Next-Generation Communications Services: Issues and Opportunities John C Klensin, Ph.D. TWNIC Conference, 2004 March
Looking Two Steps Ahead Talk is less about What is happening today or what one should do next But, instead, about What to start looking at now for the steps after the next one.
New Services or Simulation of Old Ones in New Medium Is the Internet a good telephone? –VoIP in closed networks Equipment savings ? Still simulating SS7 ? –Internet Telephony Service quality, customer support, and expectations Economics and tariff arbitrage
Internet Telephony as a Substitution Service Dubious long-term economics Good short-term Very good for equipment providers
Tariff Arbitrage Very good short-term business In the long term, one of several things happens –Regulators impose comparable tariffs on new activities –Tariffs are abolished, reducing or eliminating price advantage –New activities are prohibited –New activities are ignored; old, tariffed, way of doing things disappears
Terminal Signaling Capability as a Defining Characteristic Service models determined by twelve buttons (or worse) Consequences –Smart Central Switches –Intelligent Network –Menus and VRUs to increase call setup bandwidth
The History of Distance Communications Low bandwidth – High speed –Signal fires, drums, telegraph, telex –Telephone ? High bandwidth – Low speed –Packets of letters carried by horse or coach –Bags of letters carried by train –Tapes in the back of an automobile –Fax or other images of paper ?
Reviewing “Convergence” Carry the torch onto the coach ? Load the horse onto the train ? Tapping on the telephone handset ?
The Actual Pattern Parallel use until One technology drops off
Advanced Services in the PSTN Paradigm “Integrated Messaging” –Voice, voic , fax, pager to one phone number –Remote pickup and intercept Call forwarding and similar routing The Instant Messaging problem
What Are You Trying to Reach Surrogate for a copper pair leading to a specific terminal device? A person or function? Almost always the second
Specifying a Target Preferred medium ? Person or alternative ? How important ? Interruption levels and tracking/ forwarding
Permitting a Source Receiving a connection should be a negotiation… Do you want to be reached? By the caller? With what priority ? How much are you willing to be followed around? What do you think of the caller’s priorities?
Example: Phone call with forwarding and roaming Colleague places a call to US “office number” at 2PM. Phone rings at 3 am in Singapore Obvious questions… –Would it have been placed if destination and time were known? –Should it be received without knowing its importance? –How does one guess at time of recipient when country and city codes are meaningless?
Example: Whom am I calling? Number reaches a terminal or surrogate. People may be widely distributed. Long VRU menus seem to be our best solution, but cannot be the right answer. So –Call person or function, not a number –If we are to number people for convenience, E.164 phone numbers are probably the wrong model
Can This be Done in the PSTN? Maybe Big scaling problem, high complexity Very difficult authentication problems –Current “four digit PIN” strategy in many countries not good enough –Limit to setting from “home” phone provides poor service.
ENUM itself may be the wrong model E.164 is not only tied to phone system semantics but to complex regulatory politics While the routing environment tied to E.164 is implemented by bilateral agreements, ENUM creates the first global telephony regulation opportunity for ITU. Should it have been –Number.CityCode.3166-country-code.enum… ? –For example: tw.enum… or even enum… ?
Who Needs “Internet Telephony”? Internet → PSTN –Yes, but for how long and at what rates? Internet → Internet –Not needed; use NAPTR records with names PSTN → Internet –Not for Central Office switches: better ways –Not for dual mode phones: Internet → Internet or POTs → POTs devices –Smart-routing PBX switches? Maybe.
Networks: Central Control and Edge Control Edge-based networks permit distribution of control functions –“My server”, “my agent”… not tied to CO switch –Different people/ organizations can get different functions –Lots of competitive business opportunities Avoid both “one size fits all” and option- complexity problems
A Different Communications Model – Initiator Specifies preferences –Person or function to be reached –Preferred/ ranked contact medium Simultaneous voice, voic , fax, , assistant,… –Priority/ importance –Conditions E.g., “don’t interrupt if…”
A Different Communications Model – Receiver Specifies rules for people/ groups/ defaults –Preferred/ ranked contact medium Simultaneous voice, voic , fax, , assistant,… –Assessment of priority/ importance statements More priority from some people than they specify Less for others –Relationship between Derived priority and Acceptable medium
Generalization Extensions are almost trivial for –Multiparty communications “conference calls” and group discussions –Multimedia connections Images, text, sound, interactive remote whiteboards –Asynchronous and semi-asynchronous communications , fax, instant messages, “push” voic The special case –Two party, not prearranged, fully synchronous, audio- only,…
Some Lessons from Instant Messaging Is being interrupted a good thing? –Maybe better than by the telephone –More choices: Identification of caller, not calling number Ability to delay response somewhat, not pure real-time Controlled access to interrupt –Need more than Available or not Friend or not How to divert to , or…
Could we do this? Technology basically exists Changing styles of thinking moving away from “make it look like a telephone” may be harder Designing a rule-specifying system that is Sophisticated enough to be useful Simple enough for consumers to use –Is not trivial. But not impossible either.
What Next? If we build it, will anyone come? How bad does –information overload –Interruption overload need to get before we do something real about it?