Copyright © 2001, CIMI Corporation, Voorhees, NJ, USA. All rights reserved. This issue may be printed or stored in Web format for personal, non-commercial, use, provided that the entire issue including this copyright notice is reproduced and included. Portions of this issue may be quoted, printed, or stored providing that the subject portion is annotated with the issue identification above, and is attributed to the copyrighted material of CIMI Corporation. Other publication, reproduction, electronic storage or retrieval of this material, in whole or in part, without the express written consent of CIMI Corporation, is prohibited.
This section is not provided in this month's newsletter. Our next Market Area Focus will be published in the June issue, and will address new-age voice technology. Is it dead, or ready to take off? Remember, this section is always reserved for subscribers.
The continued slide
in the technology stocks probably isn’t making this a happy spring for
anyone in our industry, but all is not lost (yet).
The current slump has many causes, but one reason it’s sooo
deep is that companies are dumping all their bad news at once to prepare
for a fresh start. For CFOs
who have been trying to keep up with inflated growth estimates, the
“recession” is an opportunity to recognize the inevitable shortfalls
and blame macro-economics. We
expect the market to whipsaw around its current price levels, but the
bottom will be reached at the early end of our May-to-August forecast,
as we now see it.
The question will
then be how long we’ll stay at or near bottom, of course. Recognition of the need for significant regulatory reform
seems to be broadening in the financial markets and among analysts and
editors, which is good. A
Barrons article on March 12th pointed out some of the reasons
we hadn’t had regulatory reform in the last two years, and weighed in
favor of such reform for the current session.
A think-tank review of the performance of the telecom act (http://www.newmillenniumresearch.org)
referenced in Network World’s new Vortex.net newsletter contains a
number of short papers that all seem to conclude that additional
tweaking (at the very least) of the Telecom act is essential.
Congress
doesn’t seem to be moving, though.
As we predicted in our December issue, Washington is preoccupied
with other issues this spring, and there doesn’t seem to be any rush
to get any revision to the Telecom act introduced.
The 106th Congress version of the bill (S.2902) has
not been resubmitted, and we believe that significant revisions to the
old version will be introduced before it is even presented for a vote.
Thus, we are probably looking at late in the 3rd
quarter at best before passage could be expected.
The
nature of the regulatory changes made will decide the way the market
moves. So far, we can see
the following possibilities:
·
Congress would fail to move at all, or move on a very
long (2 year) timeline. In
this case, the FCC’s decision to expedite RBOC admission into the
long-distance markets would have the effect of shifting RBOC priorities
to voice services. This is
because both voice and data infrastructure would be subject to
wholesaling and unbundling, but less incremental voice investment would
be required to enter long-distance competition, and the revenue stream
is more provable in the near term.
·
Congress would move to restructure the Telecom Act in
both the voice and data sense, aimed primarily at meeting the demands of
AT&T, Sprint, and WorldCom (using the dying CLECs as straw men) to
fetter the RBOCs further. This
would have the effect of suspending the entire market, propping up the
IXCs for a couple of years, but at the expense of developing a vital
industry.
·
Congress would move to exempt data services from
regulation under the Act, in a way that simply codifies the FCC’s
“separate advanced services subsidiary” and “packet exemption”
rules into law (circumventing the Federal Court reversal of the first,
and challenges to the second). This
would restore the industry balance to where it was in June of 2000, when
the RBOCs were promoting fast DSL build-out.
·
Congress would move to exempt data services from all
unbundling and wholesaling, without need for structural separation.
This is the thrust of the S.2902 framework, and it would
essentially open RBOCs to exploit data services without any
infrastructure or service sharing.
We
think that the first and last of these options are the least likely, but
not to be ruled out. The
middle two reflect how lobbying influence might be balanced in the
current administration. In
the past, the Republicans have generally favored the RBOC position,
which would make the third option more likely.
However, a complete rethinking of the Telecom Act is not out of
the question; with campaign reform possible, parties might be
susceptible to lobbying in the near term to fill their coffers with
bucks before the tap is turned off.
The
big question here is the extent to which any changes in regulatory
posture would protect RBOC investment in data infrastructure, including
DSL access. The more the
protection, the more likely that investment balance would shift into the
data space. Without
protection, the fact that the RBOCs are already fairly well built out in
voice terms and have little data networking might induce them to focus
on competing for IXC voice customers, where better ROI (arising from the
lower incremental investment needed for voice) might be obtained.
Another
development in this space may influence the congressional agenda.
The Pennsylvania PUC has, for over a year, been responding to IXC
lobbying of its own to demand that Verizon break itself up into multiple
companies (wholesale and retail) within Pennsylvania as a condition of
offering long-distance services. Here,
the IXCs clearly have control of the process, and the question is
whether Verizon will bow to the pressure, appeal to the FCC to assert
primary jurisdiction, or resist in court.
The
RBOCs have long believed that the TELRIC pricing for wholesale services
and the Universal Services subsidies were insufficient to recover actual
cost, and that the result was that their retail profits in part
subsidized their competitors. Some
RBOCs, at least, saw this kind of separation as a logical response to
the regulatory trends—a defense against their being used to prop up
the CLECs. That an RBOC
would be forced to do it as a so-called pro-competitive step is a
reflection of how the IXCs have used the CLECs as pawns in their own
game. At this stage in the
market, a breakup of this nature would be singularly dumb, and have no
possible outcome positive to the actual consumer.
PUCs, however, are even less enlightened than Congress in many
cases, and it’s not impossible that this issue will be decided on the
question of whether Verizon’s being forced to break itself up in
Pennsylvania would have multi-state impact and thus be rightfully an FCC
issue.
We
think that the Pennsylvania case will proceed based on a combination of
Verizon’s interest and those of the new FCC chairman, Powell.
If Verizon thinks that it can gain some benefit from the
wholesale/retail split, the FCC may sit back based on its “free
market” commitments and let nature take its course.
If they don’t like separation and Pennsylvania moves to break
them up anyway, that same free market stance may deter Powell from
stepping in.
Breaking
up RBOCs would be a monumental business task, and one that would
preoccupy them for several years. If
that happens, we can be reasonably sure that they’d make no major
infrastructure or service commitments during the period of change.
What
all this adds up to is that industry pressure is building on the
regulators, but the sins of neglect we’ve sowed in the past are also
coming to roost on us. If
pressure builds slowly, it may be that some of the old threads of
manipulation (like the PA-PUC) will bite us before we have a chance to
pass legislation that addresses the problems.
We continue to believe that there will be some reform later this
year, but we must admit that the progress so far is alarmingly slow.
The stock market
responded poorly to the Fed’s half-percent March rate cut. That reflects an unfortunate but probably inevitable
truth—the stock market wants another wave of hype and nonsense and not
a real recovery, because tangible results take too long.
We believe that the Fed has made a decision (reflected in
Greenspan’s comments about the over-hyped nature of the market back in
1Q00) that they would not intervene to prop up an equity market that was
out of control. We believe
that policy will continue in force, and that it will be the best policy
in the long term. However,
it will insure that networking will have to earn its way back into
investors’ hearts. That’s why it is so important that we address the
regulatory problems.
This material is edited for Web publication; critical market statistics are removed. The full document is reserved for our newsletter subscribers.
While the prospects
for DSL empowerment are tottering in a regulatory sense, cable modems
seem to be on a regulatory fast track.
While the cable TV industry consists of thousands of separate
cable systems, most subscribers are already served by Multiple System
Operators (MSOs) who run collections of systems.
In the past, FCC rulings have limited the size of the MSOs, but a
recent Federal courts decision to overturn the FCC’s restrictions on
cable company consolidation seems likely to promote development of cable
supergiants, each with the power to service tens of millions of users.
That makes cable modem technology second on our list of important
access options.
Cable deserved a
high position before, in the view of many industry observers.
Cable infrastructure, like phone infrastructure, passes nearly
95% of the credible broadband opportunity in the US, and a slightly
lower percentage in the other industrial centers of the Pacific and EU.
Based on a combination of fiber and CATV cable, the plant is
reasonably well suited for delivery of residential and some business
services. Arguably, it
could be made to be more suitable through incremental investment
of magnitude similar to that required to create universal high-speed DSL.
One factor that’s
held cable back is its generally low level of credibility. While cable companies have made great strides in reliability
over the past six years (the average consumer reports a cable outage
once each 16 days today, versus once each 6 days in 1995), their
performance doesn’t even come close to that of the phone companies.
Consumers cannot provide any reliable estimate for the last time
they had a phone failure. This
credibility gap versus their major competitor forces cable companies to
focus on home data applications for their cable modems, because
businesses are more critical of their access networks.
This, in turn, has tended to keep the cable companies in the
low-margin segment of the market.
The regulatory
shift that might enable a massing up of cable companies may provide an
answer to this problem of credibility, since most in the industry feel
that capital strength is the basis for developing a reliable cable
network, and subscriber mass is the only way to develop capital
strength. FCC limits on the
size of cable companies, in subscriber terms, clearly limits their
ability to build their businesses and develop deep financial pockets
from which to disburse funds for modernization.
But the cable
players aren’t without their problems.
With relaxation on the restrictions on cable company size is
likely to come increased regulation of their behavior.
Nobody (even the Republicans) want a monopolistic environment
developing in the cable industry, particularly when consumers already
feel they’re being gouged by cable players in rates for basic and
advanced services. In
addition, cable companies must work out of the low-margin Internet game
if they are to stand as equals to the phone players.
It won’t be easy, and both cable infrastructure topology and
basic cable modem technology have their own unique plus-and-minus
contributions to make to the service credibility process.
A simple cable
distribution network consists of a CATV copper trunk or “run”, to
which a series of individual subscribers are connected via taps.
Each tap can be connected to a home termination, which could be a
cable-ready TV, a set-top box, or a cable modem. Since the basic topology is one of a shared, multi-drop,
link, multiple home terminations can be supported per subscriber—and
almost always are.
This simple
structure has major limitations in terms of distance, number of
customers, and total bandwidth available.
The competition posed by the satellite providers of TV service
would, even without cable data applications, have tended to drive
providers toward a hybrid fiber-coax (HFC) architecture.
Most also believe that HFC is essential in providing higher
reliability and supporting lower craft costs.
HFC cable networks
consist of a head-end location where long-haul networks deliver services
to the cable provider, and a distribution network that gets those
services connected to the customers.
The distribution network begins with a fiber system that links
the head-end to a series of fiber nodes, each of which serves a
copper-based customer delivery community.
Cable data is
supported over this infrastructure based on a set of standards that are
generally homogeneous worldwide, called DOCSIS (Data Over Cable Service
Interface Specification). DOCSIS
version 1.1 is the “current” version, but most cable modems today
support the basic 1.0 specification.
DOCSIS defines communications between a Cable Modem Termination
System (CMTS) master and a series of cable modems.
DOCSIS is a
four-layer architecture, with the lowest layer being the physical
modulation layer, the next being the information format and transmission
convergence layer, the third the media access layer, and the last the
data link encryption layer. Applications
are layered on top of these layers, so Internet access would involve
adding the IP layer plus TCP, UDP, and whatever other protocols were
used. We’re not going to
get into the dirt of the lower layers for this piece; refer to any of a
zillion works on DOCSIS for that. We do note that there are multiple
modulation layers possible, and that two (the US and the EU) are already
in use.
The information
format and transmission convergence layer of DOCSIS is based on MPEG-2,
a protocol that allows video, voice, and data to be sent independently.
MPEG-2 is baed on synchronized 188-byte frames or packets, and
each MPEG-2 packet has a program ID (PID) that identifies the
information application or target device.
A standard PID is assigned to the cable modem application, and so
data flows on that PID only. When
a cable modem is syncing up, it’s searching for a channel on which an
MPEG-2 flow is found having the CM PID value.
The modem must then determine if the channel parameters for that
particular channel match its capabilities, and continue searching until
one is found. It them
retrieves the control parameters for that particular channel, and sets
itself up as a member of that community.
The return path on
a DOCSIS system is based on both TDMA and FDMA.
FDMA allocates modems to return path channels, and TDMA slots
modems who compete for a single channel into mini-slots (16 to 128 bytes
long). These slots can be
assigned to modems, groups of modems, or just stuffed into a pool for
contention. The use of
allocated mini-slots reduces collisions and retransmissions, and
regularizes performance in the access network.
Return path bandwidth is based n a bunch of modulation issues,
but works out to be between 320 kbps and 10.24 Mbps, as determined by
the CTMS control process.
The MAC layer of
DOCSIS is responsible for access control, and it’s based on a
request/grant exchange. The
CM requests permission to send a block of information, and the CMTS
grants it by assigning a mini-slot in the TDMA return path.
This allows the CMTS to exert some control over the bandwidth
allocation to each station, and makes the cable modem model more
deterministic than a simple CSMA LAN would be, for example.
Above the MAC
layer, there’s an encryption layer whose goal is to insure privacy of
the data on the cable, given the intrinsic multidrop architecture of
cable. This isn’t a true security system, simply a mechanism to
prevent (or reduce) data interception or eavesdropping. Encryption takes place below the payload IP layer, but it is
applied only after the IP parameter exchange and authentication process
is completed.
DOCSIS 1.1 provides
some functional extensions to the basic DOCSIS.
First and foremost, it includes a packet classification and flow
management procedure that could be used to provide QoS management for
traffic of various types that involve a single CM (remember, DOCSIS
provided some traffic control between competing CMs before).
Support for MAC fragmentation of frames reduces the impact of
serialization delay in long data packets on concurrent non-data traffic
types. The new version also
includes a better encryption/privacy management system to reduce the
risk of hacking (as opposed to simple eavesdropping) and support for
efficient IP multicasting to multiple CMs (the old standard required
duplicating the packets).
Cable modems
support more broadband Internet connections than DSL modems, and the
regulatory-induced slowing of DSL deployment will probably insure that
remains true through at least the first two or three quarters of this
year. Cable companies in some areas are promoting their offerings
more aggressively, using discounts and other incentives to pull in
prospects while DSL is on hold. Some
believe that this will give cable a big leg up on the phone companies.
Others, though,
believe it’s a race to lose money.
As we’ve pointed out repeatedly, simple Internet service at any
speed is not profitable in itself, based on the financial filings of the
ISPs. Access to an unprofitable service is also unprofitable.
For the cable industry, the success of cable data or cable
anything (other than cable broadcast video and perhaps audio) is
dependent on identifying and supporting a profitable business model
beyond the Internet service model.
Can that happen?
How about voice?
Pundits say voice over cable is the next revolution, but the same
pundits have been revolting (no pun, of course) for years in many
diverse spaces and have yet to be right.
The cable companies have a number of serious problems in the
voice space:
1.
Users have consistently been reluctant to trust cable companies
with voice, especially the business and higher-end residential users
whose expenditures for voice would necessarily make up most of the
profit. Our surveys have
consistently shown that cable voice credibility is so low that discounts
of 30% or more versus normal service pricing would be needed to overcome
the barrier.
2.
Voice service return on capital for a new player is limited by
both the strike price for service (see above) and the relatively high
cost of voice infrastructure incremental to the basic video cable
service plant. We estimate
that cable voice would return about half the necessary value to make it
a viable business.
3.
Feature enhancement to voice services, while credible and
demonstrably profitable doesn’t seem to be a priority with the
equipment players. It’s
questionable whether a viable feature-driven voice offering could be
deployed today based on available technology.
Content networking
is the other play, and here the problems of the cable players may be
even worse. The primary
difficulty is that the most credible near-term revenue stream in the
content space comes from video on demand.
Not only does digital video have a very high bandwidth
requirement (beyond, in our view, the capabilities of the currently
deployed cable modem systems), it competes with the cable companies’
own broadcast video business. Would
they deploy a self-competing technology that effectively validated the
video-over-DSL market for their RBOC archrivals? Ha!
Perhaps the worst
problem the cable people face, though, is the restructuring of their
industry. It appears that
the Powell FCC has concluded that cable is the competition for DSL, or
at least that the FCC policy of an open marketplace must apply
even-handedly. Combine this
with the Federal Court decision that the FCC lacks the authority to
limit the number of subscribers an MSO has, and you have the formula for
industry consolidation. Periods
of consolidation are rarely periods of aggressive service and technology
change.
Where, then, cable
modems? We think that if
DSL deploys on a fairly large scale and validates public IP applications
beyond the Internet, some of those applications (other than video on
demand) will percolate over into the cable space and raise the ROI for
the cable players.
A dead stop in DSL
will tend to favor cable for a time, but ultimately either DSL will pick
up or residential broadband will die back.
The cable players, in either case, could strand a substantial
capital investment and let themselves open for increased pressure from
satellite competitors in their core entertainment space.
Cable is symbiotic with DSL, more than competitive, folks.
This section of our newsletter
presents the response of Canadian management pioneer Syndesis to our
service management framework. Sorry, but it's
for subscribers only!
Netwatcher is a copyrighted information product of CIMI Corporation. Subscription is on a controlled circulation basis only. Subscription information may be obtained by clicking here.