Transcript ACIF NGN for ITU

Next Generation Networks
Australian Communications Industry Forum
(ACIF)
ITU-T NGN Workshop, July 2003
Peter Darling
Manager, ACIF NGN Project
ACIF NGN Project
• ACIF’s Strategic Plan in early 2001 identified
need to work on “Next Generation Networks”
• Meetings with the Australian regulators (ACA and
ACCC) and the Industry Association SPAN
confirmed they had a similar interest.
• ACCC sponsored an initial consultancy in
second half of 2001 “to raise issues”
• ACIF held an NGN seminar in May 2002 to scope
the issues
• Attendees proposed a continuing industry
“conversation” on NGN matters.
2
ACIF NGN Project
• The ACIF Board agreed to support an ACIF NGN
Project, working through the ACIF NGN
Framework Options Group (“NGN FOG”).
• The aim of the ACIF NGN Project is to help all
involved discuss issues that cross current
boundaries, including
– Internet/telco divisions
– Regulatory issues (ACA and ACCC)
– Industry issues (including self-regulation
requirements)
– Policy issues
• An early agreement was that user requirements
must be the main driver of this work.
3
NGN FOG Work
• The main task of the NGN FOG has been to
assist understanding of the transition to next
generation network equipment. The NGN FOG
work involves consideration of issues including
–
–
–
–
–
Technical standards
End-user issues
End-to-end services
Interconnection across networks
Regulatory issues (both self-regulation and
government regulation
4
How Will “NGNs” Develop
• ACIF sees a number of options (not mutually
exclusive)
– From the existing public Internet
– From the introduction of enhanced Internet
Networks for corporate customers
– From the introduction of new networks using
packet technology
– From the upgrading of existing networks using
packet technology
5
PSTN
ISP
“Carrier Grade”
IP-based
Networks
2G Mobile
ISP
?
?
ISP
“The Internet”
(best-endeavours
network)
ISP
6
•
•
The Internet View
– Interconnected networks will mainly serve to provide end-toend connectivity, carrying packets of data end-to-end between
smart terminals, and establishing end-to-end sessions under
terminal control as required.
– Services will be provided by interaction between end-user
equipment (e.g ICQ/MSN style voice +, web access)
– Legacy carrier-based services will be in decline
The “Telco” View
– Services will primarily be provided across interconnected
networks operated by multimedia carriers, with a combination
of “smart” and “dumb” terminal equipment working with a
“smart” network which would control end-to-end services as
needed based on user requirements signalled to the network
– Current telephone networks will develop to support
multimedia, forming the basis of the NGN
– Much service development and provision will come from
public network operators, supported by end-to-end higherlayer services developed over open interfaces
7
Outcomes from ACIF Work 1
• “Next Generation Networks”will be packet based
• The predominant packet technology will be IP
– the major role for ATM will be to support IP, not as
the prime packet medium
• Next generation networks will have to support a
full range of services, including real-time
interactive services
• The base Internet Protocols are connectionless,
and do not guarantee secure and timely delivery
of each packet in a session (or call), especially
during periods of congestion
8
Outcomes from ACIF Work 2
• Quality of service requirements may be met by
ensuring congestion does not occur, by overdimensioning all paths over which packets pass.
• The business cases of current ISPs (the “Public
Internet”) do not make this approach likely for
end-to-end service, particularly over “thin
routes” such as those for customer access.
• The most likely path for the development of NGN
is from current public telecommunications
networks (telco networks)
9
Outcomes from ACIF Work 3
• There is a wide range of techniques to provide
QoS connections, including
– Over-dimensioning;
– Providing underlying connections, either
• Actual separated physical routes, or
• Virtual circuits
– Various QoS protocols, both IETF and proprietary
solutions
• These techniques cannot yet be reliably operated
end-to-end, but can (and do) work in current
networks
10
Outcomes from ACIF Work 4
• Within a network, “NGN” type infrastructure may
be used
– To an extend an existing network, based on
operational savings expected from packet
technologies. (NGN customer access may or may
not be provided)
– To provide a new network able to offer telephony,
interactive multimedia and Internet services;
– To provide managed data networks for corporate
customers, supporting telephony and data
• The economic advantages of individual corporate
networks supporting telephony and data are most
obvious
The common factor is telephony, aka VoIP
11
Outcomes from ACIF Work 5
• Voice (telephony) is likely to be a major (if not the
major) service on IP-based networks evolving
from current telco networks;
• In the medium term, these services are likely to
have to meet most of the current national
regulatory requirements for the telephone
service, including
– Number allocation;
– Portability;
– Division into local and long distance
• Many new services may have a telephony
component e.g. voice + video, which could start
voice to voice and later add multimedia
12
The Importance of Quality of Service
• The NGN should be able to support real-time
interactive services, including voice and also
multi-media services such as two-way video
communication
• The NGN will need to provide a continuous
Quality of Service (QoS) level greater than the
QoS currently provided from the Internet at times
of congestion
• Despite IETF work, there still seems to be no
clear way of guaranteeing end-to-end Quality of
Service across networks with current Internet
implementations.
13
How to Provide Required QoS?
• Provision of the required QoS level can be done
in a number of ways:– based on proprietary approaches developed by
network equipment vendors
– Based on agreed industry standards (primarily
protocols developed in the IETF),
– By ensuring that there is no link or router
congestion at any point in the end-to-end packet
paths in use during a call (or session), either for all
links or by choice of selected routes.
14
Interconnection and Interoperation
ACIF’s view of the current situation:
– Implementation of almost all work in the IETF on
techniques to provide assured QoS has been
directed to a single network rather than across
networks/autonomous systems
– many different approaches based on IETF
protocols and proprietary developments are now in
service in different networks.
End-to-end connectivity is not assured, and
priority standards work is needed.
15
PSTN
ISP
“Carrier Grade”
IP-based
Networks
2G Mobile
ISP
?
?
ISP
“The Internet”
(best-endeavours
network)
ISP
16
Standards needed urgently
• For the duration of a session there is a need to
establish, operate and maintain a “call” (session)
across multiple networks, with required
resources that may vary during the “call”
CPE
CPE
IP over ATM
MPLS
RSVP
Diffserv
17
Policy and Regulatory Implications
• Policy setting and regulation is not the job of standards
bodies. However, policies must be set based on the
technical reality of the services and underlying networks
that can be supplied economically. For this reason the
policy makers and regulators have had an active
involvement in the ACIF NGN Project, with a special Policy
and Regulatory Group.
• Is any similar approach being carried out internationally?
• How can the ITU help the policy setting process nationally,
regionally and globally?
18
The General View
• The techniques to support transport, control and
management across different networks are still to
be fully developed, and this work is needed
urgently.
• Techniques to support interconnection and interoperation between the advanced packet networks
now coming into service will not be available for
some time.
• This is an essential part of the “Broadband
Future”
19
Network Evolution
As you would expect. There were different views
as to how networks would evolve.
Two boundary views were considered, with the
feeling that the actual evolution would rest
between these two.
(Many of the participants in the ACIF work believe
the “Telco” approach will predominate, but this
may be because of our background in the PSTN
world!)
20
The “Internet” view
• Interconnected networks will mainly serve to
provide end-to-end connectivity, carrying packets
of data end-to-end between smart terminals, and
establishing end-to-end sessions under terminal
control as required.
• Services will be provided by interaction between
end-user equipment (e.g ICQ/MSN style voice +,
web access) in quasi-private mode.
• Legacy carrier-based services will be in decline
21
The “Telco” view
• Services will primarily be provided across
interconnected networks operated by multimedia
carriers, with a combination of “smart” and “dumb”
terminal equipment working with a “smart” network
which would control end-to-end services as needed
based on user requirements signalled to the network.
• Current telephone networks will develop to support
multimedia, forming the basis of the NGN.
• Much service development and provision will come
from public network operators, supported by end-toend higher-layer services developed over open
interfaces.
22
Interworking, Inter-operability and Anyto-Any Connectivity
• If the “Internet View” is the predominant
approach, end-to-end connectivity will depend on
the ability of the terminal equipment to work
together. The IT world has shown this will not be
assured, and it is likely that proprietary solutions
will compete and any-to-any connectivity could
be lost. Interworking to the PSTN should be
relatively easy to achieve, but access from the
PSTN to the new network is likely to be harder.
• The ability of the ITU and its partners to set
standards for future networks of this type will be
limited.
23
Interworking, Inter-operability and Anyto-Any Connectivity
• If the “Telco View” is the predominant approach,
end-to-end connectivity will depend on the
architecture(s) for the new networks, and the
interworking arrangements set in place.
• The role of the ITU and its partners in
establishing this standards framework will be of
major importance (and is the basis for much
current work which will have been described in
this Workshop).
24
Numbering, Naming and Addressing
• Future networks are likely to be based on Internet
Protocols. If the “Telco View” evolution path is
followed, we see that considerable work is
needed on numbering naming and addressing as
part of interworking arrangements across
networks for similar services.
25
• A call from one telephony terminal to another on
the NGN is also likely to be made by the user
dialling the E.164 number of the wanted terminal.
How is this to be routed?
2G Mobile
03 8111 6965
02 9959 9111
“Carrier Grade”
IP-based
Networks
GW
GW
03 5422 7503
26
Further Work Needed
We believe work is needed at the international,
regional and national level on interworking and
associated numbering, naming and addressing,
taking into account
– The current E.164 numbering arrangements
– Current requirements for services such as number
portability, and the techniques used to implement
them
– The need to be able to route calls to provide
maximum functionality (for example, for NGN to
NGN calls to remain within the NGN)
– Available tools such as the IETF ENUM protocol
(already specified for use in the 3GPPs)
27
Policy and Regulatory Implications
• Policy setting and regulation is not the job of
standards bodies. However, policies must be set
based on the technical reality of the services and
underlying networks that can be supplied
economically. For this reason the policy makers
and regulators have had an active involvement in
the ACIF NGN Project, with a special Policy and
Regulatory Group.
• Is any similar approach being carried out
internationally?
• How can the ITU help the policy setting process
nationally, regionally and globally?
28