17 January 2013
| Kavit Majithia
Epsilon's decision to reintroduce a bandwidth trading platform after widespread failures in the market over a decade ago has again brought about the age-old question: is bandwidth a tradable commodity?
Over a decade ago, the former
founder and board member at US exchange company Arbinet, Alex
Mashinsky, met with a group of Enron executives.
During the meeting, he looked the group of hard-nosed traders
in the eyes and told them, that while their idea to create a
bandwidth exchange platform was interesting, "it was not going
to be about data for the next 10 years, it was going to be
The board ignored Mashinsky and told him in no uncertain terms
that "there is an insurmountable amount of money in bandwidth
trading". The rest, as they say, is history.
Enron invested aggressively in bandwidth trading, and attempted
to consolidate a market that had little belief their model
would work. Enron’s hard-line approach quickly
left its investors out of pocket, and the company eventually
Although his prediction proved correct, Mashinsky regrets
voicing his concern to the Enron board. In the very same
meeting he rejected a $100 million offer from Enron to acquire
Arbinet. "Sometimes being right does not always bring you
fruition," he says. "But that’s life."
Fast forward to 2012, and there is
a CEO in town that believes the time is now right to
re-establish the bandwidth exchange platform.
Andreas Hipp, CEO at Epsilon, has taken it upon himself to
attempt to trade bandwidth again, and it has generated a host
of opinions from old and new players in the market.
Although Epsilon Capacity Exchange (eCX) works a little
differently to the platforms that were developed in the late
nineties just before the famous telecoms bust of 2001, the same
fundamentals are there.
eCX offers both buyers and sellers an online trading platform
for SDH and Ethernet bandwidth services, incorporating a
real-time inventory of routes, a pricing and trading engine and
a capacity booking and reservation function.
For Epsilon, it builds on a profitable business that is has
worked on for a number of years, incorporating its global
network exchange spanning across 60 PoPs worldwide. Hipp claims
the idea and the development of the platform is really driven
by the need to speed up a process that is dated, and can now be
facilitated by a range of technological developments.
"We have built the infrastructure that Enron did not have 12
years ago," says Hipp, who spent six months at Enron in 2001.
"We identified that we had a lot of customers with idle
capacity sitting on their networks which could have been
because of big cable investments."
In some ways Hipp’s reasoning to take the leap
into bandwidth trading makes perfect sense. Over the past few
years there has been substantial investment in subsea cable
systems, which has left carriers and service providers alike
with capacity they are not fully utilising.
"They approached us and asked if we needed any of that
capacity, which we then tried to offload for the company, in
essence becoming the middle man between carrier A and carrier B
and as the transactions increased the solution was clear
– it’s time to trade this capacity," says
Epsilon claims it had all the pre-requisites in place, and is
only really tasked with establishing the front-end optimisation
piece and operating the connectivity platform. Hipp does not
want to get involved with price, contract terms or quality of
If the platform or the idea fails, as it has done in the past,
"our business won’t be affected", Hipp says
defiantly. "There is an SLA to fall back on and
it’s still the seller’s risk and
Richard Elliott, now managing director at satellite
communications provider Apollo, was one of the first pioneers
of capacity trading. He co-founded Band-X back in 1997, and
claims Band-X’s exchange concept was one of the
first to hit the market, leading to numerous imitators
including Arbinet, Enron and Dynergy.
Its developed platform originated in voice, and the company
traded IP transit as a wholesale interconnection on the
internet with a fixed and visible margin that also traded
point-to-point circuits in a similar way to Epsilon. Elliott
"takes his hat off to Epsilon", and heralds the web-based front
end developed by the company as "better than it ever has
industry’s experience with the highly
controversial Enron continues to leave a sour taste.
The fact is, however, that even Enron’s rivals
failed in long-standing attempts to secure a highly scalable
tradable platform for capacity. Jim Poole, general manager of
global networks and mobility at data centre provider Equinix,
says voice thrived and data failed in the last decade because
of a set of well-established routes, which simply
weren’t in place when the Enron’s of
the world attempted to build a trading platform.
"Voice had a set of interconnect points and routes, and you had
people in place that ran tandems or resellers who sat behind
networks and it never flipped into the idea that voice was ever
sold as a commodity," he says. "Certainly, bandwidth has moved
closer to voice and it is slowly beginning to reach towards a
more fluid market, but I certainly do not see commercial wide
adoption as yet."
Elliott’s Band-X IP trading platform was bought
out by Mashinsky’s Arbinet during a time when the
stock market was highly reactionary. Any murmurings in the
market that suggested activity in bandwidth trading sent stocks
Mashinsky recalls that when Enron first announced its strategy
for bandwidth trading, its stock doubled in value from $30 to
$60. The story goes a long way to explaining why the sector
collapsed so rapidly. "All they said was, we are interested in
bandwidth and we are going to start playing in it. It was
pretty much the same case for any other company."
Mashinsky reveals to Capacity that Enron was not the only
company that looked to Arbinet to beef up its operation.
"Because of our exchange strategy, at one point before the
telecoms bust we were offered $1.6 billion to go public but we
never completed the transaction. Global Crossing also tried to
acquire us, but its valuation was too low."
And the potential for bandwidth trading to grow and develop was
fuelled by the bigger players’ interests in the
service, says Elliott. Before the collapse of 2001, the
industry saw the emergence of a generation of very well-funded
This helped create a pure wholesale network with the single
purpose of selling on wholesale services. Such market dynamics
played nicely into the hands of bandwidth trading because it
allowed start-ups to cheaply operate in the market and acquire
global network connectivity.
"Wholesale players were the main drivers to demand on the
network side of bandwidth trading but a significant proportion
of these went bust within the space of six months," says
Back then, growth came from the plethora of carriers buying
capacity which was then sold on in the wholesale sphere. This
inadvertently created the same fundamentals that a bandwidth
trading platform is based on.
Trading thrived, says Elliott, but when the collapse hit it was
almost comparable to what the banking sector experienced in
2008. "Almost $4 trillion was obliterated from the
telecommunications and technology sector and almost all
bandwidth trading organisations collapsed. Subsequently, we
haven’t seen a bandwidth trading sector emerge
ever since, until now."
So why now? Hipp and Epsilon clearly believe its online tool
will pave the way for capacity trading services on a global
"Enron claimed that they could offer a differentiated service,
but really they were just powering their own network under the
illusion of trading," says Hipp. "It was just too early to get
the concept going. We believe we have now reached the right
pricing levels and margins, with more people understanding what
the service is. It’s a traditional network service
that provides a point-to-point link."
With optimism naturally comes pessimism. The market still
appears at odds with regards to how responsive operators will
be to such a platform.
Tom Madaras is now senior sales engineer for CDN services at
Level 3, but was formerly the man in charge of global capacity
at Enron and worked with a range of traders to establish the
viability of an exchange platform. Madaras claims
Enron’s failure was down to the fact that
"carriers had established a good old boys network".
Controversially, he says: "Carriers kept their assets close to
their hearts and they didn’t want to give up any
sort of control. They had built and network and decided they
did not want to play with Enron, and dismissed it as a gimmick.
Ultimately, we at Enron had a great concept, a great product
but they couldn’t see the benefit of creating
liquidity in the market and the rest is history."
And despite Hipp downplaying the amount of risk it has taken,
Poole believes Epsilon is taking on something in an industry
"that is a slow beast to make changes". He says: "No-one seems
to be going out on the limb to say they will throw all their
excess capacity into this. I see itinerant intermittent
interest, not some saying, this is the future and I have to
jump on this."
The main issue with trading
bandwidth and wholesale services is "how do you make money out
of something that continuously drops in price?" says
When it came to trading minutes, small incremental improvements
on a voice trading platform only saved the carriers a few
pennies - not big returns.
According to Mashinsky, the difference is that bandwidth is a
growing business. "It will continue to grow for years and it is
clear that traffic volume will scale up because the economies
of the internet are going to change. In different ways, you are
seeing an exchange model emerging with companies like Level 3
and Netflix paying peering premiums to Comcast for the delivery
of video with Comcast dictating the pricing."
It is equally important to remember the market’s
relative infancy, according to Elliott. "Utility trading has
only really emerged in the past 25 years. The centralised
market for oil, gas and electricity only emerged in the late
seventies so there certainly is a viable market for traded
bandwidth. I spent a good number of dollars trying to perfect
it with Band-X."
Drawing on his past experiences with Eron, Madaras believes
there are still potential limitations with the bandwidth
exchange model in today’s marketplace. Namely,
that the number of companies which control the physical assets
"Bandwidth still isn’t at the point that it can be
physically traded," he says. "For all the price compression
that has occurred in the market, the number of companies that
control it on any given route is small. In some ways it is
utilised as a commodity in the metro markets. On long-haul
routes, which are already established in developed markets like
the Americas and western Europe, the paradigm is the same.
People who control those routes are still trying to manage
prices as closely as they can."
This differentiates bandwidth from the real lifeblood of
tradable products, like commodities or stock, where there are
an insurmountable number of players investing, trading and
interacting. "It’s still sitting in very few hands
and if you want to trade in real volumes you need liquidity,"
"For this platform to really work you need big players like
AOL, Amazon, Netflix and Google to enable this capability for
customers. If Amazon cloud or Amazon video services started to
buy capacity through this exchange this will blossom.
Unfortunately, sellers do not want liquidity or transparency
because that will mean a price drop, and buyers will have the
ability to switch to the cheapest sellers efficiently."
In some ways, the entire premise of an online trading exchange
platform actually goes against the fundamentals of wholesale
telecommunications, which prides itself around the idea of
networking and holding good client relationships.
"Friends do business with friends, but that’s the
problem," says Madaras. "If we continue down that road it will
never allow for liquidity and yes, people do business with
people they know but we need to move away from that
The time has come for an international body to set regulations
on how business is done, develop a costing and interconnect
model to create a standard playing field for everyone."
Aware of the
industry’s perception of bandwidth trading
platforms, Hipp urges his fellow colleagues to embrace change.
"People have even spoken to me about the impact this sort of
thing will have on jobs, but it’s not something we
should fear," he says. "The earlier companies address where the
market is going and prepare for lower margins, the more time we
have to prepare for it."
For Mashinsky, the memory of a failure to establish bandwidth
trading has long lingered in his mind. He notes that
Arbinet’s IPO in 2004 was the highest that year,
beating Google which also went public. He clearly still carries
regret when discussing his past with
"We never took off after going public, and continued to fall
below Wall Street expectations. In the end they began to focus
on international voice. I moved on to other things [now founder
of VC technology company Governing Dynamics] but did try to
convince the board that I should be reappointed CEO in
2007. The 'geniuses’ at the board opted to hire
middle management from mediocre companies. It’s
very painful when it’s something that you have
created and nurtured, then have to watch it get yanked away and
Enron filed for bankruptcy in late 2001 and it is a widely held
view within the market that its failure to establish a trading
exchange platform was a major proponent to its collapse.
Although the service was relatively small in contrast to its
interests in commodities and energy, the company invested
aggressively in developing the business to sit beside its
Here, industry experts offer contrasting views to why the
company failed over 11 years ago:
"Enron failed because it was a corrupt and
dishonest company on a grand scale. The success and
difficulties they had with bandwidth were no different
to anyone else. They thrived initially, and came down
rapidly after targeting an extremely flawed
"They were doing all types of crazy deals.
They attempted to sell future options to people like
Microsoft for bandwidth distribution because they
thought bandwidth pricing was going to go down. They
were not real deals. They were fake with the intention
to keep stock high."
"In 2000-2001 everyone was digesting the
amount of millions they had invested in networks and a
lot of operators tried to differentiate services by the
type of fibre and equipment they used. Enron told them
what they had built was a commodity but no-one cared.
There was not much response to them and they were
judged in a bad way."
"It was the good old boys network in those
days. Carriers held assets close and they
didn’t want to give up that type of
control. They did not want to play in that field with
Enron. Enron had a great idea but it
wasn’t forward thinking enough from an
"Enron was trying to take the idea of
bandwidth trading and at the same time operate a
network. It was a disingenuous way of trying to drive
traffic to network infrastructure and establish
interconnections all over the world. They tried to hide
the fact that they were powering a network and driving
pricing at the same time."
Richard Elliott has nailed it; primarily they were corrupt and dishonest and arrogantly incapable of listening to what their would be customers were telling them.
Feb 05, 2013