|
|
article continued from
page 1
In addition to dumping huge amounts
of bandwidth to each user's doorstep,
a provider may first be tempted to overbuild
and over-engineer the access network
in order to provide this guarantee.
But, Boland explains, "from the
routing perspective, hierarchical queuing
mechanisms allow service providers to
deliver voice, video and data services
without having to overbuild the access
network."
Video traffic must, by design, be sent
with very low loss and low latency,
and with guaranteed bandwidth. Boland
notes, "To a certain extent, there
isn't much QoS per se to apply when
you look at the problem in isolation.
Either the bandwidth is there or not,”
assuming that the traffic is properly
shaped by the video source and reshaped
at the last hop. But, when you add VoIP
and video, strict priority must be enforced.
However, the most overlooked element
here is high speed Internet (HSI). Delivering
superior service for one
leg of the triple play must not
come at the expense of the others. "Most
people are very dismissive, due to the
'best effort' nature of the service.
And several QoS architectures for broadband
are based on basic per-traffic-class
and per-port scheduling, enforcing prioritization
of VoIP and video, but adversely impacting
HSI." When bandwidth-hungry file
sharing applications are factored in,
there's naturally going to be disruption
without advanced QoS features like per-subscriber
fairness and weighted queuing.
Boland takes note of the fact that
most consumer Internet services have
no formal SLA, despite the fact that
this service is the mainstay for existing
broadband service providers. If there
is degraded service, the end users will
notice it, and the providers will have
more tech support calls--and more customer
churn. "So how can carriers afford
to deploy new services if this would
disrupt their bread and butter activity
for both consumers and small business
offices? And this is exactly what will
happen with a mix of video, VoIP and
HSI if no advanced QoS is used."
|
|
"The
interactivity of IP naturally lends
itself to services that are not
possible on traditional television." |
|
By design, however, just having
a very large amount of a bandwidth
delivered to the subscriber's
door is just as important as QoS.
Boland notes that managing the
bandwidth is critical. “The
routers in the access network
need to provide per-subscriber
and per-service queuing. Per subscriber/service
queuing helps to prioritize traffic
within a subscriber VLAN which
enables each service to be optimized
for delay, jitter and packet loss."
Conclusion
The growing demand for IPTV as
part of a triple-play bundle is
already placing tremendous demands
on the industry, and requiring
providers to lay out big money
for infrastructure changes. The
competitive front will be vibrant,
but Boland says, "I do not
think that the number of players
will diminish. In a few years,
I expect to see additional service
providers come to market with
a triple-play offering based on
wireless access. That will create
an interesting competitive battle
between wire and fiber companies
versus wireless."
Ultimately, though, customers
don't really care how their service
gets delivered, only that it is
good quality, and they get lots
of television channels with cool
new features. IPTV is going to
fill those consumer demands for
a number of reasons, including
the availability of pricing discounts
available for bundled service
packages, consolidated billing,
and more creative channel bundles
that would allow the consumer
to individually select, and pay
for, only the channels they want.
In addition, IPTV will allow for
more interesting features, such
as onscreen caller ID, email notification,
multiple picture-in-picture, menu
screens, and unlimited other options
not available with traditional
cable television.
|
|
|
|
|
