Federated Identity Management for HEP

Federated Identity Management for HEP

David KelseyHEPiX, Ann Arbor MI

30 Oct 2013

Overview• What is federated IdM?• A review of some IdM topics

– Since April 2012 HEPiX (Prague)• FIM4R, REFEDs and VAMP meetings• WLCG activities• Levels of Assurance - IGTF IOTA• Next steps

30 Oct 13 HEPiX IdM, Kelsey 2


SP

User

IdP

Many different permutations depending on the technology

IdP: Identity Provider

SP: Service Provider


SP

User

IdP

Then add a community

operated attribute authority (for

AuthZ), e.g. VOMS

AA

Talk at CHEP 2013• “Horizon 2020: an EU perspective

on data and computing infrastructures for research”– Plenary talk by Dr. Kostas GLINOS

(European Commission)• He suggested they are working towards

a vision of an integrated e-Infrastructure with all researchers accessing via a federated digital identifier


Helsinki (CSC) IdM meetings• 30 Sep 13 – VAMP (VO Architecture

and Middleware Planning)– http://www.terena.org

/activities/vamp/ws2/• 1-2 Oct 13 – Joint FIM4R & REFEDs

– https://refeds.org/meetings/oct13/index.html


http://www.terena.org/activities/vamp/ws2/



https://refeds.org/meetings/oct13/index.html




Introduction – FIM4R• Federated Identity Management for Research

Collaborations• An activity (EIROforum) that started 2 years ago in

Europe– To explore and document a joint vision and our common

requirements for FIM– And describe issues that make progress difficult

• Includes: Climate Science, Earth Sciences, ESA, High Energy Physics, Social Sciences & Humanities, Life Sciences, Neutron & Photon Facilities, WeNMR– And open to any others who wish to join

19 Jun 13 FIM4R, Kelsey 7

Workshops and Paper• 6 workshops to date

– link to Mar 2013 agenda (and links therein)http://indico.psi.ch/conferenceDisplay.py?confId=2230

• April 2012: We prepared a paper that documents use cases, common requirements, a common vision and recommendations

• Paper: CERN-OPEN-2012-006: https://cdsweb.cern.ch/record/1442597


http://indico.psi.ch/conferenceDisplay.py?confId=2230

https://cdsweb.cern.ch/record/1442597

Common vision statementA common policy and trust framework for Identity Management based on existing structures and federations either presently in use by or available to the communities. This framework must provide researchers with unique electronic identities authenticated in multiple administrative domains and across national boundaries that can be used together with community defined attributes to authorize access to digital resources 19 Jun 13 FIM4R, Kelsey 9


Pilot Projects

Roadmap for collaboration

REFEDS/eduGAIN produced a document to address FIM4R issues: Provides an initial list of

prioritised requirements (thanks also to Bob Jones & co.)

Addresses some perceived issues

Presents proposals to solve some of the challenges

https://refeds.terena.org/images/3/3e/AnalysisFIMDocumentv0.7.pdf

Identity Management for Research Collaborations: from Pilots to Production

Bob JonesIT deptCERN

CERN EFDA EMBL ESA ESO ESRF European XFEL ILL

EIROforum

CERN

EFDA

EMBL

ESA

ESO

ESRF

E. XFEL

ILL

Bob Jones (CERN) – October 2013


A Vision for a European e‐Infrastructure for the 21st Century

Sustainable - RIs currently in construction (FAIR, XFEL, ELIXIR, EPOS, ESS, SKA, ITER and upgrades to ILL and ESRF etc.), need to be convinced that e-Infrastructure will exist and continue to evolve throughout their construction and operation phases if they are to take the risk and invest in its creation & exploitation

Inclusive - Need an e-Infrastructure that supports the needs of the whole European research community, including the “long tail of science”, and interoperate with other regions

Flexible - Cannot be a one-size-fits-all solution

Integrated - Coherent set of services and tools must be available to met the specific needs of each community

Innovative - Essential that European industry engage with the scientific community in building and providing such services

User driven - The user community should have a strong voice in the governance of the e-Infrastructure

See https://cds.cern.ch/record/1550136/files/CERN-OPEN-2013-018.pdf

https://cds.cern.ch/record/1550136/files/CERN-OPEN-2013-018.pdf

https://cds.cern.ch/record/1550136/files/CERN-OPEN-2013-018.pdf


Prototype Centres of Excellence – Example from CERN

• This Centre of Excellence will focus on data-centric services representing a platform on which more sophisticated services can be developed

• Use the resources installed by CERN at the Wigner Research Centre for Physics in Budapest, Hungary

• Services will be accessible via single sign-on through a fed id. mgmt system– Multi-tenant compute environment to provision/manage networks of VMs on-

demand– ‘dropbox’ style service for secure file sharing over the internet– Point-to-point reliable, automated file transfer service for bulk data transfers– Open access repository for publications and supporting data allowing users to

create and control their own digital libraries (see www.zenodo.org)– Long-term archiving service– Integrated Digital Conferencing tools allowing users to manage their

conferences, workshops and meetings– Online training material for the services

http://www.zenodo.org/


Sustainability of CERN’s Centre of Excellence - role of partners

• Partners will– curate their data-sets– connect their identity federations– deploy their community specific services & portals– manage the interaction with their registered users and

associated support activities

• Beyond this first year, partners engage to fund the cost of the services their users consume according to a pay-per-usage model (to be jointly-developed with CERN during the first year)


How can we create e-infrastructures that overcome fragmentation?

• Fragmentation of users (big science vs. long tail)• Fragmentation of infrastructure (not integrated services)• Common platform (e-infrastructure commons) with 3 integrated areas

– International network, authorization & authentication, persistent digital identifiers

– small number of facilities to provide cloud and data services of general and widespread usage

– Software services and tools to provide value-added abilities to the research communities, in a managed repository

• Need a data continuum - linking the different stages of the data lifecycle, from raw data to publication, and compute services to process this data

18

WLCG Group

FIM4R Meeting, Helsinki, Finland, 2-3 October 2013, D. Kelsey, R. Wartel

Identity Federation - HEP/WLCG

http://indico.psi.ch/conferenceDisplay.py?confId=2230

19

Use cases• Use cases in WLCG are:

– Web-based (grid portals used for job submission, wikis, etc.)– CLI-based (job submission, admin tasks)

• Use cases foreseen by the experiments for federated identities– Alice: Interested - but would rather the work to focus in priority on

the Web use case– Atlas: Interested in both CLI and Web use case– CMS: No immediate adoption foreseen, but supportive of both CLI

+ Web use cases– LHCb: Interested - in particular in a CLI

19

20

Technical work• Conduct no further work on an ECP-based CLI pilot for now

– Pilot works, but ECP will not be available for some time, making deployment difficult

– Investigate alternative solutions for a CLI (supported by CILogon)

• Focus on the Web use case– Understand what WLCG users (LHC experiments) need exactly– Determine how the pilot should interface with existing services

• Very little progress since the last meeting– Support can be provided by the WLCG working group– But experiments will have to do the integration work– Not perceived as an important priority in the short term

20

21

Operational considerations• Identity federation brings more than implementation issues

– Affects security incident handling and security operations

• In the current model– Service providers implement user banning & conduct forensics– IdPs (e.g. certificate authority) perform (almost) no operational

role

• In a federated realm– IdPs implement emergency suspension and also collect essential

traceability information– IdPs will therefore have to provide operational capabilities and

actively participate in incident response

• Providing operational capability is a significant change– Requires careful planning to ensure sufficient logging, expertise,

communication channels, etc. are in place21

22

Attributes• HEP/WLCG has used user attributes for many years:

– Full name + email address

• Base attributes verified with high LoA (passport verification)• Additional attributes are aggregated later

– VO membership and role

• In the future, the same attributes will likely be used– LoA may (have to) vary however– These issues are actively discussed within the IGTF

(http://www.igtf.net/)

• HEP/WLCG needs few and simple attributes– But they need to be fully harmonised across administrative

domains22

IOTA Authentication ProfileTowards Differentiated

Identity Assurance

David Groep, Nikhefsupported by the Netherlands e-Infrastructure and SURFsara

EUGridPMA Kyiv 2013 meeting – 24David Groep – [email protected]

Identifier-only Trust Assurance profile· Cater for those use cases where

· the RPs (VOs) already collect identity data· this RP (VO) data is authoritative and provides

traceability· the ‘identity’ component of the credential is not used

· through an AP where the authority provides only· persistent, non-reused identifiers· traceability only at time of issuance· naming be real or pseudonymous (discussion on

going!)· good security for issuance processes and systems

· and where the RP will have to take care of· subscribers changing name often (in case traceability

at issuing authority is lost)· all ‘named’ identity vetting, naming and contact details

IOTA Questions for RPsSept 9, 2013

Bucharest, RomaniaDavid Groep & David Kelsey

28th EUGridPMA Kyiv – May 2013- 26David Groep – [email protected]

Basic Premises

· Subject names must be globally unique· Subject names must be persistent (for the life

time of the authority)

· It’s about people and robots· Naming of IOTA subjects must be different from

other profiles


IOTA Questions for Relying Parties (RP)

Do the RPs need the ‘real name’ of the person the certificate subject?

· Is it enough to be able to ask another entity to give you the name (e.g. the VO, community, other site or central LDAP)?What assurance level do you expect from this 3rd party?

· Do you need this information up-front (before or at use time)?· Do you need this information at end (e.g. for accounting)?· or It is sufficient to be able to ask an entity to contact the user

on your behalf (like a privacy service)?· Should pseudonyms be clearly identifiable (e.g. if they look like

a name they should be the real name)?· The email use case will not work: since emailAddress must be

embedded in the SAN, but the mail will be pseudonymous as well


IOTA Authentication Profile

· Vetting assurance level and responsibility· Given that the name may be a pseudonym and is

weakly verified, do you (RP, community) have the mechanisms in place to strongly identify the real user?

· Should we enforce obfuscated naming for all subjects?· How will you (RP, VO) deal with ‘identity spoofing’?· How do you currently enrol users in the community? Do

you use or rely on the commonName of the subject for adding people to your databases (or get a ‘warm fuzzy’ feeling in association with e.g. unsigned email)?

· Tracability by the VO· Are you set up to provide traceability for your users?· Do you have means and procedures for incident

response and mitigation?


Auditability, traceability· Access to the user information by RPs

· Do you insist that the collection of this data is verifiable?· Should the chain of processes be documented?· Should the chain of processes be audited periodically (expensive!)?· Should the chain be auditable? Or contract/sanction-based?· A user may and will have many credentials and changing names:

does this pose issues?· Do you expect e.g. the community to provide a real name up-front

with every access request (e.g. in a VOMS generic attribute)?· Do the RPs need to be able to independently trace the

user without involving the user community?· Can a registration mechanism, retrievable or callable by the

RP, satisfy this requirement (e.g. like the EGEE CIC portal having an ability to send email to the owner of a DN)

· Do the RPs need to be able to trace without involving the CA?· Which are the ‘emergency cases’ where they expect CA involvement

in tracing or contacting subscribers?


Incident response

· Do RPs expect the CA to be involved in incident response?· What is an incident where you expect CA involvement?· What is the level of involvement?· Do you see a classification of incidents?· If there are up-stream IdPs, are you ‘happy’ with just the CA

response even if upstream does not participate?· Do you expect more than pure credential revocation in case

of demonstrable credential compromise?· Do you see the CA more than a fall back to point to in case

LEA comes after you?· Do you expect/prefer suspension possibilities? Do you have

this capability at the RP level (through authorization)?· Can you get by with just your own response capability?

Next steps• IGTF discussions (next week)

– What is their role in relation to FIM4R/REFEDS?• Defining best practice for Attribute Authorities, IdPs?• Can IGTF members provide IdP services for Research

• Discussions with others re Horizon 2020– EGI, TERENA, REFEDs, …– Should we create a Federation of Research Communities?

• To join eduGAIN as a single federation• FIM4R likely to propose an Interest Group to RDA

– Next meeting of FIM4R in April 2014 • How should HEP, CERN, WLCG, … “federate”?


Questions?


Federated Identity Management for HEP

Documents

Transcript of Federated Identity Management for HEP