Internet2-NLR MPT Final Report, August 15, 2007

 [Download PDF version here]

Merger Planning Team Final Report

August 15, 2007


Preamble


Over the past several years, Internet2 and National LambdaRail (NLR) have created advanced networks, assuming leadership responsibility for meeting the networking needs of the U.S. research and academic community and for building relationships with the major networks that support the corresponding communities abroad.  


The contributions of Internet2 and NLR during the time of their existence have been enormous.  Unfortunately, as their spheres of activity have come to overlap more and more, their relationship has evolved from one of healthy complementarity into one of self-destructive competition.  The past 18 months have witnessed a distracting diversion of resources and energy away from the primary tasks of innovation, research, and education, into an excessive focus on the organizations themselves.


A merger holds the promise of bringing an end to this battle for organizational survival, enabling the entire community to focus its energies outwards towards productive ends.  A successful merger would create an organization that can draw on the best attributes of each existing organization, with enhanced flexibility and adaptability going forward.  It could work in more effective partnership with other members of the national and global cyberinfrastructure community, promoting understanding of the transformative power of advanced networking for the future of humanity.


Building on the work of others, the Merger Planning Team is of the unanimous view that Internet2 and National LambdaRail should merge on the terms described in the attached Definitive Agreement to Merge.  We believe that the research and education communities whom our organizations were created to serve will derive greater benefit from the creation and operation of a merged organization on these terms than they would derive from the continuation of two separate, rival organizations.  We believe that the terms of the merger described below are fair to all concerned and will establish a structure under which the merged organization can succeed in both the short and long term.


The merged organization proposed in the Definitive Agreement to Merge is neither Internet2 nor National LambdaRail, but a combination of the two.  Following on the recommendations of the Network Planning Team, we have not predicated our recommendation upon the anticipated elimination of any specific current activities.  We are recommending that, with a new CEO, a new Board, and a new governance structure, the new organization should undertake a comprehensive strategic planning exercise, including a top-to-bottom review of current programs, in order to determine what activities should be priorities for the long term.  There is no reason to assume, a priori, that those activities will not include a full complement of network services (IP, commercial peering, demand aggregation for price discounts on commodity traffic, Layer 2 service, and two different varieties of DWDM service), network-related activities (applications, middleware, security, etc.), and community leadership on the national and international platforms.  We recognize that this approach poses financial challenges for the merged organization. But we believe that those challenges can be overcome through a participatory planning process that will ensure that the organization is appropriately financed for the activities its members wish to support.


In the pages that follow, we first describe our recommendations for the way key issues should be resolved in the Definitive Agreement to Merge and the Bylaws of the merged organization.  We then describe the basis for our conclusion that the merged organization will be financially viable.  A set of Appendices complete our report.


Key Principles of the Merger


Mission.   


I2-NLR represents a merger of equals between Internet2 and National LambdaRail designed to capture the best of both organizations and their approaches to providing advanced networking infrastructure and related services to the U.S. research and education community. I2-NLR incorporates the missions of both organizations.


See the proposed vision statement included in Appendix A.


See Bylaws Article I.


Commitment to Full Use of the IRUs.


I2-NLR will use the NLR IRUs as an integral part of the national network infrastructure for the full life of the IRU agreements.   Failure to do so will trigger IRU clawback to/by NLR Investors.


See Bylaws Article XII, Section 3.


Dual Infrastructure Evaluation and Full Control of Optical Paths.


We endorse the recommendation of the NPT that the operation of two national optical networks over the next several years inform community-based decisions regarding architecture, operation, and financing of one or more national R&E backbones in the future.  Beyond network engineering considerations, operation of the two networks with their differential provisioning costs and service offerings will enable the community to reach a better understanding of the array of services that I2-NLR should provide to support the R&E community.  In both cases, I2-NLR anticipates operating its networks on systems of fiber and equipment that are wholly dedicated to its use.  In all events, I2-NLR must have the ability, in cooperation with its regional connector partners, to promptly establish and take down optical paths across the networks at the times and in the manners of its own choosing to support its leading edge mission.  It must have the independent ability to monitor network performance and to measure network capabilities directly, through its own employees and agents of its own choosing.  We recognize that there is a serious debate about the relative benefits and liabilities of owning and leasing optronics, in which reasonable people hold different views.  Ownership advocates point to greater security of title, greater independence from vendor fortunes, and widespread deployment of owned infrastructure among leading state and regional networks.  Leasing proponents assert such advantages as greater flexibility in technology-driven upgrades, and the opportunity to provide stronger guarantees of reliability and performance.  There are also differences of opinion about full life cycle costs associated with the two approaches.  We view operation of the two networks in parallel over the next several years as a valuable means of assessing the complementary costs and benefits of each approach and gaining insight into price elasticity as well as network externalities within the R&E environment.  The Architecture and Operations Advisory Council (AOAC) will provide advice and guidance to the Board of Trustees on the development of the I2-NLR network infrastructure and services over time.


See Bylaws Article XIII Section 3.


Partnership with Regional Connectors.


Regional Connectors (RCs) are the primary and preferred partners of the Corporation in deploying national networks, connecting members within regions to the Corporation’s networks and services, providing exchanges and collocation capabilities in a region, and collaborating with and connecting to other regional, national, and international networks.  RCs will comprise the “Investors” (current members of NLR) and other multi-entity connectors to the backbone.  In the event that an RC is unable or unwilling to provide a particular service as required by one of its members, the member, the RC, and I2-NLR will strive to develop jointly an alternative approach to delivering such a service.  The RC Caucus and the Executive Liaison Caucus shall be consulted on matters relating to the relationships and conflicts among I2-NLR, an RC, and a member. 

On and after July 1, 2009, all RCs will have access to the same I2-NLR service offerings at the same prices, subject to the RC’s willingness and ability to make appropriate connections to the national backbones.  During the Initial Period, the Investor RCs will receive certain special considerations described below. During the Initial Period, Non-Investor RCs shall be eligible to acquire static waves on what is now the NLR network via Investors only through an Investor RC connected to the NLR network.


Based on geographic considerations and the deployment and availability of fiber, we expect that over the long term there will be on the order of 15 to 25 RCs.  We recognize that there is a serious debate about how much authority within the overall system should reside with RCs, how much with individual campuses, and how much with a representative national organization.  We expect that one of the long-term challenges for the research and education community will continue to be the calibration and re-calibration of that balance.  I2-NLR will work with the RC Caucus and the Executive Liaison Caucus to develop mechanisms for transparency, inclusive participation, and sharing of best practices at all three tiers (individual campus, regionals, national), enabling processes through which expectations are set, compliance is monitored, and disputes are resolved at all three tiers.  It is essential that RC’s be supported as decentralized sources of deployment, financing, innovation and experimentation, and local responsiveness, without setting in concrete a particular geographic instantiation of regionalization that might inhibit dynamic growth or response to particular opportunities and changes.  And at the same time, we must be able to work together in a consistent and coherent manner at all tiers to support the achievement of collective national goals, many of which are long-term in nature.  


Existing Internet2 member institutions may make use of the networks under existing Internet2 rules throughout the Initial Period, and thereafter until the relevant governing rules may be modified by the Board.


We expect a future in which an increasing number of our community’s key partners will include large national and international entities.  We do not believe a single paradigm will be right for all such partnerships.  Sometimes, as in the case of the ESnet partnership with Internet2, a single comprehensive partnership relationship may be appropriate.  At other times, as in the case of NASA’s partnership with National LambdaRail, a more distributed approach will be appropriate.  In all cases we expect the development of such relationships to proceed through an inclusive and participatory process, working with RCs and not bypassing them.  The AOAC will provide ongoing guidance to the staff and Board on preferred architectures for new kinds of partnerships.


See Bylaws Article XIII Section 4.


Investor Benefits.  


In recognition of the Investors’ past contributions to National LambdaRail, they will receive certain special benefits, relative to other RC’s:

On the financial side, they will receive: (1) perpetual clawback rights as specified above; (2) the right to continue their existing connection to the NLR network for a single connection fee of $500K during the final year of their NLR membership contract, and at the rate of $500K for each twelve month period throughout the remainder of the Initial Period.  Those Investors who have paid more than $500K during 2007 in the final year of their contract shall receive a credit of the excess paid above $500K, that may be applied during 2008; (3) the right to receive a discount of 50% on their existing connections to the Internet2 network during the period extending from the closing date through 12/31/07, and to continue those connections without charge for the remainder of the Initial Period;  (4) the right to receive unlimited access to all I2-NLR shared services and to purchase carrier-provisioned waves at prevailing member prices; and (5) the opportunity during the Initial Period to enter into contracts with terms of up to five years to provision static waves on what is now the NLR network at the current NLR prices. (Note that participant fees for entities connected by the RC, including SEGP fees, will still apply.)


On the governance side, Investors will receive enhanced standing within the Board and Councils as specified in the MoA, after which they will become part of the RC constituency.


On the geographic responsibilities side, Investors will receive initial areas of geographic responsibility consistent with their NLR areas of geographic responsibility, but not to the exclusion of the ability of other existing non-Investor RC’s to provide services within their current geographic areas of activity.  During the Initial Period, any change in these patterns of service would need to be considered through a transparent and consultative process involving the RC Caucus, and endorsed by a 75% vote of the I2-NLR Board.


See Bylaws Article XII, Article XIII Section 2.


Governance.  


The following points of clarification are made to the MoA’s discussion of governance:


It is expected that the governance structure of I2-NLR will evolve continuously in response to changes within and outside the community.  To facilitate that process, the Governance and Nominations Committee (GNC) will be charged during the 2008-2009 fiscal year with responsibility for conducting a full governance review, including consultation with the community, similar to that which was followed by the Internet GNC during 2006, and to prepare a public report with recommendations that will be ready for consideration by the I2-NLR Board that will take office on July 1, 2009.  This review will consider the lessons learned by the GNC in the course of the first round of elections, input developed through community consultation, and changes in circumstance that result from the merger. 


During the Initial Period the membership of the Governance and Nominations Committee will be expanded to include two leaders within the Investor community, to be specified by the Chair and Vice Chair of I2-NLR.  


A set of new Caucuses will be created as special resources to the Board, Management, and Councils:  a Network Researcher Caucus, a Disciplinary Researcher Caucus, an RC Caucus, and an Executive Liaison Caucus (corresponding to Internet2’s “Community Leaders Forum”).  Caucuses are intended to provide constituent input into the more formal governance processes.


During the Initial Period each Advisory Council may choose to elect as its Chair any member of that Council or any member of the Board.  In order to ensure a strong liaison between the Councils and Board, if the elected Council Chair is not a member of the Board then the Council will be asked to also select a Vice-Chair who is a member of the Board.  The Council membership may expand to 16 in order to accommodate a Board member who may join the Council as Chair or Vice-Chair. 


The list of 21 Board members identified in the Memorandum of Agreement who shall serve during the Initial Period is replaced by the list of individuals included in Confidential Appendix E.


In the event any of the 21 named Board members is unwilling or unable to serve during the Initial Period, a replacement shall be nominated by the Chair and Vice-Chair of the Board and ratified by a 50% vote of the Board.


See Bylaws Article IV Sections 2, 10; Article VI Section 2.


Acceptable Use Policy.


I2-NLR is committed to limiting the use of Acceptable Use Policies to the greatest extent possible and will avoid entering into contracts that limit its flexibility and that of its membership in this regard.  In order to provide maximal flexibility for its constituents, I2-NLR will not implement any Acceptable Use Policy unless it is necessary under applicable state or federal law (such as that which is necessary to preserve the tax-exempt Code Section 501(c)(3) status of I2-NLR) or is necessary to fulfill current contractual obligations. Members or other partners acquiring lambdas or other services from I2-NLR may implement Acceptable Use Policies appropriate to their program characteristics and participants.


See Bylaws Article XIII, Section 1.


Scope of Services and Service Review.


The initial scope of activities of Internet2-NLR will encompass all the services, projects, and programs currently being carried on by both Parties. These activities of the Parties will remain and continue to operate in their normal fashion until the Internet2-NLR Board has completed an assessment of the scope of activities and, based upon that assessment, established a plan for changes to the scope of activities of Internet2-NLR going forward.


The Bylaws and Definitive Agreement to Merge provide the initial basis for the merger of Internet2 and NLR.  In addition to the ongoing work to integrate the network capacity offerings of both organizations, the Initial Board shall direct a highly participatory top-to-bottom review of the service offerings of the organization.  The purpose of this review shall be to validate the relevance and importance of all services offered, to ensure the financial capability of the organization to deliver such services in a high-quality manner, to support alignment of the organization's staffing with its mission and direction, and to provide input into any changes in the financial and membership model of the organization as required to ensure transparency of financing and member acceptance of any cross-subsidies.


See Bylaws Article XIII Section 5.


Public Goods and Internal Subsidies.


Internet2-NLR is a nonprofit membership organization that is dedicated to the provision of a range of public goods for the research and education community, with a rigorous commitment to transparency in prioritization and internal subsidization of projects, rather than a fee-for-services organization that prices all elements of its activities as “tubs on their own bottoms.”  That said, I2-NLR is committed to ensuring transparency of financing and member acceptance of any cross-subsidies.


See Bylaws Article XIII Section 5.


Executive Management of Internet2-NLR.  


Upon approval by both organizations of the Definitive Agreement to Merge, a CEO Search Committee will be formed to recruit a new CEO for Internet2-NLR.  The CEOs of Internet2 and NLR shall be ineligible to serve in this position.  The CEO Search Committee shall consist of Jeffrey Lehman, Tracy Futhey, James Bruce, and Wayne Clough, and shall be supported by a professional search consultant.  Should a new permanent CEO be unable to begin service on the Closing Date, the Internet2-NLR Board shall appoint an interim CEO who is neither the CEO of Internet2 nor the CEO of NLR.


See Definitive Agreement Section 5.15.


Financial Viability of the Merged Organization


The Merger Planning Team (MPT) has worked to confirm that it is economically feasible for a merged organization to operate pursuant to a vision statement that it has previously provided, at a cost that is affordable to the stakeholder community.  The merged organization’s operational costs and revenues will derive from network operations and from the conduct of other “core” organizational activities not directly associated with network operations.  As a matter of accounting, we expect the merged organization will track and report those two different kinds of activities separately in order to achieve the commitment to financial transparency and explicitness of subsidies, but will operate as a single, integrated organization.


With respect to network operations, the MPT has taken as its starting point the report of the Network Planning Team (NPT).  The NPT concluded its analysis of the technical issues associated with a merger of Internet2 and National LambdaRail by describing a hypothetical approach to consolidating the two networks.  The NPT approach is designed “to optimize on flexibility and long-term cost savings rather than focusing primarily on near-term cost savings.”  The key features of the NPT approach are:


Consolidation of IP connectivity services using the Cisco routers.  (The MPT recommends that peering between the two IP networks be implemented as quickly as possible pending consolidation.)


Peering services using the Juniper routers, partnering with Juniper and potentially an independent third party peering service such as TransitRail.

Layer 2 and Experimental Services optimized based on customer demand.


Two different types of wave services (at least through 2009) at two different price points reflecting two different service levels.


In conducting this exercise, the MPT has sought to develop an approach to projecting the likely revenues and costs of the merged organization that it recommends for use as a point of departure (an approach referred to as “Scenario 1” below).  At the same time, the MPT recognizes that the Boards of the existing organizations will want to understand the sensitivity of the projections in Scenario 1 to variation in the background empirical assumptions, and that the Board and management of the new organization may choose to depart from the policy assumptions underlying Scenario 1.   


Accordingly, the MPT has sought to explore a range of options available to the new organization, and the plausibility under a range of different circumstances of the proposition that the merged organization will provide genuine and cost-effective value to the research and education community, relative to continuation of the status quo.  To do so, we have relied upon a spreadsheet model that was created to show the relationship among different parameters, and how varying those parameters might affect the overall financial picture of the organization.  

The model operates across the period from 2008 through 2023.  We have considered the years after 2013, however, only to ensure that we do not construct scenarios that work in the short term at the expense of the long term.  We do not report numerical projections beyond 2013, as we believe the future is too cloudy for specific numbers to have analytic value for decision-making.


In creating different scenarios with the model, we have made a few structural assumptions.  (These assumptions were either non-parametric features of the model, or were parameters that we did not vary significantly.)  Some of these assumptions reflect empirical assessments of the environment in which the organization will operate, others reflect recommended policy commitments, and others reflect assumptions about how the new organization might conduct its affairs that we find plausible without denying the plausibility of other policy choices.  The most important structural assumptions are:


Two sets of service offerings.  The organization will carry forward the full set of service offerings described in the NPT report.  At the Layer 1 level, those offerings will cluster into two sets of services:  a “carrier provisioned network” featuring the end-to-end dynamic provisioning and carrier-class guarantees of the current Internet2 network at a higher price point, and a “community provisioned network” featuring the static waves of the current NLR network at a lower price point.  The organization will retain the ability to route traffic between networks in order to optimize traffic flow.


Technological progress.  We assume that technology relevant to the activities of the merged organization will continue to improve dramatically, and at times disruptively.  Sometimes such improvements will translate into reductions in the price of equipment needed for network operations, and such reductions can be modeled parametrically within the model.  At other times, the improvements will yield quality improvements rather than price decreases.  Perhaps most importantly, certain changes (such as the potential transition to 100 Gbps waves within the next 2-4 years, or the possible development of longer-distance transmission technologies in the middle term) could yield non-incremental changes in cost structure or in service delivery.  The model does not attempt to reflect such possibilities.


The model also includes many parametric assumptions that we have varied to create different scenarios.  The most important parametric variables are:


Fee structure for the merged organization.  We have experimented with different combinations of charges to different participants in the community.  Generally, we separated “membership fees” from “connection fees.”  In the Initial Period, connection fees hew closely to the current structure of fees with which community members are familiar; in the Long Term, we also experimented with simplified models that rely on fewer distinct charges.


Base cost structure for the merged organization.  We have experimented with different assumptions about the pattern of costs the organization will incur for programs and core operations.  


Cost structure for additional additional Layer 0/1 capacity.  We have experimented with different assumptions about future patterns of cost to the organization of acquiring additional optical wave capacity, both for community-provisioned and carrier-provisioned network services.

Price structure for Layer 0/1 services.  We have experimented with different assumptions about the average prices that the organization’s “customers” will pay for optical wave services.  We note that these average prices are the result of a number of different factors:  the terms (1-5 years) of wave contracts, the nature of contract purchasers (e.g., R&E members vs. government partners vs. commercial researchers).

Demand for layer 0/1 services.  The model assumes (in accord with existing usage patterns) that different segments of the network experience different levels of use (low, medium, and high).  We have assumed that demand will grow across all segments, but at different rates.  We have experimented with different assumptions about the rate at which overall demand will grow and with different assumptions about the distribution of that demand between community-provisioned and carrier-provisioned waves.


Core community size.  The model assumes that the total number of campus users will remain relatively stable, although it could clearly rise or fall.  We allow for greater variation in the number of regional connectors, reflecting different assumptions about the degree of aggregation that will prevail in equilibrium.


We began by creating a “model” or “base” scenario, which we have referred to as “Scenario 1.”  The key assumptions in Scenario 1 are as follows:


During the Initial Period, the population of RCs and network participants remains relatively stable, paying fees according to the terms of the Definitive Agreement.  Beginning in 2010, the population of network participants remains stable at 240, and the population of RCs remains stable at 24.


In 2008, total base expenditures for the merged organization’s various programs and activities are reduced by 9% from the sum of each separate organization’s most recent projected expenditures for that year.


After the Initial Period, base membership fees for campuses increase to an average of $50,000 per campus, Connection fees are set at an average of $500,000 base per Regional Connector (RC), each of which is assumed to connect an average of 10 campuses, plus an average of $42,000 per connected campus.  The model is neutral as to whether the per-campus charge is collected directly from the campus or indirectly from the RC.


Demand for optical waves grows modestly overall (at a rate lower than has been suggested informally by members of both the Internet2 and NLR staff).  Growth in demand from the community is directed primarily towards community-provisioned waves (75% until 2013 and 67% thereafter).


Demand for optical wave contracts is distributed equally across 1-, 2-, 3‑, 4-, and 5-year contracts.

XXX.

XXX.

XXX.


Technological progress leads to steadily falling prices, both for electronics purchases and for managed services.


In general, the MPT believes these to be a fairly conservative set of assumptions.  Appendix D shows the picture of operations and results for the period from 2008 to 2013 projected by the model under Scenario 1.


Appendix D also shows how seven different scenarios change the projections obtained under the model.  The Scenarios are as follows:


Scenario 2 (“Faster Waves Sales - Balanced”) retains the features of Scenario 1, with the following modifications:


Demand for optical waves grows more robustly overall, with demand equally divided between community-provisioned and carrier-provisioned waves.


Scenario 3 (“Faster Wave Sales – Carrier Provisioned”) retains the features of Scenario 1, with the following modifications:


Demand for optical waves grows more robustly overall, with 80% of demand going to carrier-provisioned waves.


Scenario 4 (“Faster Wave Sales – Community Provisioned”) retains the features of Scenario 1, with the following modifications:


Demand for optical waves grows more robustly overall, with 80% of demand going to community-provisioned waves.


Scenario 5 (“Faster Wave Sales, New Anchor Tenant, Reduced Costs to Community Connections”) retains the features of Scenario 1, with the following modifications:


Demand for optical waves grows more robustly overall, with demand equally divided between community-provisioned and carrier-provisioned waves until 2013, and with 75% of demand going to carrier-provisioned waves thereafter.


Before 2010, a second additional partner from outside the community commits to a significant contract for carrier-provisioned wave services.

After the Initial Period, connection fees are lowered to an average of $400,000 per RC plus an average of $42,000 per connected campus.  The model remains neutral as to whether the per-campus charge is collected directly from the campus or indirectly from the RC.  This demonstrates just one example of how the new organization may adjust the prices of services to members based on overall market demands.


Scenario 6 (“Greater Aggregation”) retains the features of Scenario 1, with the following modifications:


While the total number of network participants remains constant, the number of RCs falls to 20 by 2010 and to 15 by 2023.


Scenario 7 (“Costs Don’t Fall”) retains the features of Scenario 1, with the following modifications:


Technological progress produces increased performance but not falling prices.


Scenario 8 (“Slower Wave Sales”) retains the features of Scenario 1, with the following modifications:


Demand for optical waves grows at half the rate projected in Scenario 1.


Our analysis indicates that the new organization will enjoy a healthy financial future under a wide range of conditions.  While in most cases the organization will operate in deficit for 2008 and 2009, based on the expected levels of Net Present Value of the organization through 2013, the cumulative deficit would be well within the financial carrying capacity of the merged organization.  We believe that incurring deficits of this order would be a reasonable approach for the new Board to accept while it undergoes strategic planning and program review.


Conclusion


We unanimously recommend that the Boards of Internet2 and National LambdaRail agree to merge their two organizations into a single new organization, on the terms set forth in the attached Definitive Agreement to Merge.


Respectfully submitted,


Tracy Futhey, co-chair

David Lassner

Jeffrey Lehman, co-chair

Harvey Newman

Dan Updegrove


Appendices:

Proposed Vision Statement for the Merged Organization

Proposed Definitive Agreement to Merge

Proposed Bylaws of I2-NLR

Scenarios Illuminating Financial Viability Analysis

Confidential Appendix:

Proposed Substitute List of Names of Initial Board Members