LTE

Downlink Packet Scheduling

Abhijit Hota

The Node B scheduler (for unicast transmission) dynamically controls which time/frequency
Resources are allocated to a certain user at a given time. Downlink control signalling informs
UE(s) what resources and respective transmission formats have been allocated. The
scheduler can dynamically choose the best multiplexing strategy from the available methods,

In LTE the scheduling is tightly interacting with link adaptation and HARQ. The decision of
which user transmissions to multiplex within a given subframe may, for example, be based
on:
 Minimum and maximum data rate.
 Available power to share among mobiles.
 BER target requirements according to the service.
 Latency requirement, depending on the service.
 Quality of Service parameters and measurements,
 Payloads buffered in the Node-B ready for scheduling,
 Pending retransmissions,
 CQI (Channel Quality Indicator) reports from the UEs.
 UE capabilities.
 UE sleep cycles and measurement gaps/periods.
 System parameters such as bandwidth and interference level/patterns, etc.

Due to signalling constraints, only a given number of mobiles can be scheduled on the same
TTI (e.g. eight).

Interactions between HARQ and Link Adaptation for Packet Scheduling in the below diagram
illustrates the interactions between the different entities involved in packet scheduling (PS),
which are located at the base station (eNodeB) in order to facilitate fast channel-dependent
scheduling in shortening the round trip delay.

The basic time–frequency resource available for data transmission is the physical resource
block (PRB), which consists of a fixed number of adjacent OFDM subcarriers and represents
the minimum scheduling resolution in the frequency domain. The Packet Scheduler is the
controlling entity in the overall scheduling process. It can consult the Link Adaptation (LA)
module to obtain an estimate of the supported data rate for certain users in the cell, for
different allocations of PRBs. Link Adaption may utilize frequency-selective CQI feedback
from the users, as well as Ack/Nacks from past transmissions, to ensure that the estimate of
supported data rate corresponds to a certain BLER target for the first transmissions.

Further, the offset calculation module in the link-adaptation process may be used to stabilize
the BLER performance in the presence of Link Adaption uncertainties. It provides a user-
based adaptive offset on a subframe interval that is applied to the received CQI reports in
order to reduce the impact of biased CQI errors on Link Adaption performance. The aim of the
scheduler is to optimize the cell throughput for the given load condition under the applied
scheduling policy in time and frequency. The HARQ manager provides buffer status
information as well as transmission format of the pending HARQ retransmissions.
Among the possible scheduling policies, we can mention the:

 Fair allocation scheme in which each mobile (in DL or UL) is allocated the same
amount of available PRBs. The number of allocated PRBs per UE changes only when
the number of UE in the cell changes (handover).

 Proportional allocation scheme in which the user bandwidth is adapted to the

changing channel conditions while trying to match the signal-to-noise ratio by means
of power control.

It can also be noted that the interference system in the frequency domain is somehow highly
Dependent on the way the spectrum is used in the respective cells of the network. Frequency
Planning tricks close to the ones well known in FDMA/TDMA systems – like GSM – may be
Valid, including efficient packet scheduling under fractional re-use of the spectrum, e.g. the
Whole spectrum is not used in the whole system in case of lack of traffic, to decrease cell
edge interference.