in terms of
steps (numbers not given in the table).
Possible user capacity
The total used capacity of the BTS depends on the number of HSPA schedulers. In addition,
there are separate limits in the HSUPA scheduler for an LCG and cell levels.
Baseband capacity monitoring cases description
Note that there are differences in the monitoring between RU30 EP1 and RU30 EP2. In
EP1, only LCG-level M5006 counters are used for baseband monitoring, while in EP2
M5008 counters are also available for BTS level monitoring.
The available baseband capacity is based on a BTS-level or a HSUPA scheduler-level
license, that is, the licenses are based on R99 CEs and HSxPA processing sets and not
the subunits located in the system modules.
Therefore the monitoring is (primarily) based on following the license utilization rates
and not the consumption of subunits/baseband/DSP HW resources. If needed, the DSP
HW resource, that is, the subunit utilization level can be seen. This requires then additional
mapping between the monitored license related utilization rates and what it means
in terms of subunit consumption. This can be done following the mapping instructions
given above in the introduction section Possible system module capacity, Possible
throughput capacity, and Possible user capacity.
The BTS HW capacity can be further grouped (pooled), that is, the BTS can consist of
one to four pools (local cell groups, LCGs). Thus, there are two different cases for
capacity monitoring:
• One LCG, that is, baseband pooling is not used
License monitoring is simple, as the BTS licensed capacity = the LCG licensed
capacity.
2 - 4 LCGs, that is, baseband pooling is used
License monitoring is more complicated, as the licensed baseband capacity is distributed
among the LCGs according to the commissioning setup, common LCG pool
and/or common schedulers.
– The assigned part of the R99 CE license is distributed to the LCGs according to
the commission setup. Unassigned Rel99 CE licenses are in common LCG pool,
and licenses from that pool can be utilized by any LCG on need basis. This
common share is controlled on a BTS level.
– The licensed HSDPA throughput is shared between LCGs in case of commissioning.
The maximum value depends on the available HSDPA processing sets.
If commissioning was not done, the throughput is shared according to the default
allocation rule
– HSDPA user shares of used licenses in case of commissioning is shared
between LCGs up to maximum allocation of 100% licensed capacity. If not, the
remaining capacity is put to the shared LCG pool. If commissioning was not performed
for these licenses, then each LCG gets equal share and there is no
common pool of licensed users.
– For HSUPA, there is no common part - commissioning has to be performed so
that values sum up to 100%. If commissioning was not performed, then each
LCG gets equal share.
Monitoring sub case setup:
RU30 EP1: The monitoring is performed on an LCG level, that is, separately for each
baseband pool or for the whole BTS.
RU30 EP2: The monitoring is performed on either LCG or BTS level.
Monitoring sub case analyze:
• LCG level monitoring: If LCG resource consumption is high, then the problem
is either based on the commissioned LCG resource shortage, or full BTS license
shortage, or baseband shortage (when all baseband is licensed).
• BTS level monitoring: If BTS resource consumption is high, then the problem
is either based on the BTS license shortage or baseband shortage (when all
baseband is licensed).
Monitoring sub case notes:
RU30 EP1: The monitoring is executed on LCG level only. This is sufficient as the
provided counters show directly the utilized capacity in relation to the available
capacity in each LCG (that are obtained based on commissioned and/or common
BTS/HSUPA scheduler-level licenses).
RU30 EP2: The monitoring is executed on either LCG or BTS level with the corresponding
counters.
HSUPA consumes both UL and DL baseband processing within the WBTS, so it's necessary to ensure
that there is sufficient baseband processing resource prior to enabling HSUPA. The HSUPA baseband
processing requirements are symmetric across the UL and DL directions. Additionally, DPCCH for
HSPA needs one extra CE.
The minimum static allocation of 8 CE (for both UL and DL) is reserved as soon as HSUPA is
enabled at a WBTS
With the exception of the first 8 CE’s, the CE allocation for HSUPA is dynamic. Depending upon
both the number of HSUPA connections and load generated by the DCH traffic.
In the CE allocation, DCH has a higher priority than HSUPA => DCH capacity requests are able to
pre-empt the set of CE allocated to HSUPA (with the exception of the minimum static
allocation of 8 CE).
CE usage can cause bottleneck for HSDPA users (UL return channel being rejected) or in terms of
limiting the DL throughput (i.e. UL not supporting even the TCP akcs). 64kbps UL return channel (4
CE) seems to be enough up to ~2mbps and 16kbps is enough up to ~600kbps. Also it has bee noticed
that the UL return channel allocation share is:
• ~30% UL 16kbps (1 CE needed)
• ~50% UL 64kbps (4 CE)
• ~10% UL 128kbps (4 CE)
• ~10% UL 384kbps (16 CE)