1.  P.702 Internet Protocol version 6 - Opportunities for Service Providers and PNOs - - End-User Bandwidth on Demand - Telebit Communications A/S info@tbit.dk March 1998

2.  Partner Profile Telebit Europe’s only independent router manufacturer Design and shipment of ATM Switches and MPR Routers Technological leadership in the integration of routing and switching focusing on IPv6 and ATM University of Lancaster IPv6 centre of expertise Multi-media R&D

3.  Partner Profile

4.  UNI-C:

5.  Why IPv6 ? IPv4 Problems: Lack of class B IPv4 address space => CIDR addressing Circa 1,800 active Autonomous Systems Inject nearly 43,000 Routable Prefixes Inadequate address aggregation Ballooning BGP databases, and Router memory exhaustion Increased forwarding table look up time Ubiquitous but simplistic

6.  Why IPv6 ? Expanded Routing and Addressing capabilities –Relief of address shortage –Support of more addressing hierarchy –Addition of anycast address to give multicast efficiencies Header length fixed to 40 bytes –64 bits + 128 bits source & destination address –Easier to process in hardware and easier to compress. Improved support for Options –Use of Extension Headers and improved coding give more efficient forwarding Multihoming possibilities –Enables users to switch between providers (auto-readdressing) –Offers increased security and cost optimisation

7.  Why IPv6 ? IPv6 Mandates Auto-Address Configuration: IPv4 Configuration Process : 1) IPv4 Address 2) Default Gateway 3) Subnet Mask / Prefix Number 4) Domain Name Server and Domain Name 5) Solutions => Bootstrap (Static) & DHCP (Dynamic / Server based IPv6 Configuration Process: 1) Neighbor Discovery (stateless configuration) 2) DHCPv6 (statefull configuration)

8.  Why IPv6 ? Security: IPv4 Security Problems: 1) Denial of service attack (BGP / RIP hijacking) 2) Address spoofing 3) Use of source routing defeats address authentication IPv6 Security: 1) Mandated at the Kernel level => IPSEC 2) Authentication Header (Default to MD5) 3) Encryption ( Default to DES-CBC) 4) Security Parameter Index (Defines non-default security association) 5) Repudiation features

9.  Why IPv6 ? IPv6 QoS Advantages: QoS becoming an issue as real time services emerge: 1) Need for lower latency and jitter, but improved tolerance to lost packets 2) Less emphasis on re-transmission of lost data 3) More emphasis on timing relationships (time-stamping) 20-bit Flow Label enables identification of traffic flows Class of Service field to manage conflicts RSVP used by routers to deal with requests

10.  Why IPv6 ? Tele Denmark(DK) DSC Communications (US/DK) UNI-C (DK) European Commission (EU) France Telecom (F) Eurescom (EU) Internet For Business (UK) Microsoft (US) DFN (D) Ellemtel (S) NIST (US) University of Lancaster CSELT (IT) FAA (US) ARINC (US) AEROTHAI (THAI) Nav Canada (CAN) Politecnico di Torino EENET LATNET LITNET ARNES Customers and Partners:

11.  IPv6 - The Killer App ? Objectives: End-User Bandwidth on Demand over multiple backbone technologies (P.702) Partners: Consortia of 23 European Telcos (Sponsor) University of Lancaster (Application development) Telebit Commications A/S (Router development)

12.  P.702 - Objectives Motivators and Objectives: Most general voice/video/data services available with IP Need for user-initiated multiple service levels with QoS/CoS Integration of IP services with all backbone/trunking technologies Needed to make optimum use of network resources and facilitate growth in multimedia services. –Doesn’t ADSL promise 6 Mbps to everyone ? –But how will the backbone network carry traffic ?

13.  P.702 - How it works The Concept: © Eurescom

14.  P.702 - How it works The Demonstrator: © Eurescom

15.  P.702 - How it works 1) Customer selects video from menu, and clicks on PLAY 2) Client in host signals to the server to start video Video SelectService Select  Video 1  Default  Video 2 Bronze  Video 3  Silver Video 4  Gold  Video 5  Platinum PLAY

16.  P.702 - How it works 3) Video server starts and assigns random IPv6 flow label. This is combined with the source address => unique Flow ID 4) Client initiates RSVP sequence signalling reqd BW to server 5) Server sends RSVP PATH_MESSAGE, signalling the BW requirement to the first Decision Point (DP1). 6) PATH_MESSAGE routed using QoS Routing Table mapping BW to port no. 7) Path through network installed.

17.  P.702 - How it works QOS Routing table: Source Prefix Dest Prefix BW Interface Next Hop Max Min xxxxxxxxxxxxx64K 8K Internet:: xxxxxxxxxxxxx 2M 64KISDN:: xxxxxxxxxxxxx 8M2MATM ::

18.  P.702 - How it works 8) Client then sends reservation message back through network following the initial RSVP PATH_MESSAGE 9) Remote Decision Point then sets up connection and installs reservation 10) All subsequent packets with identical Flow ID are identified and forwarded using the Reservation Table - RSVP scaleability problems avoided by using its end-end signalling properties, rather than hop by hop reservation features - Note: Service provision under the control of policy and subscription information held by a PNO controlled NMC.

19.  P.702 - How it works

20.  P.702 - How it works

21.  P.702 - Other Technologies CISCOs tag switching /IPSILON’S IP flows: Work at an IP level with few ATM / FR / ISDN possibilities Signalling not extended to the user, so no bandwidth on demand facility Focus on priority access to finite resource - overbooking problems not resolved Do not consider ITU / ATM forum standards, thus ill suited to broader telco environments Until MPLS is ratified by the IETF, they are proprietary solutions Little opportunity for system administrator to control user access Rely on point-point soft-state connections throughout a network, and distribution of QoS information - scalability ?

22.  IPv6/P.702 - Conclusion Could revolutionise relationship between PNO & End-User Requires no specific link-layer technology Facilitates new and differentiated services: –High bandwidth on demand with existing infrastructures –Real time services (Voice over IP / Video) –Multicast services Options for spontaneous purchase of BW Could easily be deployed Standards based implementation Software developed by group available for experimentation