Page 201 - From GMS to LTE

Universal Mobile Telecommunications System (UMTS) and High-Speed Packet Access (HSPA) 187

despite the introduction of HSDPA in 3GPP Release 5. The solution towards satisfying
the increasing demand in the uplink direction is referred to as Enhanced Uplink (EUL)
in 3GPP and is also known as HSUPA. HSUPA increases theoretical uplink user data-
rates to up to 5.76 Mbit/s in 3GPP Release 6 and 11.5 Mbit/s in 3GPP Release 7. Further
enhancements have been made in subsequent 3GPP releases, but it is unlikely that these
enhancements will be implemented in practice due to the emergence of LTE. When
taking into account realistic radio conditions, the number of simultaneous users, mobile
device capabilities, etc., user speeds of 1–4 Mbit/s are reached in practice today.
For the network, HSUPA has a number of benefits as well. For HSDPA, an uplink
DCH is required for all mobile devices that receive data via the HS‐DSCHs for TCP
acknowledgements and other user data. This is problematic for bursty applications as a
DCH in the uplink direction wastes uplink resources of the cell despite reduction of the
mobile device power output during periods when no user data is sent. Nevertheless,
HSUPA continues to use the dedicated concept of UMTS for the uplink by introducing
an Enhanced Dedicated Channel (E‐DCH) functionality for the uplink only. However,
the E‐DCH concept includes a number of enhancements to decrease the impact of
bursty applications on the DCH concept. To have both high‐speed uplink and downlink
performance using an E‐DCH introduced with HSUPA only makes sense when com-
bined with the HS‐DSCHs introduced with HSDPA.
While a Release 99 DCH ensures a constant bandwidth and delay time for data pack-
ets, with all the advantages and disadvantages discussed in previous chapters, the
E‐DCH trades in this concept for higher datarates. Thus, while remaining a dedicated
channel, an E‐DCH no longer necessarily guarantees a certain bandwidth to a user in
the uplink direction. For many applications, this is quite acceptable and allows an
increase in the number of simultaneous users who can share the uplink resources of a
cell. This is because the network can control the uplink noise in a much more efficient
way by dynamically adjusting the uplink bandwidth on a per‐subscriber basis in a cell to
react to changing radio conditions and traffic load. This reduces the overall cost of the
network by requiring less base stations for the same number of users, which, in turn,
can result in cheaper subscription costs.
The E‐DCH concept also ensures full mobility for subscribers. However, the radio
algorithms are clearly optimized to ensure the highest throughput for low‐speed or
stationary use.
The main purpose of the E‐DCH concept is to support streaming (e.g. mobile TV),
and interactive (e.g. web browsing) and background services (e.g. FTP). To ensure good
performance for real‐time applications like IMS video conferencing, the E‐DCH
enhancements also contain optional mechanisms to ensure a minimal bandwidth to a
user. As these methods are optional and packet‐based real‐time applications are not yet
widely used, current implementations focus on the basic E‐DCH concept. Despite this,
Voice over IP and Video over IP services can still be used over a non‐optimized E‐DCH
without any problem as long as the cell’s bandwidth is sufficient to satisfy the demand
of all users currently transmitting data in a cell regardless of the application.
As the uplink bandwidth increases and fast retransmissions are introduced, the
E‐DCH approach also further reduces the round‐trip delay times for applications like
web surfing and interactive gaming to around 60 milliseconds.
Finally, it is important to note that the E‐DCH concept is backward compatible. Thus,
a cell can support Release 99 mobile devices that were only designed for DCHs, HSDPA