Suffering From Packet Loss?


 by: Alex Hayward
(This article originally ran in OSP Magazine)

IPTV for telephone operators has become a strategic vehicle to gain
competitive advantage leading to a complex world of Triple Play and
Quadruple Play operators. But all video is not equal. Consumers have
had years of cable and satellite TV service to refine their
expectations, so telcos must be able to satisfy these end-user
expectations while providing a service that isn’t cost-prohibitive.

One factor that impacts both of these is packet loss. IPTV operators
increase infrastructure expenditures to resolve loss issues, and yet
these same losses impact customer quality of experience. Many IPTV
operators say packet loss is the Number One cause of quality of service
complaints, but they continue to struggle with solving their packet
loss problems. What is causing this loss? Where is it occurring? And
what can be done to solve this problem?

It is well-known that The Last Mile of an access network produces
much of the packet loss experienced at the home. In fact, many packet
loss problems occur in the last six feet inside the home, where service
providers have little or no control over the system and its use. Such
seemingly innocent home appliances such as hair dryers and microwave
ovens can impact Quality of Service (QoS) at the last six feet.

Even more frustrating are some geographically unique situations
providers encounter. Imagine being an IPTV service provider in Europe
that has finally documented the cause of a customer’s packet loss
experienced as the electric trolleys that pass by at all hours of the

Because of the unpredictability of packet loss events, even the best
IPTV hardware systems cannot effectively protect against loss to ensure
QoS. No matter how much studying, measuring, and analysis an IPTV
provider performs, these external factors make it impossible for IPTV
providers to solve packet loss with hardware alone. Solving packet loss
under these conditions requires a unique approach, one that includes a
solution that can proactively protect the random problems caused by
external factors before they happen.

Identifying the Source

As previously stated, analog and electromagnetic interference are
typically caused by external factors, including weather and nearby
electric equipment, such as radios, televisions, street trolleys, and
hair dryers. Such external factors can also involve the telco’s own
equipment. Due to the intermittent nature of these causes, they can
exceed the physical layer correction models, resulting in packet loss.
DSL configuration, including noise margin and impulse noise protection,
can be engineered on a per-line basis to mitigate such noise
interference, but it may also result in reductions in available
bandwidth. What’s even more frustrating is the fact that per-line
engineering involves significant operational cost.

Transient changes in bandwidth arise from both configuration
selection as well as equipment tolerances. QoS constraints necessarily
involve selecting bandwidth limits that can produce packet loss if not
managed properly. Excessively jittery or bursty traffic can exceed
input buffers or packet processing capabilities. This, too, can result
in packet loss, or buffer underflow, if the packet is received too late
to be of use. Even if events of this nature are rare, they cause
significant issues for video, whereas they might remain unnoticed for
other services.

Equipment issues, which include such items as bad fiber connections,
spurious messaging, and improper packet handling, are also problematic.
There’s also the potential for providers to use hardware that is
standards-based yet not fully interoperable with other vendor
equipment. This is a challenge with integrating the old with the new.
When field testing for new technologies is limited, dropped packets can
occur due to many things providers have not yet experienced.

Addressing Packet Loss With Hardware or Software?

Since packet loss is a common problem, operators typically scrutinize
their transport networks to eliminate noise and reduce packet loss
wherever possible. Fibers are re-analyzed and/or re-terminated,
questionable connectors are replaced, and problematic equipment and/or
sub-assemblies are swapped out. All of these efforts are part of a
strategy to minimize packet loss in the system.

In spite of this effort, burstiness will occur. This forces some
operators to over-allocate bandwidth to prevent exceeding maximum
channel limits. They may do this to prevent occasional, or even
infrequent, buffer overflows that could lead to packet loss.
Unfortunately, this can compromise DSL reach and some lines can be
disqualified to account for intermittent impulse noise. In short, the
current system is over-engineered so providers may swap efficiency and
reach for a positive user experience. This results in more conservatism
than necessary. While operators must be careful to ensure sufficient
quality throughout the system, excessive conservatism is wasteful.

A software solution is another strategy providers can use to remedy
packet loss. Forward error correction (FEC) technology has been around
for decades and is used today in nearly all electronic transmission and
reception devices such as cell phones, set-top boxes, home gateways,
and most any other device that is involved in moving data on digital

Advanced forms of application-layer forward error correction
(AL-FEC) can be applied to today’s more sophisticated digital media
services, including IPTV. This software system sends repair packets
along with the original data packets to recover IP packets that are
lost in transmission. Since it operates on IP packets, AL-FEC can
provide protection against significantly longer burst periods of
seconds or longer, far longer than the ability of xDSL physical layer

Because it is a proactive solution that mitigates unpredictable
factors that cause packet loss, AL-FEC ensures the highest possible
level of QoS for IPTV providers and their customers while maintaining
any existing network infrastructure and devices.

All IPTV providers know about packet loss, but many struggle to
pinpoint the source of the problem. Because it is such a universal and
complex problem, most IPTV providers simply concede that some packet
loss is going to occur. They aim to minimize the frequency of these
occurrences by beefing up their network infrastructure.

However, packet loss is so unpredictable that even the most robust
networks still experience problems. With ever-increasing customer
expectations for QoS, “almost perfect” simply isn’t good enough. With
IPTV customer service calls at an all-time high, many of which are
packet-loss related, IPTV providers must find a way to keep customers
happy and infrastructure costs to a minimum.

Because of its unique ability to provide passive protection, even in
unpredictable conditions caused by factors outside of the network, such
as weather or household appliances, AL-FEC is a potent strategy to help
solve IPTV QoS problems of both provider and consumer.

So why have there not been more implementations of AL-FEC? Why are
telcos not protecting their networks and making their customers happy
right now? Historically, FEC solutions have been difficult for IPTV
service operators to implement. The FEC codecs are buried into silicon
chips, or the software is extremely low level, making it almost
impossible for the operators to deploy a solution themselves without
thousands of hours of development. This forces the operator to rely on
their equipment manufacturers, who are not error-correction experts.

FEC has never been an elegant solution. In order to protect a
network you have to protect it to the lowest denominator. An operator
who has 2% packet loss in one area, 5% packet loss in another, and a
worst case of 10% packet loss always has to protect to the 10% case;
there is not a lot of flexibility in FEC.

One of the most off-putting things for any operator considering
deploying an FEC solution is that you have to deploy to the whole of
the network or none at all. Since the coding FEC performs changes the
actual source stream delivered to the set-top box (STB), all STBs have
to be upgraded with the necessary software at the same time. With IPTV
deployments reaching 200,000 subscriptions and up, this is no small
upgrade – and is enough to give even the hardiest of IT groups a severe
case of the heeby-jeebies.

Digital Fountain’s ToughStream makes it easier for an IPTV operator
to deploy FEC across their network. Its many features give the operator
the opportunity to implement a roll-out plan that allows them to
address problem areas first and then bring on other areas in a
controlled manner.


About Author

Comments are closed.