Cost-based tariff system SURFnet Walter van Dijk
Mar 30, 2015
Cost-based tariff system SURFnet
Walter van Dijk
Highlights SURFnet
SURFnet is the Dutch National Research & Education Network (NREN) Services, innovation, knowledge Not for profit Task organisation of SURF = ICT collaboration of higher education & research Institutions fulfil two roles: - strategic control / policy
- customer
A small operation serving a large community: 85 employees 180 connected institutions 1 million end-users Turnover 31 million Euro 12 million: innovation subsidies 19 million: tariffs incurred to customers
Guiding Principles
Government funding solely used for innovation
Running of services fully incurred to customers
Not for profit structure
Positioning: exclusively on the demand side of the market
Global collaboration on innovation (GN3+, GLIF, Internet2, CERN)
Various partnerships for innovation with industry
Tariff model
Costs of most services are fixed and largely independent of actual usage
Connection fee covers about 70% of income from tariffs Distribution of connection fee based on cost-sharing model
Other 30% comes from additional services, charged on a per service/customer basis: SURFlightpaths, SURFmailfilter, SURFcertificates (TCS), SURFdomainnames et cetera
Avoid complexity If service = infrastructure: included in connection fee (e.g. network, federation, eduroam) If vast majority of customers uses a service: included in connection fee If only selection of customers want a service: charge as additional service Tariffs for additional services are cost-based Yearly evaluation of cost schemes and tariffs
The why of tariff differentiation
Some services are only used by specific subset of customers e.g. SURFmailfilter: some customers do it themselves or use cloudservices e.g. SURFcontact: HD videoconferencing, developed for academic hospitals e.g. SURFlightpaths: heavy users in astronomy, high-energy physics and lifesciences
NREN proposition is relative to commercial offerings Opting for lump-sum model could discourage customers that require only IP-connectivity
versus
Beware of tariff differentiation
Tariff leads to higher barrier for experimenting/using a service negative incentive for services that we want our
customers to use: e.g. security services
Negative spiral can be induced: fixed cost shared by fewer and fewer customers
Cost allocations
Fairness of tariff differentiation requires accurate cost calculations: Of course: other factors than costs will also determine tariff
ABC starts with determining direct and indirect (‘overhead’) costs
KIS: 1) administration of all direct costs per service (HW, SW,
manpower)2) attribute indirect manpower where possible to services (AA &
CS)3) attribute overhead (HRM, Finance, office etc.) to services
ActivityBased Costing
Step-1: assessment of direct costs, per service (example)
Step-2/3: inclusion of indirect costs, per service (example)
Tariff history
Focus on costs and tariff differentiation (separate charges for additonal services) resulted in significant decreases of connection fee:
2008: 7,5% 2009: 5% 2010: 4% 2011: 4% 2012: 4% 2013: 4%
Number of ‘additional services’ has gradually increased:
2008: SURFlightpaths 2009: SURFmailfilter, SURFcertificates (TCS), SURFcontact
2010: SURFinternetpayments, SURFcertificates (code-signing) 2011: SURFfederation (only for SP’s) 2012: On-Demand lightpaths
Tariffs for some services were decreased: 2009: SURFmedia, SURFdislocations 2010: SURFdomainnames
Tariff differentiation in perspective
Tariff differentiation started at ±20% in 2009
Percentage increased to 35% in 2012
Percentage currently decreases due to: Decommission of services like SURFmedia, SURFgroepen en SURFcontact Decision to stop charging service providers that use SURFfederation/SURFconext Decision to decrease the tariff for certain services like SURFdislocations &
SURFdomainnames
Tariff differentiation revisited
Renewal of contracts for 2013-2016 timeframe accentuates importance of tariff differentiation: as a means, not a goal
Newly developed services require a business plan, including proposed tariff model: constitutes business-case
Trend of differentiation percentage expected to decrease further: Financing of NGE Multi Service Port by not claiming room for 4% decrease of tariffs in
2013 collective decision by universities
Financing of “Point-of-Presence redundancy” by not claiming room for 4% decrease of tariffs in 2014 collective decision by universities
Customer interaction on tariffs
Two separate roles:
Strategic policy control via SURF (College Board members) Customer function via SURFnet (ICT-management, CIO’s)
Annual meeting with representatives of customer sectors