TOC 
Individual SubmissionG. Huston, Ed.
Internet-DraftIAB
Expires: January 30, 2005August 2004

Considerations on the IPv6 Host density Metric

draft-iab-ipv6-host-density-metric-000.txt

Status of this Memo

By submitting this Internet-Draft, I certify that any applicable patent or other IPR claims of which I am aware have been disclosed, and any of which I become aware will be disclosed, in accordance with RFC 3668.

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts.

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."

The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt.

The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html.

This Internet-Draft will expire on January 30, 2005.

Copyright Notice

Copyright (C) The Internet Society (2004). All Rights Reserved.

Abstract

This memo provides an analysis of the Host Density metric as currently used to guide registry allocations of IPv6 unicast address blocks. This document contrasts the address efficiency as currently adopted in the allocation of IPv4 network addresses and that used by the IPv6 protocol. It is noted that for large allocations there are very significant variations in the target efficiency metric between the two approaches. The memo notes that the IPv6 address assignment efficiency metric would benefit from a detailed technical review, particularly relating to large scale deployments of public infrastructure.

Document Revision Notes

The following changes have been made to the draft:

draft-iab-ipv6-host-density-metric-000:

Initial draft




Table of Contents

1.  Introduction
2.  IPv6 Address Structure
3.  The Host Density Ratio
4.  The Role of an Address Efficiency Metric
5.  Network Structure and Address Efficiency Metric
6.  Considerations
§.  Normative References
§  Author's Address
A.  Comparison Tables
§  Intellectual Property and Copyright Statements




 TOC 

1. Introduction

Metrics of address assignment efficiency are used in the context of the public Internet as a part of the address allocation function. Though the use of an address assignment efficiency metric individual networks can be compared to a target model in an objective fashion. The common use of this metric is to form part of the justification of an address allocation request, demonstrating that the network has met the target address efficiency metric and that the allocation of a further address block is justified.

Public IP networks have significant differences in purpose, structure, size and technology. Attempting to impose a single metric across this very diverse environment is a challenging task. Any address assignment efficiency metric has to represent a balance between stating an achievable metric for any competently designed and operated service platform, while not specifying a metric that allows for an address usage rate that imperils the protocol's longer term viability. There are a number of views relating to address assignment efficiency, both in terms of theoretic analyses of assignment efficiency and in terms of practical targets that are part of current address assignment practices in today's Internet.

This document contrasts the address efficiency as currently adopted in the allocation of IPv4 network addresses and that used by the IPv6 protocol. It is noted that for large allocations there are very significant variations in the target efficiency metric.

The conclusion drawn here is that the IPv6 address assignment efficiency metric would benefit from a detailed technical review, particularly relating to large scale deployments of public infrastructure. approaches as required.



 TOC 

2. IPv6 Address Structure

Before looking at address allocation efficiency metrics it is appropriate to summarize the address structure for IPv6 global unicast addresses.

The general format for IPv6 global unicast addresses is defined in RFC3513 [1]Hinden, R. and S. Deering, Internet Protocol Version 6 (IPv6) Addressing Architecture, April 2003. as follows (Figure 1).


   |         64 - m bits    |   m bits  |       64 bits              |
   +------------------------+-----------+----------------------------+
   | global routing prefix  | subnet ID |       interface ID         |
   +------------------------+-----------+----------------------------+

     

Figure 1. IPv6 Address Structure

Furthermore, within the current policy framework for allocation of IPv6 addresses in the context of the public Internet, the value for 'm' in the figure above is defined as 16 bits, such that the global routing prefix is 48 bits in length, the per-customer subnet ID is 16 bits in length and the interface ID is 64 bits in length.

In relating this address structure to the address allocation function, the efficiency metric is not intended to refer to the 128 bit IPv6 address, nor the 64 bit routing prefix, but is limited to the 48 bit global routing prefix. This allocation model assumes that each customer is allocated a minimum of a /48 address block, and, given that this block allows 2**16 possible subnets, it is also considered that a /48 allocation will be used in the overall majority of cases of end-customer address assignment.

The following discussion makes the assumption that the address allocation unit in IPv6 is an address prefix of 48 bits in length, and the address assignment efficiency in this context is the efficiency of assignment of /48 address allocation units.



 TOC 

3. The Host Density Ratio

The "Host Density Ratio" is first described in RFC 1715 [2]Huitema, C., The H Ratio for Address Assignment Efficiency, November 1994., and subsequently updated in RFC3194 [3]Durand, A. and C. Huitema, The H-Density Ratio for Address Assignment Efficiency An Update on the H ratio, November 2001..

The "H Ratio", as defined in RFC1715, is:

                  log (number of objects)
              H = -----------------------
                     available bits
     

The rationale for this approach was to mathematically model hierarchical forms of address assignment, where the number of levels of hierarchy increases with larger address blocks. Such hierarchies are common in IP network deployments, where, for example, address blocks may be assigned per region, and each regional block further divided into an address block per Point of Presence (POP), and from each POP address block there is a per customer address allocation. If the network is to allow for various forms of dynamic change, and if there is a desire to avoid renumbering within the network and renumbering customers' networks, then due allowance must be made for change at each level of the network hierarchy.

The RFC draws on a number of examples to support the assertion that this metric reflected a useful measure of address assignment efficiency, and furthermore that the optimal point for such a utilization efficiency metric lies between 0.14 and 0.26

[side note - to be removed from final]

Interestingly the table in RFC1715, indicating a range of addressed objects for a 64 bit address range was given as between 9 E+8 and 4 E+16, while 128 bits yielded values of 8 E+17 through to 2 E+33. This data was used to support the argument that 64 bits of address space was insufficient. Given that we are now operating in a mode where the IPv6 address unit is somewhere between 48 and 64 bits in effective length (as distinct from 128), there is a somewhat ironic twist in this original observation.

This metric has a maximal value of log base 10 of 2, or 0.30103.

The metric was 'normalized' in RFC3194, and a new metric, the "HD-Ratio" was introduced, with the definition:

                    log(number of allocated objects)
              HD = ------------------------------------------
                   log(maximum number of allocatable objects)
     

HD values are directly proportional to the H ratio, and the values of the ratio range from 0 to 1. The RFC then applied this HD-Ratio metric to the examples given in RFC 1715, and on the basis of these examples, postulated that HD-Ratios of 0.85 or higher forced the network into some form of renumbering, while 0.80 or lower was considered to be an acceptable network efficiency metric.

The HD ratio is referenced within the IPv6 address allocation policies used by the Regional Internet Registries, and the policy documents specify that an HD-Ratio metric of 0.8 is an acceptable objective in terms of address assignment efficiency for an IPv6 network.

By contrast, the generally used address efficiency metric for IPv4 is the simple ratio of the number of allocated (or addressed) objects to the maximum number of allocatable objects. For IPv4 the commonly applied value for this ratio is 0.8 (or 80%).

A comparison of these two metrics is given in Table 1 of Attachment A.



 TOC 

4. The Role of an Address Efficiency Metric

The role of the address efficiency metric is to provide objective metrics than can be used by both the allocation entity and the recipient to determine whether an allocation is warranted, and provide some indication of the size of the allocation that should be undertaken. The metric provides a target address utilization levels that indicates at what point a network's address resource may be considered to be "fully utilized".

The objective here is to allow the network service provider to deploy addresses across both network infrastructure and to customers in a manner that does not entail periodic renumbering, and in a manner that allows both the internal routing system and inter-domain routing system to operate without excessive fragmentation of the routing space. This entails use of an addressing plan where at each level of structure within the network there is a pool of address blocks that allows expansion of the network at that structure level without requiring renumbering of the remainder of the network.

It is recognized that an address utilization efficiency metric of 100% is unrealistic in any scenario. Within a typical address structure that address space is exhausted not when all address resources have been used, but when one element within the structure has exhausted its pool, and augmentation of this pool by drawing from the pools of other elements would entail extensive renumbering. While it is not possible to provide a definitive threshold of what overall efficiency level is obtainable in all IP networks, experience with IPv4 network deployments suggests that it is reasonable to observe that at any particular level within a hierarchically structured address deployment plan an efficiency level of between 60% to 80% is an achievable metric in the general case.

This IPv4 efficiency threshold is significantly greater than that observed in the examples provided in conjunction with the HD-Ratio. It is noted that the examples used in the HD-Ratio are drawn from, among other sources, the PSTN, and this warrants some additional examination. There are a number of differences between public IP network deployments and PSTN deployments that may account for this difference. IP addresses are deployed on a per-provider basis with an alignment to network topology. PSTN addresses are, on the whole, deployed using a geo- political distribution system of "call areas" that share a common number prefix. Within each call area sufficient number blocks from the number prefix must be available to allow each operator to draw their own number block from the area pool. Within the IP environment service providers do not draw address blocks from a common geographic number pool, but receive address blocks from the regional internet registry on a 'whole of network' basis. Internally within an IP network the interior routing protocol, if uniformly deployed, admits a hierarchical network structure that is only two levels deep. Additional levels of routing hierarchy may be obtained using various forms of route confederations, but this is not a common deployment technique. The most common form of network structure used in large IP networks is a three-level structure using regions, individual Points of Presence (POPs), and end-customers.

It should also be noted that large scale IP deployments typically use a relatively flat routing hierarchy. In order to improve the dynamic performance of the interior routing protocol the number of routes carried in the interior routing protocol is commonly restricted to the routes corresponding to next hop destinations for iBGP routes, and customer routes are carried in the iBGP domain.



 TOC 

5. Network Structure and Address Efficiency Metric

An address efficiency metric can be expressed using the number of levels of structure (n) and the efficiency achieved at each level (e). If the same efficiency threshold is applied at each level of structure the resultant efficiency threshold is n**e. This then allows us to make some additional observations about the HD-Ratio values. Table 2 of Appendix A indicates the number of levels of structure that are implied by a given HD-Ratio value of 0.8 for each address allocation block size, assuming a fixed efficiency level at all levels of the structure. The implication is that for large address blocks the HD-Ratio assumes a large number of elements in the hierarchical structure, or a very low level of address efficiency at the lower levels. In the case of IP network deployments this is not commonly the case.

As noted above the most common form of structure used in IP networks is a three level structure. For larger networks a four level structure may be used, where the network is the union of a number of distinct operating entities, each of which use a three level internal structure.

Table 3 of Attachment A shows an example of address efficiency outcomes using a per-level efficiency metric of 0.75 and a progressively deeper network structure as the address block expands, up to a 5 level structure for the larger blocks ("limited levels").

It is illustrative to compare these metrics for a larger network deployment. If, for example, the network is designed to encompass 8 million end customers, then the following table indicates the associated allocation size as determined by the address efficiency metric.

   Allocation:  8M Customers  Allocation    Relative Ratio

   100% Allocation Efficiency   /25               1
   80% Efficiency (IPv4)        /24               2
   HD-Ratio                     /19              64
   75% with Limited Levels      /23               4
     

It is noted that the HD-Ratio produces a significantly lower efficiency level than the other two metrics, and the limited level model appears to point to a more realistic value for an efficiency value for networks of this scale.



 TOC 

6. Considerations

The HD ratio as a model of network address utilization efficiency produces very low efficiency outcomes for networks spanning of the order of 10**6 end customers and larger.

The HD-Ratio makes the assumption that as the address allocation block increases in size the network within which the addresses will be deployed adds additional levels of hierarchical structure. This increasing depth of hierarchical structure is not a common feature of public IP network deployments.

The fixed efficiency model uses the assumption that as the allocation block becomes larger the network structure remains at a fixed level of levels, or if the number of levels is increased, then efficiency achieved at each level increases significantly. There is little evidence to suggest that increasing number of levels in a network hierarchy increases the efficiency at each level.

It is evident that neither of these models accurately encompass IP network infrastructure models and the associated requirements of address deployment. The fixed efficiency model places an excessive burden on the network operator to achieve very high levels of utilization at each level in the network hierarchy, leading to either customer renumbering or deployment of NAT to meet the target efficiency value in a hierarchically structure network. The HD-Ratio model specifies a very low address efficiency target, and while this places no particular stress on network architects in terms of forced renumbering, there is the concern that this represents an extravagant use of address resources. If the objective of IPv6 is to encompass a number of decades of deployment, and span a public network that ultimately encompasses many billions of end customers, then there is legitimate cause for concern that the HD-Ratio may be setting too conservative a target for address efficiency.

It is recommended that further study of address efficiency metrics and the relationship between network structure and address efficiency models considered as part of such a study.

This document has also noted the choice of a fixed length of 16 bits for the subnet ID in the IPv6 unicast address architecture. While this choice has been used in the block of unicast address space spanned by the IPv6 address prefix 2000::/3, it should not be assumed by vendors or network operators that this particular subnet scheme will be used for other unicast address blocks. Prior to the opening of further unicast address blocks for end user assignment it is considered to be a useful exercise to evaluate the effectiveness of this fixed length subnet scheme, and compare it to an subnet scheme with a variable length and a smaller minimum value.



 TOC 

7 Normative References

[1] Hinden, R. and S. Deering, "Internet Protocol Version 6 (IPv6) Addressing Architecture", RFC 3513, April 2003.
[2] Huitema, C., "The H Ratio for Address Assignment Efficiency", RFC 1715, November 1994.
[3] Durand, A. and C. Huitema, "The H-Density Ratio for Address Assignment Efficiency An Update on the H ratio", RFC 3194, November 2001.


 TOC 

Author's Address

  Geoff Huston (editor)
  Internet Architecture Board
EMail:  execd@iab.org
URI:  http://www.iab.org


 TOC 

Appendix A. Comparison Tables

The first table compares the threshold number of /48 end user allocations that would be performed for a given assigned address block in order to consider that the utilization has achieved its threshold utilization level.


Fixed Efficiency Value  0.8
HD-Ratio Value          0.8


Prefix      Size       Fixed  Eff.       HD-Ratio Eff.  Ratio
                  Allocation     %                   %
                       Ratio

/48            1            1 100%              1 100%      1
/47            2            1  50%              1  50%      1
/46            4            3  75%              3  75%      1
/45            8            6  75%              5  62%      1
/44           16           12  75%              9  56%      1
/43           32           25  78%             16  50%      1
/42           64           51  80%             27  42%      1
/41          128          102  80%             48  37%      2
/40          256          204  80%             84  33%      2
/39          512          409  80%            147  29%      2
/38         1024          819  80%            256  25%      3
/37         2048         1638  80%            445  22%      3
/36         4096         3276  80%            776  19%      4
/35         8192         6553  80%           1351  16%      4
/34        16384        13107  80%           2352  14%      5
/33        32768        26214  80%           4096  12%      6
/32        65536        52428  80%           7131  11%      7
/31       131072       104857  80%          12416   9%      8
/30       262144       209715  80%          21618   8%      9
/29       524288       419430  80%          37640   7%     11
/28      1048576       838860  80%          65536   6%     12
/27      2097152      1677721  80%         114104   5%     14
/26      4194304      3355443  80%         198668   5%     16
/25      8388608      6710886  80%         345901   4%     19
/24     16777216     13421772  80%         602248   3%     22
/23     33554432     26843545  80%        1048576   3%     25
/22     67108864     53687091  80%        1825676   3%     29
/21    134217728    107374182  80%        3178688   2%     33
/20    268435456    214748364  80%        5534417   2%     38
/19    536870912    429496729  80%        9635980   2%     44
/18   1073741824    858993459  80%       16777216   2%     51
/17   2147483648   1717986918  80%       29210829   1%     58
/16   4294967296   3435973836  80%       50859008   1%     67
/15   8589934592   6871947673  80%       88550676   1%     77
/14  17179869184  13743895347  80%      154175683   1%     89
/13  34359738368  27487790694  80%      268435456   1%    102
/12  68719476736  54975581388  80%      467373274   1%    117
/11  1.37439E+11  1.09951E+11  80%      813744135   1%    135
/10  2.74878E+11  2.19902E+11  80%     1416810830   1%    155
/9   5.49756E+11  4.39805E+11  80%     2466810933   0%    178
/8   1.09951E+12  8.79609E+11  80%     4294967296   0%    204
/7   2.19902E+12  1.75922E+12  80%     7477972397   0%    235
/6   4.39805E+12  3.51844E+12  80%    13019906166   0%    270
/5   8.79609E+12  7.03687E+12  80%    22668973294   0%    310

     

Table 1 Comparison of Fixed Efficiency threshold vs HD-Ratio Threshold

One possible assumption behind the HD ratio is that the inefficiencies that are a consequence of large scale deployments are an outcome of increased number of levels of hierarchical structure within the network. The following table calculates the depth of the hierarchy in oder to achieve a 0.8 HD ratio, assuming a 0.8 utilization efficiency at each level in the hierarchy.



HD-Ratio 0.8

Prefix      Size    HD-Ratio  Structure Levels

/48            1            1     1
/47            2            1     1
/46            4            3     1
/45            8            5     2
/44           16            9     2
/43           32           16     3
/42           64           27     3
/41          128           48     4
/40          256           84     4
/39          512          147     5
/38         1024          256     5
/37         2048          445     6
/36         4096          776     6
/35         8192         1351     7
/34        16384         2352     7
/33        32768         4096     8
/32        65536         7131     8
/31       131072        12416     9
/30       262144        21618     9
/29       524288        37640    10
/28      1048576        65536    10
/27      2097152       114104    11
/26      4194304       198668    11
/25      8388608       345901    12
/24     16777216       602248    12
/23     33554432      1048576    13
/22     67108864      1825676    13
/21    134217728      3178688    14
/20    268435456      5534417    14
/19    536870912      9635980    14
/18   1073741824     16777216    15
/17   2147483648     29210829    15
/16   4294967296     50859008    16
/15   8589934592     88550676    16
/14  17179869184    154175683    17
/13  34359738368    268435456    17
/12  68719476736    467373274    18
/11  1.37439E+11    813744135    18
/10  2.74878E+11   1416810830    19
/9   5.49756E+11   2466810933    19
/8   1.09951E+12   4294967296    20
/7   2.19902E+12   7477972397    20
/6   4.39805E+12  13019906166    21
/5   8.79609E+12  22668973294    21

     

Table 2 Number of Structure Levels assumed by HD-Ratio

An alternative approach is to use a model of network deployment where the number of levels of hierarchy increases at a lower rate than that indicated in a 0.8 HD ratio model. One such model is indicated in the following table.



Per-Level Target Efficiency: 0.75

Prefix      Size   Levels  Allocations  Efficiency

/48            1       1             1   100%
/47            2       2             1    50%
/46            4       2             2    50%
/45            8       2             4    50%
/44           16       2             9    56%
/43           32       3            13    41%
/42           64       3            27    42%
/41          128       3            54    42%
/40          256       3           108    42%
/39          512       3           216    42%
/38         1024       3           432    42%
/37         2048       3           864    42%
/36         4096       3          1728    42%
/35         8192       3          3456    42%
/34        16384       3          6912    42%
/33        32768       3         13824    42%
/32        65536       3         27648    42%
/31       131072       3         55296    42%
/30       262144       3        110592    42%
/29       524288       4        165888    32%
/28      1048576       4        331776    32%
/27      2097152       4        663552    32%
/26      4194304       4       1327104    32%
/25      8388608       4       2654208    32%
/24     16777216       4       5308416    32%
/23     33554432       4      10616832    32%
/22     67108864       5      15925248    24%
/21    134217728       5      31850496    24%
/20    268435456       5      63700992    24%
/19    536870912       5     127401984    24%
/18   1073741824       5     254803968    24%
/17   2147483648       5     509607936    24%
/16   4294967296       5    1019215872    24%
/15   8589934592       5    2038431744    24%
/14  17179869184       5    4076863488    24%
/13  34359738368       5    8153726976    24%
/12  68719476736       5   16307453952    24%
/11  1.37439E+11       5   32614907904    24%
/10  2.74878E+11       5   65229815808    24%
/9   5.49756E+11       5   1.30460E+11    24%
/8   1.09951E+12       5   2.60919E+11    24%
/7   2.19902E+12       5   5.21839E+11    24%
/6   4.39805E+12       5   1.04368E+12    24%
/5   8.79609E+12       5   2.08735E+12    24%

     

Table 3: Limited Levels of Structure



 TOC 

Intellectual Property Statement

Disclaimer of Validity

Copyright Statement

Acknowledgment