Future of digital identity for SURFnet Project : GigaPort3 Projectyear : 2013 Projectmanager : Remco Poortinga – van Wijnen Authors (s) : Maarten Wegdam, Bob Hulsebosch, Martijn Oostdijk, Timber Haaker Releasedate : 17 december 2013 Version : 1.1 final Summary This report is the extended summary of a study that Novay did for SURFnet on the value-adds that SURFnet (or the SURF family) can provide the coming 3 to 5 years in the area of digital identity. This study helps SURFnet to make choices on which value-adds to pursue, and where to invest innovation budgets in. We assume the reader is familiar with the state-of-the-art in digital identities and current services from SURFnet. Determining which SURF family member is most suited for a certain value-add is out of scope for this report. Therefore, when this report refers to SURFnet one could also read “SURFnet or another member of the SURF family” This publication is licensed under Creative Commons “Attribution 3.0 Unported”. More information on this license can be found at http://creativecommons.org/licenses/by/3.0/ COLOPHON Programme line : GigaPort3 Part : SI-Infra Activity : Next generation Trust Infrastructures and Skunk works Deliverable : 2013-401-d1 The Future of Identity – studying the changes in the federated identity and privacy landscape Access rights : Public External party : Novay This project was made possible by the support of SURF, the collaborative organisation for higher education institutes and research institutes aimed at breakthrough innovations in ICT. More information on SURF is available on the website www.surf.nl. ii SIX MATTERS ONE SHOULD KNOW ABOUT FUTURE OF IDENTITY. Scenario The area of digital identity and (federated) identity management is diverse and continuously changing. To ensure the (continuing) usefulness of SURFnet’s services in the area of digital identity in the near future for its constituency, Novay was asked to study the value-adds that SURFnet (or the SURF family) can provide the coming 3 to 5 years in the area of digital identity. What is it? This report provides the extended summary of the findings of the study Novay did, highlighting main trends and possible value-adds. Who is it for? Although initially intended for SURFnet, the outcome may also be of interest to other NRENs providing services in the area of digital identity, or to readers with a general interest in developments in digital identity. How does it work? N/A What can one do with it? Information from this report may be useful for determining the focus of innovation in the (short to middle term) future. More information [email protected] iii PAGE INTENTIONALLY LEFT BLANK iv MANAGEMENT SUMMARY This report presents the outcome of a study on the potential value-adds that SURFnet can provide in the area of digital identity for the coming 3 to 5 years. Digital identity includes but is not limited to identity federation. The reason to do this study is to help SURFnet to make choices on which value-adds to pursue, and where to invest innovation budgets in. Central to the approach is a series of workshops and interviews with internal and external experts and stakeholders, to get a wide range of inputs on what relevant technologies, customer demand and value-add for SURFnet could be. The outcome of the study is: • Identity federation: here to stay and new features are needed – identity federation is a current flagship service for SURFnet. It will remain a value-add for SURFnet, and the study resulted in a list of possible new features to increase the value-add. • eID stelsel NL: use, possibly interfederate but do not replace SURFconext – The eID Stelsel NL is a relatively new initiative from the Dutch government to integrate various eID solutions, including DigiD and eHerkenning. Even though it is not clear if, how, and when eID stelsel NL will become a reality, this is a potential major development in the Netherlands. The outcome of the study is that eID stelsel NL is unlikely to be a suitable replacement for SURFnet’s own identity federation the coming 3 to 5 years. Using eID stelsel NL for verification, however, is an obvious opportunity, and inter-federation with SURFconext may be possible. • Identity-related technologies: what opportunities to select? – There are new technologies in various stages of maturity that SURFnet could use to create a value-add. Criteria, besides the obvious customer demand, are that i) technology is able to combine the - often contradicting requirements of security, privacy and enablement/convenience and ii) if a technology can be deployed because of the combination of services that SURFnet offers, e.g., DNS and certificates. Technologies that came out of the study are: o Certificate pinning: DANE but possibly also others o Anonymous credential systems: go beyond smartcards and explore use-cases o Facilitate a trusted trail for (raw) research data • Privacy: go beyond checklists, provide support and maybe services/software – SURF(net) already works on privacy as a focus point, chiefly by empowering its customers by educating them on what they could, should, shouldn’t and must do. However, SURFnet could choose to do more than provide these checklists and reports, especially SURFnet can provide actual personnel to help its customers. A third possible value-add is to also provide privacy-related services. Whether or not to provide value-add beyond the current checklists, reports and education is mostly a strategic choice, and not only for SURF(net), but also for its customers. • Rich & trusted attributes in a connected world: a new value-add for SURFnet – In a hyperconnected world there is much more value in personal information than the relatively ‘low level’ information exchange currently facilitated by SURFnet’s identity federation. The role of SURFnet could be to facilitate this by being a broker for this trusted exchange of rich attributes. In addition, there could be a role in defining and/or adoption of standards relevant for research & education. v Table of Contents COLOPHON ............................................................................................................................. II SIX MATTERS ONE SHOULD KNOW ABOUT FUTURE OF IDENTITY. .......................................................... III MANAGEMENT SUMMARY ........................................................................................................ V EXTENDED SUMMARY .............................................................................................................. 1 READING GUIDE ........................................................................................................................................................................ 1 THE MAIN TRENDS .................................................................................................................................................................... 1 CATEGORIES OF POSSIBLE VALUE-‐ADDS ............................................................................................................................... 2 IDENTITY FEDERATION: HERE TO STAY AND NEW FEATURES ARE NEEDED ................................................................. 3 SHORTER-‐TERM & CLEAR VALUE-‐ADDS RELATED TO IDENTITY FEDERATION ............................................................... 3 STRATEGIC & NO-‐CONSENSUS VALUE-‐ADDS RELATED TO IDENTITY FEDERATION ....................................................... 4 NEEDS RESEARCH & LONGER-‐TERM VALUE-‐ADDS RELATED TO IDENTITY FEDERATION ............................................. 5 EID STELSEL NL: USE, POSSIBLY INTERFEDERATE, BUT DO NOT REPLACE SURFCONEXT ........................................ 6 IDENTITY-‐RELATED TECHNOLOGIES: WHICH OPPORTUNITIES TO SELECT? ................................................................. 7 PRIVACY: GO BEYOND CHECKLISTS, PROVIDE HELPERS AND MAYBE SERVICES/SOFTWARE ...................................... 8 RICH & TRUSTED ATTRIBUTES IN A CONNECTED WORLD: A NEW VALUE-‐ADD FOR SURFNET ................................. 8 CONCLUDING REMARKS ........................................................................................................................................................... 9 ACKNOWLEDGEMENTS ............................................................................................................................................................ 9 vi EXTENDED SUMMARY The recently published report “25 years of innovation”1 shows that innovation projects from SURFnet and its partners have resulted in reliable and innovative ICT services for higher education and research. An important area of innovations for SURFnet is collaboration infrastructures, which includes an identity federation that SURFnet operates. This identity federation, SURFconext, is a flagship service for SURFnet. Looking to the future, the value-add of SURFnet in the area of digital identity may be in other services than running an identity federation, however. This report presents the outcome of a study on the potential value-adds that SURFnet can provide in the area of digital identity for the coming 3 to 5 years. Digital identity includes but is not limited to identity federation. To quote the aforementioned report on 25 years of innovation: “SURFnet’s innovations are driven by technology combined with customer demand”. But what will be the relevant technologies? And what will be the customer demand? The approach we took is to first analyse trends in the area of digital identity, and then to analyse possible value-adds that SURFnet can provide based on these trends. Central to the approach was a series of workshops and interviews with internal and external experts and stakeholders, to get a wide range of inputs on what relevant technologies, customer demand and value-add for SURFnet could be. The outcome of this study will help SURFnet to choose what innovations to invest in for the coming 3 to 5 years. READING GUIDE We assume the reader is familiar with the state-of-the-art in digital identities and current services from SURFnet. In this report it is out of scope if another SURF family member may be more suitable to provide a certain value-add. Therefore, when we refer to SURFnet one could also read “SURFnet or another member of the SURFnet family”. THE MAIN TRENDS To be able to analyse possible value-adds for SURFnet, we distinguish ten main trends that are relevant for digital identity in higher education and research. These are depicted below, where arrows indicate that one trend influences another. For some trends we make explicit that it is an on-going trend, e.g., “more mobile”. Others are new, for example: “government controlled eID that can also be used by non-government organizations (consumer-2-business, business-2-business)”, but the outcome of which is less clear. 1 SURFnet, 25 years of innovation, 2013, http://www.surf.nl/binaries/content/assets/surf/en/2013/201309_report_25yearsinnovation_surfnet.pdf FUTURE OF IDENTITY PAGE 1 Figure 1: Ten main trends in digital identity for higher education and research. CATEGORIES OF POSSIBLE VALUE-ADDS We use these ten trends to motivate possible value-adds that SURFnet can provide. We grouped these value-adds into five categories: 1. 2. 3. 4. 5. Identity federation eID stelsel NL, which is the new initiative for a government controlled eID framework in the Netherlands Identity-related technologies Privacy Rich & trusted attribute exchange We discuss these five categories at different levels of detail; especially the identity federation category is discussed in more detail since the study resulted in a relatively long list of new features that can be considered for the current flagship service. PAGE 2 FUTURE OF IDENTITY IDENTITY FEDERATION: HERE TO STAY AND NEW FEATURES ARE NEEDED A main question for this study was if within 3 to 5 years operating an identity federation is still a relevant and innovative service for SURFnet. The short answer is yes. And while five years ago, when Novay helped SURFnet to make a roadmap for the SURFfederatie , the hub-and-spoke model was a major area of discussion; this is less the case now2 . The trend appears to be more towards hub-and-spoke federation models and/or use of proxies for specific purposes than to fully meshed models3. There are several opportunities for value-adds directly related to identity federation, which typically mean new features for, or changes to, the federation. Figure 2 below divides these into three subcategories. Figure 2: Further categorization of possible value-adds related to identity federation. SHORTER-TERM & CLEAR VALUE-ADDS RELATED TO IDENTITY FEDERATION Shorter-term features which combine a clear need with a (relatively) clear view on how to implement them are: - Multiple levels of assurance: well-defined and step-up – Cybersecurity is becoming more important, combined with more sensitive SaaS applications and importance of privacy means that the identity federation should offer higher levels of authentication assurance. To balance security with convenience, this should be multiple and well-defined levels (contrary to only higher levels). SURFnet’s value-add can be to define these levels (technology and process), and provide services to facilitate this. SURFnet is already working on a Step-up-authenticationas-a-Service to address this. - Federation to mobile and other native applications: OAuth, inter-app and application specific passwords – SAML WebSSO is not a mobile-friendly protocol, and the identity federation needs to be able to connect to mobile and other native apps in another way. SURFnet has done several studies how to solve this, and OAuth2 is an obvious candidate. In addition inter-app communication solutions that support single sign-on and application-specific passwords can be considered. - Bring your own authentication: less tokens and use social login and mobile as authentication means – Although Bring-your-own-Identity is a trend, converging to “a single trust framework/federation to rule them all” is not likely to happen any time soon. Authentication solutions can be re-used however, between federations and silo-identities to reduce the 2 E.g. in case of large amounts of local logins a mesh model has benefits and a hybrid proxy/mesh model could be considered. 3 See e.g. the (teaser) for the Forrester report “Brokered Cloud Identity Is Scaling New Heights” (Eve Maler a.o., October 2013). FUTURE OF IDENTITY PAGE 3 amount of tokens that someone has to carry and passwords they have to remember4. SURFnet is already taking this route in the design of Step-up-authentication-as-a-Service and the replacement of SURFguest. The latter could become a pattern on how social login can be used, possibly also for different use-cases. This pattern is that social login is basically used as an authentication means (i.e., username/password), without relying on the attributes that are provided (with email address as a possible exception). The benefit is that people do not have to remember an additional username/password and SURFnet is not responsible for storing those. When trusted attributes are needed then these have to be verified separately by SURFnet or a third party. This is also the difference with a bring-your-own-identity solution, which would include trusted attributes like name or affiliation. Additionally, users increasingly own smartphones, which can be used as a something-you-have authentication means in a twofactor authentication solution. Mobile as a bring-your-own authentication means will therefore likely become of primary interest to SURFnet and its customers, especially for students but also for employees. STRATEGIC & NO-CONSENSUS VALUE-ADDS RELATED TO IDENTITY FEDERATION A second sub-category of possible value-adds has, compared to the value-adds listed above, a more strategic character and lacks clear consensus among experts and stakeholders whether or not SURFnet should prioritize these: - Shift control from institutes to users – The trend analysis makes clear that users are becoming more in control and will have a more “loosely coupled” relationship with their institutions than in the past. For example, users will choose what hardware and which cloud services they use, contrary to an IT department. This control shift paradigm impacts the identity federation, which in its current implementation is more identity provider (or institution) centric. Also, on-going efforts such as the Code of Conduct for service providers do little to empower users. An example is that users cannot decide who they want to trust, which attributes to share and what services they can use, this is up to identity providers (and service providers). There may be (lawyers at) institutes that are worried about legal implications of giving users more control, but it is likely that if users do not get more control, they will find ways around the identity federation, for example by relying on social logins. A more user-centric implementation would empower users through self-service to quickly decide whom (not) to trust. Examples are that the user (contrary to an identity provider) decides which service provider to trust with his or her attributes, or a trusted employee that is empowered to vouch for the identity of another and unknown user to enable this user to use some service. - Harden the federation and especially the hub – Increasing importance of cyber security combined with increasing dependence on the identity federation because of its success, means that the federation is increasingly a security risk. Above we already discussed levels of assurance, but here we refer to confidentiality, integrity and availability of the federation as a whole. The federation hub is especially vulnerable here. A successful attack can cause major damage to the reputation of SURFnet and cause major business continuity issues. A choice is whether or not to pro-actively harden the hub. This includes improving resilience to distributed denial of service attacks, adding more monitoring capabilities to detect misuse (including anomaly detection), and improving business continuity during e.g. upgrades. Also on the organizational side it is a choice whether or not to invest in additional security measures like the security at the identity providers (audits, incident management, governance). The first step would be to perform a risk assessment and consider different counter measures on their implications. - A second high-security federation – In general there is a trade-off between security and convenience. This convenience has two sides: ease-of-use for users (e.g. no dedicated software needed on client devices) and for connected parties (e.g., easy to connect to the hub). The current federation balances this, typically favouring convenience over security since for most use-cases convenience is considered more important. However some institutions, including academic/teaching and other hospitals, have significantly higher security requirements, at least when it comes to patient health data (e.g., NEN75105). Although we did not explore this in much detail, we believe there is a possible value-add to better cater for these customers, which may mean introducing a second, higher-security, federation. The first 4 Of course, in a secure manner without more than one organization having access to the same password. 5 NEN7510 is a Dutch standard for information security in healthcare, see http://www.nen7510.org/ PAGE 4 FUTURE OF IDENTITY step would be to further explore the need and implications. - Make the federation itself more privacy sensitive – The federation itself is a privacy risk due to its proxy architecture that allows it to track which users log in at what service provider and to see what attributes are shared with it. Although there was no consensus between the experts and stakeholders that contributed to this study whether this is a major concern and/or potential value-add of SURFnet, we mention it here as a strategic choice. A typical example is to carefully consider what log data to store and how long, since this includes a lot of personal data which can be hacked or misused6. A second example is to encrypt the attributes so the proxy cannot inspect them. NEEDS RESEARCH & LONGER-TERM VALUE-ADDS RELATED TO IDENTITY FEDERATION The third sub-category of possible value-adds related to identity federation are those for which it is less clear if they are feasible, may require more research or will not be realized on a shorter term (e.g., not next year): - Internationalizing trust, but first do ‘plumbing’ – In both research and education the increasing need to federate beyond the Netherlands is clear. To do this requires not only what we call here plumbing, but also trust and agreement on policies between all parties. Plumbing refers to connecting identity providers and service providers (configuration, meta-data discovery etc). There has been quite some effort on large-scale international federations over the past years, especially on inter/confederation (eduGAIN). eduGAIN is successful in connecting NREN federation operators, but actual usage is limited. A major reason for this is that agreeing on policies and creating trust between all parties involved (identity providers, service providers, users, federation operators) on such a large scale is very difficult. Although eduGAIN does more than plumbing, especially through the introduction of a ‘Code of Conduct’7 for Service Providers, we do not expect this to sufficiently address the wider trust issues between all the different parties. An approach to handle this is to separate the challenge of connecting identity providers and service providers, a.k.a., the plumbing, from the actual establishing of trust (a.k.a. the policies). This plumbing can be done as a mesh (with typically some form of metadata registry), as a hub-and-spoke or as a hybrid form. Once the connections are there, it becomes easier to create different and smaller trust frameworks for specific purposes/communities that can run on top on this plumbing. The longer-term ambition could remain to standardize these trust frameworks, but to get there “perfect is the enemy of good” is sound advice. - Cross-sector federation – Many research collaborations involve organizations that are not universities or research institutes, i.e., they cannot connect to SURFnet’s identity federation and thus their employees cannot use their organizational account to authenticate. Cross-sector federation overlaps with earlier described possible value-add internationalization because it requires ‘plumbing’ to connect identity providers at a wider scale and because it is very difficult to scale the needed trust. It also overlaps with the possible value-add of shifting control from institutes to users, described earlier, because putting the user more in control may help to scale the trust. Existing and often more informal trust fabrics may be re-used, e.g., in research projects (also known as collaboration organizations), reputation and social trust (e.g., social networks). First steps could be to select use-cases, study state-of-the-art and find ‘launching customers’. There are two other possible value-adds related to identity federation that we want to mention, but do not fit very well in the categorization used above. These are phishing prevention and single logout. With respect to phishing prevention, there is little doubt that there is a need for this, but how to do this is less clear as was concluded in a study on trust mechanisms for server authenticity that was conducted by Novay earlier this year8. With respect to single logout, it is clear users want this and that it is better for security. However, how to implement this is not clear. At a minimum SURFnet could consider how to better educate users on single logout (i.e., close all browser windows). 6 SURFnet could also be legally forced to provide access to this log data, e.g., when a student is a suspect in a criminal investigation. 7 Code of Conduct, http://www.geant.net/uri/dataprotection-code-of-conduct/v1/Pages/default.aspx 8 Martijn Oostdijk and Maarten Wegdam, Exploring Innovation in Trust Mechanisms, SURFnet/Novay, June 2013, Available from: http://www.surfnetters.nl/rijswijk/doc/2013-404a-TS-NGTI-Exploring%20Innovation%20in%20Trust%20Mechanisms1.0.pdf FUTURE OF IDENTITY PAGE 5 EID STELSEL NL: USE, POSSIBLY INTERFEDERATE, BUT DO NOT REPLACE SURFCONEXT The eID Stelsel NL is a relatively new initiative from the Dutch government to integrate the various eID solutions that citizens and companies can use to access government e-services, including DigiD and eHerkenning. It will provide an overall framework for trusted digital identities in the Netherlands, including consumer-2-business authentication. Part of eID stelsel NL is a more secure DigiD, called DigiD Hoog. This is a contactless smartcard, distributed via city halls. In addition to authentication, eID stelsel NL provides delegation (in Dutch: machtigingen) functionality, allowing a person or organization to authorize another person/organization to act on their behalf. It also provides basic attributes, but it is not clear at this point if these would be accessible for nongovernment organizations (and if so, which ones). Even though it is not clear if, how, and when eID stelsel NL will become a reality, this is a potential major development in the Netherlands which is gaining momentum. Figure 3 below depicts the choices for SURFnet. It lists the short-term actions, and options how SURFnet’s identity federation could relate to eID stelsel NL. There was a clear consensus among experts and stakeholders that replacing the SURFnet identity federation with eID Stelsel NL is not a realistic option, at least not the coming 3 to 5 years. A main reason is that it isn’t likely to address NREN specific needs, such as certain identity attributes and internationalization. Also users may not want to use eID stelsel NL for work-related services. Using eID stelsel NL for identity vetting however is an obvious opportunity, and inter-federation with SURFconext may be possible. Providing more details on the different options is premature at this point in time, given the early stage of development of the eID stelsel NL. Figure 3: eID stelsel NL and SURFnet’s identity federation. PAGE 6 FUTURE OF IDENTITY IDENTITY-RELATED TECHNOLOGIES: WHICH OPPORTUNITIES TO SELECT? As mentioned before, SURFnet’s innovations are driven by technology combined with customer demand. There are new technologies in various stages of maturity that SURFnet could use to create a value-add. We first discuss which criteria could be used to select which technologies to invest in, and then discuss specific technologies. Of course, the main criterion is if there is an actual need for the technology. Additionally, a technology that is able to combine the - often contradicting - requirements of security, privacy and enablement/convenience has a plus. Identity federation is an example of a technology that combines security with enablement. A third criterion is if a technology can be deployed because of the combination of services that SURFnet offers, e.g., DANE since SURFnet is both a DNS provider and a certificate authority. Below are the technologies that came out of the study as most relevant. Some of these were also part of the trust mechanisms study from earlier this year8 and not necessarily about identities of persons but also (or more) about identities of machines or organizations. - Certificate pinning: DANE but possibly also others – Especially since the Diginotar incident it is widely recognized that the current system of certificate authorities has its limits since a security incident in any of the roughly 6839 certificate authorities can rapidly spread throughout the Internet. Certificate pinning does not replace the current system, but rather improves trust in the system by providing extra information to the end user and/or ‘machine’ about a certificate. There are different ways to do this, including DANE, certificate transparency and pinning in the network (for details see the trust mechanisms study8). DANE builds upon DNS to provide extra certificate information, and is probably the most mature. SURFnet is also already working on DANE. Certificate transparency is less mature, but may be interesting; SURFnet is also already working on this in collaboration with other NRENs, Google and large Certificate Service Providers in the context of a GN3+ task. Pinning in the network is an idea that came up while doing the trust mechanisms study and there is probably the least mature. However it is also the easiest to deploy since - contrary to the other certificate pinning mechanisms – it does not require changes to clients. Future steps are to explore these mechanisms and assess their feasibility. - Anonymous credential systems: go beyond smartcards and explore use-cases – The last few years there have been advances in anonymous credentials, which, compared to traditional certificate and federative identity systems, are more privacy friendly since the issuer of the certificates cannot see which service providers the user visits and service providers cannot track users. Examples are Idemix (by IBM) and Uprove (by Microsoft). SURFnet (and Novay) participate in the IRMA smartcard project, which is a research project to prototype and pilot this kind of technology. Participating in research projects to explore this technology is good for the image of SURFnet (innovative, combine security with privacy) but especially in the external workshop the comment was that smartcards may not be the most relevant application for higher education and research, and that motivating use-cases are needed. SURFnet is already addressing this by pushing for and performing research on secure alternatives to smart cards. - Facilitate a trusted trail for (raw) research data – There are several reasons why there is, or at least seems to be, an increasing need to provide a trusted trail for research data, typically using signing, time stamping etc. These reasons are the recent incidents of academic fraud in the Netherlands, the (many) devices that produce research data (one of the ten trends described above) and the desire for more sharing of research data. This is something SURFnet could facilitate. The next steps could be to further explore if there is indeed a need and if so look into existing standards and tooling. 9 Zakir Durumeric a.o., Analysis of the HTTPS Certificate Ecosystem, IMC’13, October 23–25, 2013, Barcelona, Spain. http://dx.doi.org/10.1145/2504730.2504755. FUTURE OF IDENTITY PAGE 7 PRIVACY: GO BEYOND CHECKLISTS, PROVIDE HELPERS AND MAYBE SERVICES/SOFTWARE Even though not everyone agrees how much users actually care, the importance of privacy is an undisputed trend. SURF(net) already works on privacy as a focus point, chiefly by empowering its customers by educating them on what they could, should, shouldn’t and must do. However, SURFnet could choose to do more than provide these checklists and reports, especially SURFnet can provide actual personnel to help its customers. For example, a part-time privacy officer that is both trusted and has dealt with similar challenges at other institutes. The market will increasingly offer this as well, but likely these professionals are not specialized in higher education and research, and are less trusted than someone from SURFnet. There was less consensus in the different workshops and interviews if SURFnet should also provide privacy-related services, for example, secure storage of personal data or as a trusted third party provide anonymisation of personal data10. SURFnet may have, or can easily acquire, the expertise to properly run such services compared to the institutes themselves. However, by offering such services the institutes might feel less responsibility themselves, and it may be too risky for SURFnet. An alternative to offering such a service is offering software that each institute can run. Whether or not to provide value-add beyond the current checklists, reports, and education is mostly a strategic choice; not only for SURFnet but also for its constituency. RICH & TRUSTED ATTRIBUTES IN A CONNECTED WORLD: A NEW VALUE-ADD FOR SURFNET The current identity federation facilitates the exchange of personal data, but in quite a limited manner: it is ‘low level’ information, such as name or email address and it is (mostly) at login time only. In a hyperconnected world there is much more value in personal information11 than is currently facilitated by SURFnet’s identity federation. This requires a trusted flow of attributes from, to and within the R&E sector. Examples are: - Student John Doe successfully completed his course Cybercrime at the University of Nijmegen, worth 4 ECTS. A virtual life-long educational dossier. Dutch government certifies residence status of a foreign student. Central registry certifies that someone is a medical doctor. These are ‘richer’ attributes or can even be complete dossiers. They have more complex semantics than what is exchanged now in the federation. Also this information is not needed at login-time only, but typically should be available through a back channel, which the current federation does not facilitate or can provide user control for. Important requirements are i) that the attributes are trusted and ii) that there is transparency and consent for users. With respect to trust it is important that the provenance of the attributes is clear (authentic source), and also privacy law requires transparency and traceability of personal data12. With respect to transparency and control for users it is important that users are empowered for this and there needs to be a benefit for the users, and not only for the service providers. The role of SURFnet would be to facilitate the above, to be a broker for this trusted exchange of rich attributes. In addition, there could be a role in defining and/or adoption of standards of attributes that are relevant for the research and education sector. Note that there is already an international initiative for sharing student data (e.g. on courses and diplomas) across international borders, the Groningen Declaration13; it is worthwhile to investigate whether SURFnet can play a role in facilitating the kind of data exchange envisioned by the participants in this initiative. Possible solution directions are personal datastores (compare to Qiy14), centralized consents (like UMA15) or a virtual dossier (in Dutch: verwijsindex, like Verwijsindex Jongeren or Landelijk Schakelpunt). 10 An example is “Parelsnoer”, which provides anonymization for academic hospitals to be able to share medical data for research, see http://www.string-of-pearls.org/ 11 Boston Consultancy Group, The value of our digital identity, Dec 2012. 12 See Article 9 of the Wet Bescherming Persoonsgegevens. 13 See http://www.groningendeclaration.org 14 Qiy – www.qiy.nl PAGE 8 FUTURE OF IDENTITY CONCLUDING REMARKS This study resulted in a long list of potential value-adds for SURFnet in the area of digital identity. The next step for SURFnet is to prioritize a short list to invest in. At the start of the study we considered the possibility that SURFnet’s identity federation may become less relevant the coming 3 to 5 years, but the outcome of the study is that it will remain a flagship service for SURFnet. The study also resulted in a list of new features that SURFnet can consider, including confirmation that some ongoing innovations for or related to SURFnet’s identity federation indeed will provide a value-add (step-up authentication as a service, federating mobile apps). With respect to eID stelsel NL there is potential, but the consensus is that there remains value-add for a federation for higher education and research. Considering the potential, eID stelsel NL deserves attention from SURFnet (and higher education and research in general). With respect to identity-related technologies the study confirmed that certificate pinning technologies (DANE, certificate transparency and pinning in the network) are relevant for SURFnet. Facilitating a trail of (raw) research data however is a possible value-add that was not already considered by SURFnet. With respect to privacy the outcome is that there appears to be a need with institutes for expertise and/or services, but it is more of a strategic choice if and what role SURFnet (or SURF in general) wants to play here (services and/or organizational/personnel). Possible the biggest ‘new’ value-add that came out of the study is a role for SURFnet as facilitator/broker for the exchange of rich and trusted attributes. There is a lot of potential value here for all stakeholders and SURFnet can ensure that this value is realized in a privacy-sensitive and user-centric manner. ACKNOWLEDGEMENTS Quite a lot of experts and stakeholders provided input to this report during one of sessions/workshop or because we interviewed them. This is an incomplete list of persons: Dave Kearns, Serge van den Boom, Hans van Looy, Maarten Stultjens, Klaas Wierenga, Alf Moens, Evert Jan Evers, Stefan Suurmeijer, Gertjan Filarski, Ronald Dekker, Meine Veld, Hans Zandbelt, Victoriano Giralt, Ken Klingenstein, Diego Lopez, Harold Teunissen, Niels van Dijk, Pieter van der Meulen, Roland van Rijswijk - Deij, Erik Huizer, Floor Jas, Alexander Blanc, Eefje van der Harst, Joost van Dijk. Remco Poortinga - van Wijnen was our main contact at SURFnet for this assignment and provided input during all phases of the work. 15 User Managed Access – http://kantarainitiative.org/confluence/display/uma FUTURE OF IDENTITY PAGE 9
© Copyright 2024 ExpyDoc