In my role, I have the pleasure of speaking with people
in all different lines of business directly or indirectly related to what we do
here at Juniper-everyone from our service provider customers to press and
financial analysts who have their own unique stake in our business. It's the
latter two groups that have been asking questions about whether service
providers can or will "run their networks hotter" to preserve capital expenses
in these trying economic times. I'm not sure who coined the term, but it seems
to be gaining traction as our esteemed competitor referenced it on their last quarterly earnings conference
call.
At first glance, the idea makes sense: service providers
build their networks with some capacity "headroom" so that spikes in traffic can
be accommodated seamlessly. Instead of investing in more capacity, why not just
reduce the amount of extra capacity headroom built in and run networks closer to
their maximum capacity? Obviously this would have an impact on service quality
in times of congestion, but everyone has to make sacrifices in these times.
Customers will deal with the occasional loss of service quality,
right?
Not exactly. Let's look at this a little closer. Most
service providers serve one or more of three basic markets: wholesale,
enterprises and consumers. Many of the large global providers serve all three;
smaller players will often focus on one or two segments. Each of these markets
in turn uses services that can be broken down according to the value they
provide to the end user and whether that value is based primarily on cost or
quality. Cost based services are typically "best effort," and examples of these
include basic Internet access (enterprises and consumers) and pure bandwidth
services (wholesale).
Value-based services on the other hand require an
assured user experience. Sometimes this experience is guaranteed contractually
via SLAs (managed services, virtualized services); other times the nature of the
service, consumer fickleness and the pure competitiveness in the market serve as
a virtual SLA of sorts. For example,
residential consumers of IPTV and VOD services will quickly switch back to cable
if they perceive quality issues.
Will service providers run their cost-based services
hotter? Probably some will. They can't keep it up for long given the continued
growth of network traffic, but realistically it will
happen.
Value-based services that depend on an assured
experience on the other hand cannot "run hotter." Service providers simply can't
afford it - meaning they can't afford breaking an SLA in the enterprise and wholesale markets, and they
can't afford losing residential consumers to competitors with higher quality
services.
Interestingly-and this is where I can't help but get a
Juniper pitch in-if you look at the types of networks that support each of these
service types, it's the value-based services that require the high-performance
networking infrastructure Juniper delivers. Cost-based services require another
scale when run on IP-which is also Juniper's DNA! So when you hear people talk
about running networks hot, it would be good to consider where this type of
cost-saving maneuver would happen-in legacy networks, supporting declining
businesses, or in the high-performance IP networks that support the services of
most value (which today only makes up less than 10% of all capital
expenditures)? I won't answer that question here, but you might consider AT&T's announcement on March 10, 2009 an interesting indicator.
Message Edited by ac on 04-24-2009 11:48 AM
Rami Rahim (RamiRahim)
on:
The Next Chapter in the Converged Supercore
