Onion or Parfait

1
Onion or Parfait?
Loring Wirbel Editorial Director,
Communications CMP Media LLC

• Layers of Fraud in Collapse Layers of Innovation
  in Recovery

2
How can a hangover last so long?
3
Necessary Dominos

• Internet content to ISP
• ISP to CLEC
• CLEC to IXC
• IXC to infrastructure
• Infrastructure to access
• OEMs to components

4
Inevitable Domino 1The Dot-Com

• Funding of bad business plans (pets.com, Funerals
  Online) made crash inevitable
• Funding of companies with no business plan
  (Razorfish, Pseudo.com) made crash imminent.
• Since broadband demand was dependent on dot-coms,
  the telecom crash was foreordained at time of
  dot-com crash of April 2000.

5
Inevitable Domino 2The ISP

• Too many ISPs were directly dependent on
  physical-layer buildout (Flashcom), while not
  contributing to phy-layer maintenance.
• Founders had touch of savings-and-loan
  crookedness to methods, e.g., free DSL
• Direct dependence on CLECs growing solely off
  footprint, not profitability within one region.

6
Inevitable Domino 3The CLEC

• When specialized CLECs (Rhythms, et. al.) were
  funded by conspiracies like Enron/Salomon,
  failure was assured.
• Investors said fiber buildout was paramount in
  metro, even though too much fiber was in ground.
• Geographical footprint expansion demanded over
  service and support.
• Last-mile access dependent on ILEC adherence to
  1996 Telco Reform Act Ha!

7
Inevitable Domino 4 The IXC

• Long-haul dependent on massive increase in data.
• Fiber swaps of imaginary earnings used to pump up
  Qwest, Level 3, Sprint
• Circuit-switched voice becomes loss-leader, but
  data growth refuses to grow at rate to insure
  profits.
• IP-based soft-switching doesnt happen

8
Inevitable Domino 5 Infrastructure OEM

• OXC market highly overvalued Corvis greater
  market cap than GM???!!!!
• No exponential IXC growth means no reason for
  Sycamore, Ciena, etc.
• Cisco, Nortel see collapse coming in 3Q2000, yet
  increase manufacturing in absurd game of
  chicken
• Result? Billions in useless inventory.

9
Inevitable Domino 6Access OEM

• Vanished CLECs, lazy ILECs mean no DSL, no reason
  for DSLAMs.
• Heavy hand of RIAA, MPAA kills Napster and most
  near-term appeal for consumer broadband.
• Collapsing economy continues to make cable modem,
  residential gateway a luxury item for disposable
  income.

10
Inevitable Domino 7Components/Semis

• Photonic market dead by early 2001.
• Phy-layer chips at 10G and above problematic by
  mid-2001.
• Network processors, VoIP DSPs, 3G DSPs, all
  become irrelevant as routers and gateways are
  sold on eBay.
• . and the rest of the semiconductor industry
  comes crashing down.

11
This was visible by early 2001 Why the denial?

• March 2001 Optical Fibers Conference, CTIA filled
  with irrational exuberance
• Rip Van Winkle Syndrome Dont confuse me with
  the facts! What right do you have to use that
  alarm clock?

12
First Layer of Fraud Crooked CEOs and Their
Partners in Crime

• Long-time players like Bernie Ebbers knew
  precisely how unsustainable the continued growth
  model was.
• Convince, confuse, expand, cash out, move to the
  islands, and rip off the little guy
• Only a small percentage in this category

13
Second Layer of FraudInvestment Banks, VCs

• Knew very well the vast bulk of companies would
  not perform, as the Blodgett e-mails indicate.
• Financed startups for short-term cash-outs to
  bamboozle the everyday investor, make the
  Masters of the Universe insanely rich

14
Third Layer of FraudExecs in Startups,
Co-Dependent Companies

• Most knew very well the growth was phony, did not
  want to be the first to give up game of chicken
• Motivated by perks and ESOPs
• Giving up the game would entail (as public
  company) telling shareholders, We need to plan
  for no growth for several quarters. Right.

15
Fourth Layer of FraudIndependent Analysts

• From Grubman (semi-indy bank-related) to Gilder
  (cult guru), pundits either were willing
  participants in fraud, or blinded by own
  stupidity.
• Pontificators of angels in the photons bear
  criminal responsibility for misleading investors
• What about media? Industry Standard, Wired, Red
  Herring, Fast Company.

16
Fifth Layer of FraudDay Trader (You?)

• Runups of some stock types were patently absurd,
  even at time of runup.
• Internal bank analysts had little respect for
  yahoos playing the market.
• Those complaining of 401k decimation have few
  legs to stand on if they participated in like
  activity.
• (Example Enron Energy trading itself was a
  Ponzi-like derivatives game. All supporters,
  employees, and co-dependents of Enron should have
  realized this.)

17
Escape from the Jungle

• Dont look to VC-funded startups the zombies
  are dying
• Dont look to large OEMs inventory overhang
  remains
• Look to ad-hoc, bottom-up specialists!

18
Consumer Need Broadband Access is Popular

• Cable modem deployment has hit 10 million
• Always-on is as popular a feature as the
  bandwidth
• WiFi boom is related to both the shared
  neighborhood characteristics from a broadband
  access point, and the mesh topology of ad-hoc
  routing

19
AnswerAd-Hoc Access

• WiFi cowboys cobble together networks from the
  ground up
• Coaxial, copper specialists boldly go where MSOs,
  CLECs fear to tread example CoaxMedia
• PON specialists bring fiber to the end user as
  cheaply as possible

20
CaveatsPay As You Grow

• WiFi libertarians must contribute to cost of
  access point no surprise if MSO and CLEC
  enemies are made!
• Converse WiFi top-downs (Boingo, Cometa) cannot
  build justification from the center outward
• Ad-hoc access will remain a bootstrapped
  business, with no gold mines in sight!

21
Business NeedClusters and Grids

• Why buy a SMP server if workgroup clusters fill
  the need?
• Serial interconnect becomes the gating factor
• I/O cards built on common form factors become the
  economic way to increase compute functions

22
AnswerCard and Chip Business for Hyper
Transport, Rapid I/O, PCI-Express, etc.

• Standard links, standard backplanes lead to
  low-cost clusters for Web servers, workgroup
  servers, etc.
• What starts as a box-to-box serial phenomenon can
  migrate inside the box through standards such as
  HSBI
• Still good opportunity for smaller board-level
  OEMs and phy-layer chip cos.

23
CaveatsBeware the Server OEM

• When Sun, HP, IBM define the grid function calls
  for the OS, they can define the partners
• Will the server manufacturers take over the
  interconnect business?
• All Layer 1-3 serial functions subject to cost
  commoditization, even price-bombs

24
Business NeedCommon Layer 2 Link from LAN to WAN

• Ethernet everywhere movement has dominated all
  thinking from serial transport to wide-area
  public network service
• Easy scaling from 10 Mb to 100 Mb to 1 Gb to 10
  Gb allows for several players to address
  innovation at phy-chip, MAC-chip, module, and
  board level

25
AnswerEthernet Replaces ATM, TDM, Frame Relay

• Simplicity of design allows better inventory
  control of line cards
• Difficulty of isochronous time-stamping means the
  inevitable death of Sonet will simplify transport
• Can Ethernet serve as a Layer 2 stand-in for all
  serial link functions?

26
CaveatsPrice Bombing Meager Functions

• The broader the application base, the more
  commoditization can rule
• If you send Ethernet out to do a Sonet or ATM
  job, expect a second-rate QoS
• Multi-Protocol Label Switching can help improve
  the IP/Ethernet/opto stack, but Ethernet
  everywhere has definite limitations

27
Business Need (Enterprise and Carrier)Upgrade
Existing Nodes

• Businesses will retain existing routers,
  switches, add-drop multiplexers until they fall
  apart
• Node OEMs may go out of business, or fail to be
  timely in line-card upgrades
• CapEx budgets will not allow for new node
  platforms until 2005 or later

28
AnswerLine-Card Upgrade Specialists

• Standard backplanes mean good opportunities
• Even proprietary backplanes, midplanes may belong
  to bankrupt OEMs
• A feasible business can be built through small VC
  rounds of funding, or even bootstrapped/angel
  financing

29
CaveatsOverthrow and Obsolescence

• Its good to serve legacy markets, but dont stay
  in TDM, ATM, or Sonet too long!
• Who owns the intellectual property for the line
  card you develop? Has someone acquired rights
  for a bankrupt OEM? Can an existing OEM sue you,
  or undercut you in price?
• Challenging the owner of a backplane technology
  directly is a sure way to get clobbered a
  glancing blow is more effective

30
ConclusionThe Punk-Rockers Model of DIY (Do
It Yourself) Works!

• Dont waste your time waiting for Cisco and
  Nortel to start re-hiring
• Dont create a business plan for a VC, with
  expectations of tens of millions
• Develop a solution to a defined problem, find a
  cheap way to bring it to market, and bootstrap it
  or fund it to the lowest extent possible

31
and dont forget to keep business ethics
paramount,this time around!

• Would momma approve of your business plan?