Bandwidth, Contention, and Profits
Combating forces working against efficient, profitable delivery of
quality data services.
By: Bruce Bahlmann - Contributing Author (your
feedback
is important to us!)
Created:
February 11, 2003
This paper is the product of
Broadband Market Research which is available from Birds-Eye Network Services.
First in a series of monthly technology
columns I've agreed to write for Broadband Properties...
Whether your company offers data service to businesses, residential
consumers, apartment dwellers, or university faculty and staff the effort
required to provide quality data services becomes increasingly complex as
you add customers or users. Contention is one of the most significant forces
working against your efforts to deliver cost effective quality data services
to your customers. Contention forces you to use more and more of your
available bandwidth to control who has access to your shared service.
Contention is created when you have a resource that two or more entities
want a certain amount of but each entity can only obtain a fraction (if any)
of that amount. The phone company works this way – as if everyone in a city
picked up his or her phone to call someone only some portion would get
through. Most data services are also provided in this fashion, only the
level of contention (or over subscription) can become impossible to keep in
check. At what level of contention would an average customer begin to
complain or become unsatisfied with your service, and begin looking at
alternatives? There is no simple answer to this question – however the right
answer is “it depends”. It depends on what market your in, where your
customers are located (geographically), what day it is, what time it is, how
long people typically use the service, whether you charge by the bit or for
flat rate, etc.
 There
is a direct correlation between bandwidth and networking costs. As the
bandwidth of links increase, equipment, personnel, and ISP or Internet
provider costs also increase. Typically the slowest link is the one you need
to regularly pay for – the link that connects your customers to the
Internet. However the size of this slowest link ultimately determines how
many customers your service can support. For example, lets say you provide
services to an MDU and have purchased at T1 from your Internet provider.
Depending on how you divide this T1, will determine your profitability. For
example you could sell 12 apartment residents a symmetric 128kbs service
over that T1 – as a result you have zero contention (very happy customers)
because everyone is allotted a dedicated slice of your T1. But how do you
squeeze more money out of that T1? You do this by increasing contention.
The phone company uses something like 82% contention – meaning that there
are only enough lines available for 18% of its customers to use their phone
at the same time. This 82% is based on many years of well-honed usage
statistics and seems to work reasonably for all but the most extreme
circumstances. Data service is much different from phone service in a
positive way in that it can be shared and in a very negative way in that any
one customer can entirely consume all (or a major potion) of the available
bandwidth.
What kind of contention should you size a data network for? Start with
20% and gradually work your way up to 50%. Beyond 50% you may be playing
with fire as when you pass 50% it requires constant vigilance of your
network engineers in order to maintain reasonable customer satisfaction.
Using 50% contention in our earlier example of the T1, we can calculate the
number of customers such a network could support:
0% contention = bandwidth / service = 1,536 bps
/ 128 bps = 12 customers
50% contention = (1 + %Contention/100) *
(bandwidth / service)
= (1 + 50/100) * (1,536 bps / 128 bps) = 18 customers
Before the telecom meltdown, most large ISPs sized their links for 50%
contention. By using this 50% mark, ISPs were able to leave enough headroom
for the uncorrelated usage spikes that happen daily on shared networks. Now
that the bottom has fallen out of the market, the amount of contention is
being pushed well beyond 50% to the point where it is beginning to impact
service quality and customer satisfaction on a daily basis. If you relate
this to the 88% contention used by phone companies, running a network in
excess of 50% contention is similar to building a phone network that would
max out each day when people returned home from work. Imagine what your
satisfaction for your phone service would be if you would hear the words
“all circuits are busy, try again later” on a regular basis.
Rural area data networks could probably get away with contention as high
as 88 percent - especially if they have a captive audience (the only
alternative is dialup). However, in larger cities and metropolitan areas
where alternative options are numerous your best bet is to provide quality
service at a competitive price now over trying to improve services at the 11th
hour. If you must keep contention in check how do you turn a buck?
Think of this purchased bandwidth as a leaky bucket you want to keep
draining. Your Internet provider supplies a steady stream of water to keep
the bucket full and each leak represents a customer using the service. As
long as the water level is decreasing (but not empty) you are making money.
When everyone is offline, the water coming into the bucket spills over the
top – result, wasted opportunity to capitalize on your bandwidth investment.
There are few providers that seem to exploit the concept of continually
using the bandwidth they have purchased. Your challenge for profitability in
this space is to diversify. Rather than relying on a one dimensional
offering that serves only individual subscribers for all your income, look
at patterns of usage to determine when your bandwidth is idle (no
contention), then execute plans to exploit your bandwidth in other ways at
those idle times. Make your goal to sell your bandwidth 24hrs a day, 7 days
a week to a range of customers whose usage patterns rarely overlap. For
example, when your MDU customers go to work, sell that excess bandwidth to
commercial customers. When they go to bed at night, sell that bandwidth to
companies offering backup services, etc. Keep your options open and your
service offering multi-dimensional. Your quest to diversify your customer
base and exploit your available bandwidth is the way to keep your bucket
leaky and your profits soaring.
Check out these other Birds-Eye.Net papers/products regarding
this article:
Can Birds-Eye.Net help you or your Company?
Receive your Birds-Eye.Net articles and white
papers hot off
the presses by adding our RSS feed to your reader.
|