SYSTEMS DESIGN OPTOELECTRONICS
few months, Chinese manufacturer
Huawei has been promoting the idea
of F5G for some time, and has put
forward a proposal based around a
mixture of bre-optic xed links and
wireless-access based on WiFi 6
based on the IEEE 802.11ax protocol.
Sue Rudd, director of networks
and service platforms at Strategy
Analytics, says: “ETSI is creating
F5G in order to coordinate inputs to
multiple of cial standards bodies.”
Those bodies include the 3GPP
group responsible for the 5G
releases, the ITU and IEEE, which
have de ned a number of the existing
bre and WiFi protocols and the
Broadband Forum, which recently
issued in the TR-470 standard its
own architectural guidelines for a
5G infrastructure that handles traf c
from both wireless and wired access
networks.
The key to the deployment of
the wired side, for the F5G group
and others, is focused on a passive
optical networks (PON) architecture
in which signals passed through the
underground bre are handled using
optical combiners and splitters rather
than active electro-optical switches
and routers. Deployment based on
existing standards such as 1G PON
has been patchy with penetration in
many larger economies is only on a
small scale.
However, one reason for expecting
a more extensive deployment of a
PON infrastructure around the world
lies in the needs of 5G itself.
Equipment vendors such as Calix
Communications argue organisations
who put the bre in the ground and
offer access will acquire a signi cant
user base not just from internetservice
and TV providers but the 5G
operators who need a way to connect
millions of small-cell basestations.
Splitters peel off optical signals from
a trunk to direct bre connection
to groups of homes and picobasestations.
That original 1G PON standard,
now well over a decade old, has
spawned numerous variants that have
gone through the ITU process aimed
at handling aggregate data rates
of 10Gbit/s and higher. First came
XG-PON, with 10Gbit/s downstream
at the 1577nm wavelength and a
2.5Gbit/s upstream on 1270nm.
XGS-PON boosts that upstream link to
the full 10Gbit/s. In parallel, an IEEE
working group created 10G EPON,
which is super cially similar to 10G
XGS-PON except that it supports either
1Gbit/s or 10Gbit/s upstream and is
intended to be backward compatible
with 1G PON for an easier upgrade by
having 1Gbit/s and 10Gbit/s share
the same channels using time-division
multiplexing. The downside is a loss
of aggregate bandwidth if the bre is
used for both.
The ITU has since added NGPON2,
which uses wavelength division
multiplexing to offer handle multiple
downstream and upstream channels
in parallel, with support for more
than four users through time-division
multiplexing. Google extended the
work in a different way for its socalled
Super-PON, using a concept
rst proposed in the mid-1990s. It
is aimed at connections as long as
50km with the help of bre ampli ers
in the delivery path. This option is
primarily for green eld deployments
as it does not coexist with older PON
equipment in the way that most of
the other WDM options do. Huawei
appears to favour the ITU options
such as XG and XGS-PON in its
equipment though the F5G group itself
has not as yet speci ed a favoured
option. Among the few chipset
vendors with 10G offerings, Broadcom
is hedging its bets with support for
both the IEEE and ITU options.
More work to do
The work does not stop at 10Gbit/s.
Huawei and others are working on
50Gbit/s extensions including a
possible version that uses multilevel
signalling over a single wavelength
based on the same kind of technology
that goes into high-speed ethernet.
Cost may prove to be a stumbling
block for the higher-speed variants as,
to handle coexistence on a passive
network, the laser modules may need
to be tuneable. Vendors are looking to
silicon integration to help bring costs
down though this will depend heavily
on operators adopting the F5G or
something similar.
At launch ETSI had a number of
operators backing the F5G concept
including several based in China with
the remainder being mainly operators
in Europe and the near East such
as Altice Portugal, Bouyges Telecom
and Turk Telekom. But there were
no US operators. For F5G and the
bre-generations concept itself to
become as in uential as the mobile
generations, the work will need
far more acceptance among larger
players. The cost savings promised
by moving to a converged network
may be enough to bring other
players onboard. But there is still the
chance that the F5G group may see
competition from other groups who
favour a different crop of protocols.
Below: Huawei
has put forward
a proposal based
around a mixture of
fi bre-optic fi xed links
and wireless-access
www.newelectronics.co.uk 24 March 2020 31
/www.newelectronics.co.uk