[VOIPSEC] 4G Issue Map: signaling complexity - more

Thu Aug 31 11:28:49 CDT 2006

Some thoughts:

- ENUM is but an open protocol for a deeper value added service, best kept
invisible from the user.
- Users don't buy protocols. ENUM allows somewhat efficient access to
old-school dialtone-based SIP peers, that's it.
- Nobody should know or care about ENUM apart from us.
- SPEERMINT is good enough if you take care.

A good service that would get cash-money traction in the real world >>right
now<< would offer the following:

-  An Internet telephony service (ITSP) provider providing a quality circuit
through OEM DSL or through other means.
-  The ITSP provides a SIP registrar that does far-end NAT traversal for
both SIP and media.
-  The ITSP automates/hides peering between subscribers, including offering
relay support for NAT traversal, perhaps using ICE.
-  The ITSP encourages use of intelligent codecs such as speex for
ingress/egress to/from the PSTN, and so is resource efficient.
-  The delay between the OEM DSL termination and the egress to the PSTN is
minimal.
-  The underlying DSL connection is real-time VBR.
-  The ITSP would be federated with others so that their aggregate members
peer automatically.  I include Skype and Google in that.
-  The ITSP offers an out-of-band provisioning system over SSL so that MACS
becomes automated.
-  The fee is PSTN minutes + country-code minutes, first 40 gigs per month
is free, a fee per gig after that for people using video.

That network itself need not be on the public internet, but it could access
nodes on the public internet through a NAT relay.
That would address SPIM and SPIT fears and also address security, and it
would be very cheap.

An ITSP that does that would get a bizdev deal with my employer in minutes,
and access to a real channel.

Some notes:
      - I recently hacked my residential speedstream router to support real
time VBR DSL to my OEM DSL provider. It now works.
      - I get 4m down 800K up real-world cross-continent with very
low jitter.  It costs me $36/mo canadian with fixed IP.       - Such a 5m/1m
ADSL circuit using Speex as the underlying codec for efficiency could serve
a campus of up to 10,000 users.
      - My provider took my suggestion and now uses it for their default
config.  They will soon be able to provide free member-member
        VOIP to their customers with a low fixed rate for PSTN telephony.
www.ncf.ca is my provider.
      - It has a paid staff of six.
      - I estimate they will need three people to manage a good telephony
offering, plus a sales person.  That's a staff of ten.
      - They could cover a geographic area of 2000 square kilometers with 2M
people using only that for infrastructure.
      - Their annual total budget for everything would be in the mid six
figures.
      - This means that their overhead is tiny, and even thin margin means
money.
      - Their customers would have full stereo calling and video calling and
presence.
       - They could offer this service for 10% of the cost of a fractional
T1 and make lots of cash.
      - There is no other service apart from the good dumb pipe.

If the carriers think Skype is scary, they're mistaken. THAT is the
immediate threat, the bargain-bin walled garden.  Once someone with branding
like Dell does it, through OEM, it's over.

On 8/31/06, Henry Sinnreich <hsinnrei at adobe.com> wrote:
>
> > ENUM is going to a failure for 3 main reasons
>
> Well, there may be another reason:
>
> Tying advanced communication to the broken phone numbers addressing
> system instead of using Internet style URIs is pathetic. My two SIP
> phones and several SIP UAs on the desk allow me to speak to anyone else
> on the Internet, see their presence, IM, do video, etc.
>
> My Skype clients and Skype cordless phone can connect me to >100M users
> without phone numbers. Skype deserves a mention here since it is living
> proof that P2P networks with their own addressing can be very
> successful. This is useful to remember just in case the DNS get
> perverted by walled garden networks...
>
> >3. Nobody wants to pay for an ENUM service.
>
> Right. All IP-IP calls are free anyway.
>
> Forgive me for shouting: THE EMPEROR HAS NO CLOTHING.
>
> Thanks, Henry
>
> -----Original Message-----
> From: Simon Horne [mailto:s.horne at packetizer.com]
> Sent: Wednesday, August 30, 2006 9:15 PM
> To: Geoff Devine; Michael Slavitch; Paul E. Jones
> Cc: bill at flanagan-consulting.com; Henry Sinnreich; Voipsec at voipsa.org
> Subject: Re: [VOIPSEC] 4G Issue Map: signaling complexity - more
>
>
> ENUM is going to a failure for 3 main reasons
>
> 1. Most VoIP networks are very little security and almost no call party
> authentication. Some VoIP protocols are securable to some extent but
> others
> like SIP are far far more complex. Just detecting the vulnerabilities of
>
> protocol is not good enough, there needs to be solutions to combat
> Internet
> telephony SPAM or Phishing before services like ENUM or public URI could
>
> even be considered. The guys offering ENUM services are fighting a
> losing
> battle convincing SIP vendors to implement TLS (although not a complete
> solution) but it seems in the main part, the problem is being avoided
> and
> the easier fortified bridged VoIP Islands approach seems to be is the
> way.
>
> 2. Some will argue there's no revenue in providing ENUM services. There
> is
> more money charging fractions of cents for providing bridging services.
> But
> the big question is, after spending years building the bridges and
> installing expensive equipment, are customers willing to pay for a
> service
> that they were told was "free" or "near free". The Boeing in-flight
> Internet access is a prime example. The service was great, had great
> promise just no-one wanted to pay for it.
>
> 3. Nobody wants to pay for an ENUM service. APRA have priced themselves
> out
> of whatever market they were trying to create, It's so easy to setup an
> ENUM service. There are already www.e164.org and www.e164.info for
> instance
> who offer the service for free.
>
>
> Now with that said, what are the opportunity costs of not tackling the
> security problems of standard based protocols like SIP and not opening
> networks up to the public internet? Possibly a lot more than the cost of
>
> providing expensive VoIP bridges that perhaps no-one wants to pay for,
> it
> could be billions...one easy example is Google and E-Bay recent
> partnership
> to provide click to call and pay per call services for web advertisers.
> Huge opportunity lost...
>
> Simon
>
> At 09:11 AM 30/08/2006, Geoff Devine wrote:
> >I also think ENUM is going to be a total failure.  People will stay
> with
> >their trusted service providers to avoid getting SPAM phone calls at 3
> >AM from somebody claiming to be General Mubuti in Nigeria but who is
> >actually trying to scam their bank account or sell them erectile
> >dysfunction drugs.  ENUM-like services that reference private databases
> >administered by those trusted service providers will be the norm rather
> >than a free-for-all e164.arpa.
> >
> >
> >
> >Geoff
> >
> >
> >
> >
> >
> >________________________________
> >
> >From: Michael Slavitch [mailto:slavitch at gmail.com]
> >Sent: Tuesday, August 29, 2006 5:16 PM
> >To: Paul E. Jones
> >Cc: stuart jacobs; bill at flanagan-consulting.com; Henry Sinnreich;
> >Voipsec at voipsa.org; Geoff Devine
> >Subject: Re: [VOIPSEC] 4G Issue Map: signaling complexity - more
> >
> >
> >
> >One more short note before closing off my contribution to this thread:
> >
> >
> >
> >While the telecoms carriers are looking at the cable companies, a far
> >bigger threat will come from companies like Akamai:
> >
> >
> >
> >http://www.akamai.com/
> >
> >
> >If many-to-many conferencing, SIP ENUM peering and one-to-many live
> >video needs geographically distributed CPU's, such vendors have all
> that
> >is needed in hand.
> >
> >
> >
> >_______________________________________________
> >Voipsec mailing list
> >Voipsec at voipsa.org
> >http://voipsa.org/mailman/listinfo/voipsec_voipsa.org
>
>
>