Executive Council Meeting - Monday 9th July 2007
Draft Minutes
A special meeting of the APNIC EC was held at the Royal Plaza on Scotts Hotel at Singapore on the 9th July. Randy Bush, the co-Chair of the APNIC Fees Working Group was invited to attend this meeting, as well as consultants from KPMG who have been engaged to provide advice to the APNIC EC on APNIC fee structures.
Present
- Akinori Maemura (Chair)
- Che-Hoo Cheng (EC)
- Kuo-Wei Wu (EC)
- Kusumba Sridhar (EC)
- Ming-Chen Liang (EC)
- Wei Mao (EC)
- Vinh Ngo (EC)
- Randy Bush (co-chair, APNIC Fees WG)
- Paul Wilson (Director-General)
- Geoff Huston (APNIC Secretariat)
- Irene Chen (APNIC Secretariat)
- Connie Chan (APNIC Secretariat)
- Michael Hiller (KPMG)
- Bert Wu (KPMG)
Introductory Comments
Paul Wilson opened the meeting with introductory remarks, noting that the general issue of APNIC fee structure is the topic of the meeting. This meeting has come about as an outcome of a previous EC discussion to try and bring this topic to a conclusion by the end of 2007. Paul noted that the situation is one that has not made significant progress in the past year or so, and the previous time we were in this situation (when considering the APNIC secretariat location) we used the assistance of a third party neutral body to analyse the options. The same approach is being used here, and KPMG has been engaged to discuss with us what the issues are, and to investigate what directions may offer some form of resolution. The overall intent is to use KPMG's assistance in making what appears to be a difficult decision for APNIC.
Akinori Maemura commented that this is a topic that has been active within APNIC for several years and we have not made substantial progress. The other big issue at hand now is IPv4 address depletion and the probability that the APNIC business model will be changing as a result. We need to take some big steps to move forward here, including the issues of IPv4 exhaustion. This meeting offers us the opportunity to set a direction for some progress with the assistance of KPMG.
Michael Hiller noted that KPMG have already circulated material to APNIC meeting, including a comparative analysis, membership survey and situation analysis. The objective from this meeting is to identify those issues that are seen as critical by the EC and to establish the general direction of approaches that the EC would see as productive to investigate.
KPMG Presentation
Fee structure has been a topic of considerable debate within APNIC over the years. Many approaches proposed and debated.
This study is using a 2 part approach of initial analysis and consultation, followed by option analysis in further detail.
Membership information
Membership growth in Very Small and Associate membership shows a compound growth 15%, but larger category membership growth is less than 10%
Compare membership growth, fee growth and operating financial outcomes. APNIC is at a critical point, as the current financial structure is not sustainable any longer.
Member survey
There have been 4 member surveys conducted so far, with the most recent one completed in early 2007. Results from this recent membership survey were presented at the APNIC member meeting in March 2007.
The activities with the highest positive feedback were APNIC's role in infrastructure support, the reporting and analysis of Internet activities, support for developmental activities in the region, database, quality, reliability and useability, efficient and effective service delivery. Overall member satisfaction is high here, and there is an expectation of continued excellence in service delivery by the members
It was noted a need for improvement in support for the policy process, resource allocation process, clarity of roles of various related bodies (NRO and ASO), accessibility and affordability in the open policy meetings. There is an expectation from the membership that APNIC services should be improved in these areas.
The survey asked for member views on allocation by APNIC of financial resources to various activity areas - services (including allocation process streamlining, training and education expansion), communications (accessibility, representation and web access) and technical (research and development, infrastructure and resource certification). These are current expectations on the part of the membership and again pose the issues of managing the secretariat's resources and available budgets to encompass these members' expectations regarding service delivery.
It was noted that larger members have a more positive perception of APNIC services,
Discussion about the "satisfaction" result vs category, and noted that there is a discrepancy here in that while APNIC is able to deliver services to high expertise customers, the service is not so readily delivered to smaller customers, and there are issues relating to the accessibility and complexity of the APNIC delivered services to these smaller customers.
There was a request relating to the survey of how many results here are driven by the responses from the larger members - what is the breakdown of the survey results if you take out the larger member categories?
The meeting discussed what would be the 'desired' profile of the member survey? Does this point to an area of potential improvement here that APNIC could concentrate on? What is the desire outcome when looking at APNIC as a whole? Are we happy with this outcome, or are we interested in improving this profile.
A question was raised as to survey participation and KPMG responded that this was a high survey participation rate for surveys of this sort, and the outcomes are reasonably illustrative of the general member perceptions
A question was raised about budgetary process and whether there was the option of beneficiary-based budget analysis. Budget has been largely based to date on prioritization of activities to fit within the budgeted expectation of available funds, rather than setting membership fees and related APNIC charges to match the proposed budgeted objectives.
It was noted by the Fees WG co-chair (Randy Bush) that the focal issue here was all about APNIC fees, not the budget determination process. APNIC is about in the middle of the RIRs in regard to scale of operation. There is the widespread view, supported by the member survey, that AP{NIC is generally doing the right thing in terms of service delivery and support activities, and the question for this meeting is about sustainable funding models, rather than a question of "what" to fund. This is not an exercise of activity stripping.
The comment was made that in terms of fees and service activities there is a need to determine the balance between smaller and larger APNIC members.
The membership survey also exposed the perspective that longer standing members are more comfortable / happier with service delivery from APNIC
Financial performance indicators 2001 - 2006
KMPG analysis of APNIC's finances noted a trend in recent years of APNIC's expense growth out-pacing revenue growth and the foreign exchange movement has been a significant contribution here, in addition to the increasing membership base and activity levels. The 2007 APNIC budget is projecting a loss of some USD 500K. The foreign exchange contribution is significant because of the income in USD currency and expenses incurred predominately in AUD currency. Over the past 6 years revenue growth (expressed in USD) is around 6% and total expense growth is 22%, taking out foreign exchange contributions. The major revenue source remains the membership fees, with per-allocation fees, resource application fees and interest income being the secondary sources. Salary (human resource) is a major expense, at around 43% of total expenses. The majority of these expenses are paid by APNIC in AUD.
There are serious issues around stability and risk of APNIC's income streams. Membership funds have increased, but as a total proportion of the APNIC income, this income steam is dropping and in proportionate terms other income streams are rising at a higher rate.
The highest relative category of income in membership in revenue terms is in the small members, and the question was raised - where should APNIC concentrate its service delivery in order to improve its position in membership satisfaction? Also it was noted that this breakdown does not include the per-allocation fee.
Comparative analysis with other RIRs
It was noted that APNIC is in a median position with respect to overall size, etc, but APNIC has the lowest income / revenue ratio. This implies that APNIC is operating with slimmer margins in terms of financial contingency provision than any other RIR at present.
Kuo Wei Wu asserted that the EC understands the situation and wants to take the approach of looking at the other RIRs and making the decision as to whether we want to continue in this way or look at "another way" that may be modelled on another RIR in that case.
Change Management Considerations
KPMG undertakes a considerable amount of work in change management, and there are elements of change management that need to be considered. These include the generation of a compelling case for change, a clear vision and a defined plan, which, if lacking contribute to outcomes of inertia, confusion and diffusion of effort. Then the change process requires resources and organizational capacity, the lack of which has an outcome in frustration and fatigue. Then motivation and community feedback is required to maintain the pace of change and dispel doubts as to the adopted direction of change.
APNIC, in the fee discussion process, appears to be in the initial case of agreeing that we need a compelling case, but lacking a clear vision or a defined plan at this stage.
Questions to the EC
KPMG requested that each member of the APNIC EC consider this issue by phrasing responses to following questions:
- Governance - the role of the EC in setting fees, as opposed to the personal role as a member; role of the DG in fee setting
- APNIC's ability to respond to the member survey - need to balance member survey outcomes and financial considerations
- The components in a new fee model - suggestions solicited, such as the basis for the calculation of the fees (e.g. resource usage), suggestions about the update of the fee structure (eg in terms of change with CPI or forex or...
- Criteria which should be used to assess each fee structure option
- The desirable decision process for the EC and membership
- Risks
- Other comments
Discussion: how does the fee model change the service delivery focus? Noted that RIPE's focus on revenue from new members has meant that RIPE has made it very easy to join and be serviced, and their business model is strongly based on such growth and new members. It was observed that as you structure the revenue sources you adjust the focus of the organization on those sources, so there is a real feedback loop between the fee structure and the service profile that needs to be considered here in the overall discussion
Reponses to KPMG's questions
Kuo-Wei Wu:
- Suggest that we discuss about the general mechanism, then discuss the next step within the framework of a particular mechanism.
- We have issues of exchange rate variations, which is different to the other RIRs. In this meeting we should resolve this exchange issue in the long term
- Consider APNIC as a whole rather than large or small members or similar. General perception of an eagerness to make APNIC financially stable, but not too much time to encourage different voices as this will lose focus on the APNIC issues. If we have a good mechanism for service delivery than this will be good for all. The EC is a good platform as it has NIR and small member individuals on the Council. The EC has a process already and the EC should develop this and take it to the membership
- We should bring this fee structure discussion to a halt - it has been 3 - 4 years and this is too long and we should come up with a solution now and come up with a result in New Delhi
- [What is the process?] We already have a mechanism - do all the members agree? Do we want to change the relative balance between small and large? We (EC) should agree on this today
Kusumba Sridhar:
- Either you satisfy the larger membership or the smaller membership and this is seen as an exclusive choice
- NIR reps on the EC or not is not a question - the EC looks at a "global perspective"
- The issue here is change, we must change here. Either we change the service delivery mechanism, or change the nature of APNIC itself. There is a need to understand this change process and be prepared to look at the entirely of APNIC and its role.
- Suggest a short term approach and a longer term approach here as being distinct responses here
- Noted that there is an invitation in the fees WG inviting papers and sees this as a non-terminating process
- Can we discuss fees in a split manner? Should we split fees against membership category? Can we split the membership into 2 different groups and structure the fees in two parts?
- 2008 is likely to be worse that 2007 given the exchange rate forecast and the growing deficit
- Suggest to stop the member contribution phase and move to a conclusion
- If this is not possible let's have a short term proposal if this not concluded.
- Each new paper in the fees group is a new perspective from a new angle - there has been no real analysis of these proposals
- This (fees WG) appears to be a non-terminating process.
- Wants to phase in the approach
Ming-Chen Liang:
- Nothing will be universally acceptable.
- Maybe a short term approach is a less than dramatic impact from the current situation and also look at the longer term
- Consider a small incremental change to the model in the short term
- There are impacts from the budget here. Most of the APNIC income is from existing members.
- In the longer term, APNIC is a service oriented organization and charges should be aligned to service delivery.
- Also wants to advocate a realignment of APNIC with NIRs as the major mode of operation
- [Criteria for evaluation of the option?] APNIC is a service desk. We should look at this as a service proposition here.
Che-Hoo Cheng:
- Wants better sustainability of APNIC financials
- The current model was developed in entirely different circumstances over a decade ago
- Che-Hoo had investigated the history of the fee debate over the year, and noted a history of 'quick fixes' extending over years.
- It may not be possible to establish a permanent solution - the next 3 - 5 years is very uncertain
- We need a good model here that is 'fairer' in terms of the interests of the smaller members of APNIC. The motivation here is to improve the service operation of APNIC for smaller members.
- Wants a more consistent model for non-NIRs and smaller members
- Understands the rationale for "discounts" for NIRs
- Would like to see the use of a currency for fees that is more stable here in this region
- More ability for people to decide whether they need a NIR or not - it is not clear at the moment
- Criteria: fairness, consistency, stability
- Process: we can always put up one or more proposals for the members to decide - run the risk that all the proposals to be rejected by the members. One way is to use KPMG (supports this model) as the proposal comes from a disinterested third party there is the process for the EC to decide to put forward a proposal to the members
- There has been too much discussion already - we need to come to a conclusion here!
[Michael: maybe there is a need for the EC to circulate its ideas as it evaluates options]
Wei Mao:
[presentation]
- EC to put forward principles for fee, KPMG to implement these as a practical plan, DG to provide data to KPMG, the EC to review KPMG output and DG to pass this to the APNIC community
- Maintain efficient operation of APNIC, financial sustainability under change of membership profile and exchange rates, pass on efficiency of operation to members, charge based on service delivery (as distinct from resource holdings)
- Propose a shift to service-based charges. Advocates NIR service delivery.
- Propose a "budget committee" to review EC budget, and the member fee based on the adoption of the budget. i.e. set fees to match projected expenditure
- Propose votes to fees relationship
- Shows a graph illustrating fees paid per member for each RIR (including membership and per address fees) [DG points out that the graph omits members of NIRs, and therefore does not provide an accurate comparison with other RIRs ]
Vinh Ngo:
- We are running a business and trying to sustain this business.
- We have developing communities and developed communities. We have NIRs and small members. There are factors of what can be controlled and what are uncontrollable.
- Focus on existing members and provide services that fix revenue for services.
- Uncontrollable factors are exchange rates and new memberships
- the member survey is positive with noted room for improvement - but revenue is way behind. This trend will see APNIC go bankrupt. Therefore revenue must meet expenses
- Membership fees is all we have (70% membership, 20% IP allocation)
- KPMG is being consulted in terms of expertise of modelling fair and reasonable fees for our members
- We must ensure that the structure will recoup the cost of running APNIC and ensure sustainability of APNIC
Maemura Akinori:
- Service-based fee structure requires a balance
- Align fees to cost of the service
- Smaller members are more resource intensive in terms of resource analysis
- Charge by AUD as a means of higher levels of stability in terms of revenue vs expense - the payer takes the current exchange variation
- Transparency in budget planning is important
- Criteria: fairness and transparency
Randy Bush
- Largely agrees with Vinh - this is just business
- When income is less than expense than you need to adjust income
- The long term is not understood - we can't plan for it
- We need to look at a short term fix and in the long term we need to plan
- The NIRs are a major factor here
- We need to open up minds of members as to alternatives here
- The NIRs constrain what parts of our fee structure we can borrow from other RIRs
Geoff Huston:
- Use conventional business management tools. Firstly phrase the activities and budgetary objectives, then set the fees to meet the objectives.
- There is no margin here - we can't borrow money if APNIC consumes all its reserves (since we have no other equity/assets).
- What's the business today?
- What's the business tomorrow?
- Who are the beneficiaries?
- Who pays?
- A pure service transaction model as the basis for revenue is fundamentally tough as it does not align to the basic cost and benefit proposition. Many of APNIC activities which are seen as valuable by the membership are mutual in nature, capacity building through education and training, deployment of infrastructure elements such as root servers and reverse DNS, Research ad analysis, Governance, are all activities that do not fit within a service transaction model.
- The real direction here is to head to business planning and use the activity plan as the phrasing of immediate objectives and then structure membership income to enable the objectives to be achieved. One important issue here is to understand the basic rationale of APNIC as a mutual benefit membership association, rather than a third party service bureau.
- We should consider setting fees based on AUD or similar to ensure stability of the matching of revenue and expense, and have payment options in currencies and exchange rates.
- Criteria for this fees discussion should be one that terminates quickly.
- One approach here is to use business planning process through the use of member surveys, definition of an activity plan, setting budgetary objectives and settings the fee schedule to achieve these objectives
- There is a need here to address the fees issue now in order to clear some space to consider the medium term challenges we face with the IPv4 / Ipv6 issues.,
Paul Wilson:
- The original APNIC model based on capacity to pay. This fell apart in the 1997 financial tumble with voluntary self-assessment triggering downgrading
- We adopted the address-based model to represent capacity to pay - more addresses in their resource holding implied greater capacity to pay implies greater fee levels
- We see the "all I want is an IP address, why the fee?" response. The issue here is that APNIC is a registry, not just an address allocator. The database records that APNIC maintain are for the benefit of everyone. "Uniqueness" in a registry context is a common benefit. The capacity to pay and the size of the organization are relevant considerations here in general terms
- Other activities are also community based in nature - the benefit or "value" of any of these activities is not the same for all members; it is directly related to the "size" of the organisation
- APNIC is a mutual membership association - we are here for each other for communal benefit, not as an individual benefit
- The budgetary process has enjoyed a high level of trust in the EC process to review the budget
- The concern about the budget process appears to be based in a concern that this is indeed a case of raising the fees, so the attention shifts to the budgetary process
- The intention here is about sustainability of APNIC, not increasing fees for one sector with another at another sector's benefit.
- Budgetary review or activity plan review does entail far more active support and buy-in from the EC - this will involve more EC work to set these objectives with discipline and rigor.
APNIC EC Session
The meeting met without KPMG consultants present to consider various principles for fee options and the criteria for assessment of such fee options
It was observed that while it was interesting to see some alternative models of fees and the effect of turning the parameters of various models, the basic question related to the underlying objective here and the associated deliverable from this activity? The questions to be considered include: Who has the responsibility for selling this to the membership? Who should be assembling the material and presenting it to the membership? How is "consensus" of the membership assessed? Given that many parties would tend to favour the option that reduced their own individual fee payments, then how is APNIC going to achieve consensus across the membership in the matter of fee structure?
One approach considered was to put the option of some form of status quo, perhaps as a relative formula rather than a set fee amount as an option to compare against other proposed fee options.
The meeting considered fee structure principles, or parameters that could be adjusted in a fee model, and agreed on the following list:
- ability to set fees against adopted expense budget, including setting of budgeted reserve amount
- choice of fee currency
- resource holding ageing system, similar to RIPE NCC
- service "intensity" metric
- flatten fee differentials
- NIR/Confederation Per-address fee
- NIR/Confederation member charging vs aggregate holding charging
- Discounted fees (LDC /NIR / ??)
- Voting structure - KPMG's comments invited rather than options modelled
Discussion of various approaches included the observation that there are activities undertaken by APNIC that produce outcomes of common benefit across the region, there are resource holding where the recipient derives benefit from the held resource and there are service transactions with APNIC.
It was noted in discussion of a service transaction-oriented fee structure that many of the Internet's activities that are infrastructural in nature and essential to the continued operation of the Internet are now falling to the RIRs' activity agenda simply because of the lack of any alternative through either industry activity or public sector activity. It was commented that the NIRs are not in a position to take this general infrastructure activity agenda up as there is a consideration of critical size and capability to undertake such roles that are beyond the scope of an NIR.
Discussion of mechanism for adoption - is this a membership vote or an EC decision? It was commented that the powers of the EC are powers of decision that would otherwise be taken by the membership, and the EC has the ability to make such decisions between member meetings (with a provision for membership review at those meetings). APNIC now has e-voting so the criticality of "between member meetings" is no longer a major consideration. The question was posed: Why not have a vote by the members? Discussion considered whether such a vote would be an up / down vote on a single option, or an either / or vote between 2 models, or a single vote across many models? It was also noted that in the APNIC By Laws as they exist right now is that membership votes are in meetings and electronic voting is prior to the meeting, so membership votes are still associated with the membership meeting schedule.
The group considered assessment criteria and agreed on the following list:
- "Fairness"
- Difference to current model, and financial impact to 7 current member categories and NIRs / Confederation members
- APNIC Sustainability
- Impact of new NIRs
- Impact of cessation of IPv4 allocations - KPMG's comments invited rather than detailed assessment
- Impact of no IPv6 subsequent allocations - KPMG's comments invited rather than detailed assessment
KPMG feedback
The meeting resumed with KPMG, who then presented the issues that they had identified during discussions in the morning.
Issues to be considered:
- financial sustainability of APNIC's operations
- business focus
- narrowing expense / revenue margins
- minimize impact of exchange rate variations
- increase revenues
- prudent management of expenses
- Service Delivery
- Member expectation
- Visibly address service expectations are we happy with the current profile?
- Tie-in to service delivery / service intensity?
- Short and Longer term
- Exchange rate movements
- Split fees
- Can small and large members expectations be met in the same fee structure?
- Should we split fee structures between small and large
- Fairness
- Fairer to all
- Current model is good at servicing expert members
- Need to establish what the smaller members are looking for
- Aim to improve service delivery for the smaller members
- NIRs
- Take into account the NIR activities
- Continuous formulae, abolish tiers and abolish per address fees
- Charge same as for APNIC members, discounted in recognition of work NIRs undertake
- Fees - algorithm or formulae?
Need to quantify some of the expectations as indicated in the most recent member survey - how realistic are these expectations? What would it cost?
APNIC concept of mutuality and common beneficiaries
Financial Management considerations:
- fiscal prudence
- activity-based budgetary planning : Secretariat and EC commitments
- EC to set financial parameters
- Fees should be an outcome of a budgetary process
Need for longer term planning (5 years)
- Impacts of IPv4 address runout and size of IPv6 allocations implies a shift away from regular allocation transactions and a greater emphasis on registry maintenance functions and consideration of possible changes in the IPv4 address management model to match evolving industry requirements
Capacity of LDCs in relation to the fees
Account of historical allocations
Change the IPv4 and IPv6 scaling factors
Proposed Evaluation Criteria:
- Sustainability
- Consistency and fairness
- Impact assessment
- Transparency
- Service-Based
- Manage Risk
Potential Risks
- APNIC sustainability
- Impact assessments
- Unfair burdens and disparities continue
- Platform for change
- Member input
- Model complexity and acceptance
- Communications program and implementation of change
Next Steps for KPMG:
- Options to be developed
- Work with the Fee Structure WG and EC
- Put forward a number of options
- Use APNIC meeting in Sept as a sounding board
- EC to decide outcome, or membership to vote
KPMG would focus on deliverables for the 7th September (APNIC meeting week). The EC requested that KPMG look to producing some outcomes by the 24th August to allow for some time to prepare.
Next Steps for APNIC EC and Fees WG
There was concern voiced over the scope of the proposed work. It was seen as a significant deliverable, and doubt was expressed that meaningful outcomes could be produced in the time available that would lead to any resolution in time for the 2008 budget process. This was seen as being an unacceptable outcome in terms of APNIC's immediate financial situation in the 2008 - 2009 period.
The critical question posed here was what is the major immediate issue we are facing here and can they be addressed in some interim manner to allow due consideration of the larger issues of fee structures and principles of equity and fairness across APNIC's membership.
There are current and near term (two to five year) fiscal issues, particularly the fact that, because of currency exchange rates for the US Dollar, needed expense increases to meet service needs while retaining employees, etc., APNIC is likely to have over a US$400k loss in fiscal 2007. This means that pragmatic fiscal changes need to be made this year, while the EC and the membership also do long range business analysis and planning of the larger issues that APNIC is facing.
Given that there are currently two major fiscal causes of the budget shortfall, APNIC income is denominated in the falling US Dollar while it incurs expenses in Australian dollars, and costs of the operation rise with inflation factors, the EC asked APNIC financial management what the current prudent course would be. The answer was twofold:
- Change the fees from being denominated in US Dollars to Australian Dollars so finances are predictable and budgetable, and
- Raise fees for 2008 by 6% across the board, to address current budget shortfall i.e. with no change in the fee model, just raise them all in parallel.
It was also suggested that any subsequent fee increases be capped at no more than 6% per annum.
This proposal would need to be discussed by the APNIC membership as a whole as soon a possible, and formally reviewed and approved at the September Members' meeting if it is to take effect in January 2008.
The Fees WG co-chairs have asked the Secretariat to give the Fees WG a meeting slot in September, as well as time to discuss all this at the Members' Meeting then.
It was noted that this was proposed as a financially prudent action, pending resolution of working out sustaining outcomes. This proposal will allow APNIC the necessary time to work through these longer term change issues. A further comment was made that the EC should have a statement that makes a commitment in terms of dates that resolves the longer term fee structure conversation
Discussion on the basic parameters of the fee structure for APNIC included consideration of a number of quite fundamental perspectives that members have brought to this discussion over the extended period that this discussion has been active in APNIC. This includes:
- NIR / Confederation per-address fees,
- NIR / Confederation competitive interests with respect to APNIC
- (lack of) Transparency of NIR operations and costs for other APNIC members
- Service model and fees based on service "transactions" (transactions in a very broad sense)
- Does the EC have some capability to drive APNIC priorities in activities?
- IPv4 runout - what's the impact here?
- IPv6 address allocation policy - what are the service requirements here?
Action Items
- Fee structure WG to be advised of action proposed by the EC
- Secretariat to prepare material on the proposal above to present to the member meeting in Delhi
- EC to prepare a statement that makes a commitment in terms of timelines and activities that attempts to resolve the longer term fee structure conversation

