Authentication and Authorisation for Research and Collaboration Policy and Best Practices Harmonisation An overview of the “NA3” activities David Groep Nikhef AARC Kick-Off Meeting June 2015 https://aarc-project.eu Policy challenges What does assurance mean? And to whom? How much differentiation of LoA can people handle? How can we address incidents that propagate through the federated space? Can we get policy coordination to scale? What is the place of third-party commercial and eGov providers? How does that help guest identity? What’s a sustainable distribution of responsibilities amongst AAI participants? https://aarc-project.eu How can we share necessary accounting? Objectives Develop recommendations for best practice in the areas of identity and attribute assurance, and identify the minimal set of policies and best practices that permits grouping of identity and attribute providers. Objectives • provide a level of assurance (LoA) framework that meets the requirements of resource providers and can at the same time be supported by institutions (identity providers); • identify a distributed approach to handling security incidents in a federated environment; • specify scalable policy negotiation mechanisms between identity providers, attribute providers and service providers to facilitate resource providers; • investigate terms of usage for delivering commercial services. https://aarc-project.eu 3 Supporting and encouraging policy development Structured to support, leverage, and enhance emerging and nascent activities • Assurance level development: driven by infrastructure needs and application use cases DPCoC adoption, eIDAS levels, NIST LoA revision process, IGTF LoA generalisation, VoT, … • Incident response support development and adoption of SirTFi, and OpSec trust frameworks • Service operational models and sustainability (“who might operate which parts of the AAI?”) leveraging DARIAH developments • Scalable policy negotiation SCI Security for Collaboration Infrastructures, REFEDS EntityCategory work • Accounting and processing of data developing R&E legal guidance (and comparing that to practices in e.g. STORK), existing development of e-Infra policies on accounting [but this is a most challenging task] https://aarc-project.eu 4 Best Practices for LoA TNA3.1 – Mikael Linden (CSC) • Gather needs from research infrastructures and communities • Gather current LoA practice from R&E federations • Initial milestone coming up to collect at least a minimal level to ’promote’ (M7) • That level should be useful to relying parties – but also feasible to provide by authentication and attribute sources! https://aarc-project.eu 5 Assurance Assurance: Implementation Assurance: The profile Communities AARC Assurance: Assurance The Players Federations (IdPs) REFEDS eduGAIN Whole ecosystem https://aarc-project.eu 6 Assurance timeline AARC May 15 – 17 GN4 15 – 17 • Collect Requirements for assurance • Define a way to address assurance that works for the user-communities and for the federations • Define the service model that would offer this type of profile https://aarc-project.eu GN4 2016-2017 • Transition from pilot to production 7 Internal LoA players, how to coordinate? AARC SA1 Proof of concept AARC JRA1 Architecture AARC NA3T1 Training AARC NA2 ? GN4p1 SA5T1.4 https://aarc-project.eu Deliverable DNA3.1 M23: LoA recommendations External LoA players, how to collaborate? NIST 800-63 STORK Kantara IAF eIDAS LoA InCommon IAP Low adoption https://aarc-project.eu Stable 6/2015, published 9/2015 IGTF R&E R&E federation R&E federation practices federation practices practices Currently fragmented Incident Response in federated environments TNA3.2 – Romain Wartel (CERN) Goals • Review existing practices • Propose recommendations for incident response • Generic security incident response procedure for federations Input and expected collaboration • Sirtfi, EGI CSIRT, NRENs, REFEDS, eduGAIN, FIM4R, RDA, etc. Milestones and Deliverables • DNA3.2: security incident response procedure for federations (M20) Draft execution plan • Identity key partners and contacts • Gather requirements, operational security standards and practices • Propose and discuss a practical and reasonably simple procedure • ‘Consensus is more important than actual content’ https://aarc-project.eu see e.g. https://wiki.refeds.org/display/GROUPS/SIRTFI 10 Recommendation for service operational models for enabling crossdomain sustainable services TNA3.3 – Peter Gietz, Martin Haase (DAASI) ● What Policies are used in today's R&E federations? – – ● Legal aspects, e.g. privacy / personally identifiable information; levels of assurance Organizations matters, e.g. server security, currentness of account data What Best Practices are recommended in today's R&E federations? – – – – IdP / SP operations: SAML metadata management Mapping of entity categories and their requirements/policies to concrete operational processes Promotion of federated access by campus users to services What issues does interfederation pose over regular federation (i18n, attribute release, ...)? https://aarc-project.eu Sustainability and Results ● ● Sustainability – Policies for homeless user accounts lifecycles – Policies for translating social network identities into SAML federation users – Policies for attribute authorities / gathering user consent – Operation model Results – recommendation for possible operating models for sustainable crossdomain collaboration – input for the various proof-of-concept implementations in SA1 – Deliverable DNA3.3 - Recommendations for service operational models for enabling cross-domain sustainable services, M21 https://aarc-project.eu Development of scalable policy negotiation mechanisms TNA3.4 – David Kelsey (RAL) Develop scalable policy negotiation mechanisms between identity providers, attribute providers and service providers • Note: bi-lateral negotiation between SPs and IDPs/AAs will not work! Effort funded: 19 PM (CERN, FOM-NIKHEF, GRNET, LIBER, RENATER, STFC) Input required • MJRA1.4 – First version of blueprint architecture (M15) this is the one we expect, not the one in the DoW … Milestones and Deliverables • DNA3.4 - Recommendations on the grouping of entities and their deployment mechanisms in scalable policy negotiation (M24) Draft Execution Plan – for discussion • The first work will be the identification of the entities that need to be classified and expressed, such as the specific categorisation that may be needed for non-identity attribute providers and for credential translators (Propose M9) • proceed to formulate recommendations on the grouping of entities and on the actual deployable mechanisms that can be evaluated in the proof-of-concepts of SA1 (Propose M12) • The final recommendations will then be provided to the operational infrastructures and federations. (DNA3.4) (M24) https://aarc-project.eu 13 Accounting and the processing of data TNA3.5 – Marcus Hardt (KIT) • Rationale of the task • Information from monitoring and accounting (m&a) is personal data • For users it is not obvious if and under which circumstances his data are stored at a computer center and / or leave the site and are stored / processed by third parties • Goals of the task • • • • Identify stakeholders (VOs, Sites, Funding Organisations, ...) Identify acceptable policies for [storage & release & processing] of m&a data Identify reasonable requirements for the above Develop a workflow for acceptable (by legal and technical) requirements for processing of data • Cheap and quick example: • 1: Aggregate on a VO-level prior to publishing, • 2: Publish only to parties with reasonable and verified requests • 3: Maximum storage periods (e.g. right to be forgotten) • Define scheme to inform users about who will be given which data on which granularity https://aarc-project.eu Effort and partners available in NA3 Partners • CERN, CSC, DAASI, GRNET, KIT, LIBER, Nikhef, RAL, RENATER, SURFsara • 66 PM available • with slight emphasis (>10PM) on incident response (“SirTFi”), scalable policies, and LoA But the only way to validate the proposed best practices (and make them useful) is • Incorporation in pilots of SA1 • Adoption by REFEDS and operational work in GEANT and at the NREN level • And: we desparately need the IdPs, SPs, e-Infra’s and communities involved We can probably reach the federation folk, the e-Infras like EGI and PRACE, and ESFRI-sized communities – but we rely on training and dissemination to reach the others! https://aarc-project.eu 15 Time line, pre-set mile stones, and deliverables Tod ay 3rd Quarter 4th Quarter M7 1st Quarter 2nd Quarter 3rd Quarter M18 4th Quarter M20 1st M21 M232ndM24 Quarter Quarter Start Finish 01/05/2 015 30/04/ 2017 Recommendation on relevant assurance levels 30/11/2015 Requirements of data to protect 30/11/2015 Recommend. processing of personal data 31/10/2016 Best practices collective SIR 31/12/2016 Recommend. Recommend. LoA service recommen on grouping operational of entities dations models 30/04/2017 31/03/2017 31/01/2017 Two early milestones • On a relevant minimum assurance level to promote – in particular for use in any SA1 pilots • Collection of the types of data (attributes, accounting records, log data, over-the-wire monitoring, &c) and the participants (AAI IdPs, individual SPs, e-Infrastructures, communities, other roles) that need to be addressed by the recommendations We expect a more continuous stream of actual output: both recommendations and actual use but will rely on both the SA1 pilots and ‘external’ entities (REFEDS, e-Infrastructures, GEANT4) to work with the NA3 policy recommendations and best practices! https://aarc-project.eu 16 Thank you Any Questions? davidg@nikhef.nl https://aarc-project.eu © GÉANT on behalf of the AARC project. The work leading to these results has received funding from the European Union’s Horizon 2020 research and innovation programme under Grant Agreement No. 653965 (AARC). https://aarc-project.eu