1 prop-031-v001: Proposal to amend APNIC IPv6 assignment and utilisation requirement policy Policy SIG 8 Sep 2005 APNIC20, Hanoi, Vietnam Geoff Huston.

Slides:



Advertisements
Similar presentations
ARIN Public Policy Meeting
Advertisements

End of the Internet Predicted! Torrent at 11. The Oracle Bones of IPv4 Some personal divination by Geoff Huston APNIC.
1 IPv6 Address Distribution Mechanisms Geoff Huston APNIC.
Transitioning to IPv6 April 15,2005 Presented By: Richard Moore PBS Enterprise Technology.
IP Addressing Introductory material.
1 Computer Communication & Networks Lecture 17 & 18 Network Layer: Logical Addressing Waleed Ejaz.
IPv4 Addresses. Internet Protocol: Which version? There are currently two versions of the Internet Protocol in use for the Internet IPv4 (IP Version 4)
Draft Policy GPP Network IP Resource Policy Advisory Council Shepherds: Scott Leibrand and Rob Seastrom.
1 Where did all those IPv6 addresses go? Geoff Huston APNIC April 2005 ARIN XV Discussion Panel Presentation.
IPv6: The Future of the Internet? July 27th, 1999 Auug.
IPv6 Addressing – Status and Policy Report Paul Wilson Director General, APNIC.
Oct 28, 2004CS573: Network Protocols and Standards1 IP: Routing and Subnetting Network Protocols and Standards Autumn
2010-8: Rework of IPv6 Assignment Criteria David Farmer ARIN XXVI.
IAB/IESG Recommendations on IPv6 Address Allocation Bob Hinden at RIPE Sept Brian Carpenter at ARIN Oct Alain Durand at APNIC Oct
IPv6 Elite Panel Addressing the IPv6 Internet Paul Wilson APNIC.
Allocations vs Announcements A comparison of RIR IPv4 Allocation Records with Global Routing Announcements Geoff Huston May 2004 (Activity supported by.
IPv4 Address Lifetime Expectancy Geoff Huston Research activity supported by APNIC The Regional Internet Registries s do not make forecasts or predictions.
IAB/IESG Recommendations on IPv6 Address Allocation Bob Hinden at RIPE Sept Brian Carpenter at ARIN Oct Alain Durand at APNIC Oct
1 The Geography and Governance of Internet Addresses Paul Wilson APNIC.
IPv6 Interim Policy Draft RIPE 42 Amsterdam, The Netherlands 1 May 2002.
IPv4 Addresses. Internet Protocol: Which version? There are currently two versions of the Internet Protocol in use for the Internet IPv4 (IP Version 4)
© 2006 Cisco Systems, Inc. All rights reserved.Cisco Public BSCI Module 8 Lessons 1 and 2 1 BSCI Module 8 Lessons 1 and 2 Introducing IPv6 and Defining.
1 IPv6 Address Space Management Report of IPv6 Registry Simulation Policy SIG 1 Sept 2004 APNIC18, Nadi, Fiji Geoff Huston.
IPv4 Address Lifetime Presented by Nurani Nimpuno, APNIC Research activity conducted by Geoff Huston and supported by APNIC.
Measuring IPv6 Deployment Geoff Huston George Michaelson
ICS 156: Lecture 2 (part 1) Today:  IP addressing  Data link protocols and ARP  Notes about lab.
IPv4 Unallocated Address Space Exhaustion Geoff Huston Chief Scientist APNIC November 2007.
A proposal to lower the IPv4 minimum allocation size and initial criteria in the AP region prop-014-v001 Policy SIG APNIC17/APRICOT 2004 Feb
SANOG-7 Internet Evolution and IPv6 Paul Wilson Geoff Huston APNIC.
Policy Proposal to amend ARIN IPv6 assignment and utilization requirements ARIN XVI Los Angeles October 2005.
Draft-vandevelde-v6ops-addcon-00.txt IPv6 Unicast Address Assignment Considerations Gunter Van de Velde (editor) Tim Chown Ciprian Popoviciu IETF 65, March.
Addressing Issues David Conrad Internet Software Consortium.
1 The How of Where Geoff Huston APNIC Some Observations on IPv6 Addresses.
Address planning. Introduction Network-Level Design Considerations Factors affecting addressing scheme Recommended practices Case studies 6/4/20162.
1 AS Consumption Patterns Geoff Huston APNIC May 2005.
IPv6 - The current reality behind the promise Tony Hain IPv6 Forum Fellow Technology Director – NAv6TF Technical Leader – Cisco Systems
1 IPv4 Address Lifetime Presented by Paul Wilson, APNIC Research activity conducted by Geoff Huston and supported by APNIC.
AS Numbers NANOG 35 Geoff Huston APNIC. Current AS Number Status.
Overview of Policy Proposals Policy SIG Wednesday 31 August 2011.
1 Application of the HD ratio to IPv4 [prop-020-v001] Policy SIG 1 Sept 2004 APNIC18, Nadi, Fiji.
APNIC 32 AMM Policy SIG Report Andy Linton Thursday 1 September 2011.
1 IANA global IPv6 allocation policy [prop-005-v002] Policy SIG 1 Sept 2004 APNIC18, Nadi, Fiji.
CSE5803 Advanced Internet Protocols and Applications (13) Introduction Existing IP (v4) was developed in late 1970’s, when computer memory was about.
Global IPv6 Address Interim Policy Draft Open Issues and Discussion Summary Address Policy SIG / 13 th APNIC Meeting Kosuke Ito Global IPv6 Interim Policy.
Policy SIG Report APNIC AGM Friday 29 August 2008 Christchurch, New Zealand 1.
TCP/IP Protocol Suite 1 Chapter 4 Objectives Upon completion you will be able to: IP Addresses: Classful Addressing Understand IPv4 addresses and classes.
1 HD Ratio for IPv4 RIPE 48 May 2004 Amsterdam. 2 Current status APNIC Informational presentation at APNIC 16 Well supported, pending presentation at.
Copyright (c) 2002 Japan Network Information Center Proposal for IPv6 Policy for Essential Infrastructure in the AP region Izumi Okutani IP Address Section.
IP Addressing Introductory material.
2002網際網路趨勢研討會IPv6 Tutorial
ECSE-6600: Internet Protocols
Where did all those IPv6 addresses go?
Addressing 2016 Geoff Huston APNIC.
IP Addressing Introductory material.
IP Addressing Introductory material.
IPv4 Addresses.
IP Addresses in 2016 Geoff Huston APNIC.
The IPv4 Consumption Model
Stephan Millet, Geoff Huston
IPv6 Address Space Management Report of IPv6 Registry Simulation
An Update on Multihoming in IPv6 Report on IETF Activity
IP Addressing Introductory material
Requiring aggregation for IPv6 subsequent allocations
IPv6 Address Space Management Report of IPv6 Registry Simulation
RIPE October 2005 Geoff Huston APNIC
IPv6 Address Management Past, Present and Future
End of the Internet Predicted!
IPv4 Address Lifetime Expectancy
Unique Local IPv6 Unicast Addresses
Removing aggregation criteria for IPv6 initial allocations
Presentation transcript:

1 prop-031-v001: Proposal to amend APNIC IPv6 assignment and utilisation requirement policy Policy SIG 8 Sep 2005 APNIC20, Hanoi, Vietnam Geoff Huston

2 The Proposal 1.Add a /56 end-site allocation point (in addition to /64 and /48) 2.Default end-site allocation for SOHO end sites to be a /56 3.Evaluation for subsequent allocations to be based on an HD-Ratio value of End-site allocation size for HD-Ratio calculation based on a /56 unit

3 Presentation 1. Motivation 2. Impact analysis 3. Implementation

4 1. Motivation Analysis of overall lifetime and deployment size of IPv6

5 Current Address Allocation Policies RIR to ISP(LIR): − Initial allocation: /32 (minimum) − Subsequent allocation : /32 (minimum) ISP(LIR) to customer: − Only 1 interface ever: /128 − Only 1 subnet ever: /64 − Everything else: /48 (minimum) ISP(LIR) to each POP: − /48

6 Address Efficiency – HD=0.8 Prefix /48 countend-site count /32 65,536 7,132 /31 131,072 12,417 /30 262,144 21,619 /29 524,288 37,641 /28 1,048,576 65,536 /27 2,097, ,105 /26 4,194, ,668 /25 8,388, ,901 /24 16,777, ,249 /23 33,554,432 1,048,576 /22 67,108,864 1,825,677 /21 134,217,728 3,178,688 /20 268,435,456 5,534,417 /19 536,870,912 9,635,980 /181,073,741,824 16,777,216

7 Squeezing in Bigger Numbers for Longer Timeframes The demand - global populations: −Households, Workplaces, Devices, Manufacturers, Public agencies −Thousands of service enterprises serving millions of end sites in commodity communications services −Addressing technology to last for at least tens of decades −Total end-site populations of tens of billions of end sites i.e. the total is order The supply – inter-domain routing −We really may be stuck with BGP −Approx 200,000 routing (RIB) entries today −A billion routing (RIB) entries looks a little too optimistic i.e. a total entry count is order 10 7 The shoe horn −Aggregation and hierarchies in the address plan

8 Longevity Shifting a technology base due to address scarcity leads to a scarcity solution, not necessarily a superior solution It would be preferable to provide for ample address supply over the entire anticipated technology lifecycle −i.e. still have ‘ample’ addresses at the end of the lifecycle Long-end IPv6 lifecycle estimate of 60 – 100 years

9 Putting it together Aggregation and hierarchies are not highly efficient addressing structures The addressing plan needs to accommodate both large and small The addressing plan needs to be simple (16 bit subnets) + (HD = 0.8) + (global populations) + ( years) =?

10 HD Ratio for Bigger Networks Prefix /48 count end-site count /21 134,217,728 3,178,688 /20 268,435,456 5,534,417 /19 536,870,912 9,635,980 /18 1,073,741,824 16,777,216 /17 2,147,483,648 29,210,830 /16 4,294,967,296 50,859,008 /15 8,589,934,592 88,550,677 /14 17,179,869, ,175,683 /13 34,359,738, ,435,456 /12 68,719,476, ,373,275 /11 137,438,953, ,744,135 /10 274,877,906,944 1,416,810,831 /9 549,755,813,888 2,466,810,934 /8 1,099,511,627,776 4,294,967,296 /7 2,199,023,255,552 7,477,972,398 /6 4,398,046,511,104 13,019,906,166 /5 8,796,093,022,208 22,668,973,294 /4 17,592,186,044,416 39,468,974,941 /3 35,184,372,088,832 68,719,476,736 /2 70,368,744,177, ,647,558,364 /1 140,737,488,355, ,318,498,661

11 Multiplying it out A possible consumption total: a simple address plan (/48s) x aggregation factor (HD = 0.8) x global populations (10**11) x 60 years time frame = 50 billion – 200 billion = /1 -- /4 range RFC 3177 (Sept 2001) estimated 178 billion global IDs with a higher HD ratio. The total “comfortable” address capacity was a /3.

12 Is this enough of a margin? /4 consumption −A total of 1/16 of the of the available IPv6 address space /1 consumption −A total of 1/2 of the available IPv6 address space Factors / Uncertainties: −Time period estimates (decades vs centuries) −Consumption models (recyclable vs one-time manufacture) −Network models (single domain vs overlays) −Network Service models (value-add-service vs commodity distribution) −Device service models (discrete devices vs ubiquitous embedding) −Population counts (human populations vs device populations) −Address Distribution models (cohesive uniform policies vs diverse supply streams) −Overall utilization efficiency models (aggregated commodity supply chains vs specialized markets)

13 If this is looking slightly uncomfortable… then we need to re-look at the basic assumptions to see where there may be some room to shift the allocation and/or architectural parameters to obtain some additional expansion space

14 Where’s the Wriggle Room? IPv6 Allocation Policies −The HD-Ratio target for address utilization −The subnet field size used for end-site allocation IPv6 Address Architecture −64 bit Interface ID Interface ID 64 bits Subnet ID 16 bits Global ID 48 bits

15 1. Varying the HD Ratio /32 / % 2.1% 51.4% 31.2% Prefix Size Utilization Efficiency

16 Comparison of prefix size distributions from V6 registry simulations

17 Observations 80% of all allocations are /31, /32 for HD ratio of 0.8 or higher −Changing the HD ratio will not impact most allocations in a steady state registry function Only 2% of all allocations are larger than a /27 −For these larger allocations the target efficiency is lifted from 4% to 25% by changing the HD Ratio from 0.8 to 0.94 Total 3 year address consumption is reduced by a factor of 10 in changing the HD ratio from 0.8 to 0.94

18 What is a “good” HD Ratio to use? Consider what is common practice in today’s network in terms of internal architecture −APNIC survey of ISPs in the region on network structure and internal levels of address hierarchy and will present the findings at APNIC 20 Define a common ‘baseline’ efficiency level rather than an average attainable level −What value would be readily achievable by large and small networks without resorting to renumbering or unacceptable internal route fragmentation? Consider overall longer term objectives −Anticipated address pool lifetime −Anticipated impact on the routing space

19 2. The Subnet Identifier field RFC 3177: The subnet field Recommendation /48 in the general case, except for very large subscribers /64 when it is known that one and only one subnet is needed by design /128 when it is absolutely known that one and only one device is connecting Motivation reduce evaluation and record-keeping workload in the address distribution function ease of renumbering the provider prefix ease of multi-homing end-site growth allows end-sites to maintain a single reverse mapping domain Allows sites to maintain a common reverse mapping zone for multiple prefixes Conformity with site-local structure (now unique locals)

20 Alternatives for subnetting Consider variable length subnetting -Allows for greater end-site address utilization efficiencies -Implies higher cost for evaluation and record keeping functions -Implies tradeoff between utilization efficiency and growth overheads -Likely strong pressure to simplify the process by adopting the maximal value of the range

21 Alternatives for subnetting Consider /56 SOHO default size −Maintain /128 and /64 allocation points, and /48 for compound enterprise end-sites −Processing and record-keeping overheads are a consideration here −End-site growth models for SOHO are not looking at extensive subnetting of a single provider realm −Renumbering workload is unaltered −Multi-homing is not looking at prefix rewriting −Fixed points maintains reverse mapping zone functions −Allow for overall 6 – 7 bits of reduced total address consumption

22 3. The Interface Identifier This 64 bit identifier is now well embedded in the address architecture and protocol specification for IPv6 Considerations for change here have extensive implications in terms of overlayed services of auto-configuration and discovery functions

23 Where’s the Wriggle Room? The HD ratio −If using HD = 0.8 consumes 1 block of address space −Using HD = 0.87 consumes 1/2 as much space −Using HD = 0.94 consumes 1/10 as much space −i.e. moving to a higher HD ratio will recover up to 3 bits here The subnet field −/56 SOHO default subnet size may alter cumulative total by bits /10 -- /17 total consumption given original demand estimates Is this sufficient margin for error / uncertainty in the initial assumptions about the deployment lifetime for IPv6?

24 Now or Later? Do we want to create early adopter rewards and late adopter restrictions? Should we attempt to operate with more stable policies across the anticipated technology lifecycle? RFC3177 Therefore, if the analysis does one day turn out to be wrong, our successors will still have the option of imposing much more restrictive allocation policies on the remaining 85%.

25 2. Impact Analysis üGreater confidence in address availability across anticipated technology lifecycle üFairness of allocations across the anticipated technology lifecycle ûHigher overheads in profiling end site allocations ûPotential renumbering in end site growth cases ûHigher overheads in network address planning for HD ratio value of 0.94

26 3. Implementation Part of a global coordination effort across all RIRs −Possible review of policy proposal following consideration from other RIR forums