Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 76 Next »

This page is about a service, offered by SURF, that makes it easier for research collaborations to set up and manage access to (data and compute) services they need for their projects. Using our service saves those collaborations time and enables them to spent more time on research instead of managing infrastructure. Everybody involved benefits: researchers, providers of services for researchers and institutions. Anybody is invited to read about the service, but it is geared towards Dutch led research projects.

Don't want to read the information below but want to know how this can help you? Or have questions? You can send us an email and we'll set up a call!

Why the SCZ project?

Research is more and more about collaboration, also confirmed in the Dutch NWO 2019-2022-strategy. Researchers that want to collaborate (internationally) and providers of resources who want to offer research facilities to collaborative organisations therefor face the question: how to provide secure access to resources? Researchers have typical access needs that aren't taken care of by the current solutions, and they have documented them in FIM4R-documents (Federated Identity Management for Research). People from around the world have been thinking about how to solve the access issues. In the European AARC-project (Authentication and Authorisation for Research and Collaboration) the specific identity and access challenges researchers face are addressed, and they made a clear video about the problem:

AARC crafted a blueprint architecture that addressed those challenges. SCZ is basically an implementation of that blueprint. The SCZ-project enables you to:

  1. Simply create a group (CO, team) in our Membership Management Service.
  2. Choose and connect the services you need for your collaboration.
  3. Invite your collaborators: as soon as someone accepts the invitation, accounts get created automatically for all connected services.

  4. Easily manage your group: adding or removing a member adds or inactivates accounts in connected services, so only allowed people have access and resources and data are secure and not misused.
  5. Connect to web and 'non-web' services (think of resources accessed via SSH or WebDAV ): with SRAM those can be tied to institutional accounts, improving access revocation.

  6. Work with people from all over the world, either through their institutional account or one of the available guest identity providers SRAM offers.

  7. Simply manage authorization. Group membership in SRAM  is converted to attributes that can be used by the connected services to decide who can do what.

  8. Improve your security by providing step-up authentication (two-factor etc)

No more need for zero hours (nul-uren) accounts that take forever to arrange for and stay in existence far too long and often incur unnessary cost (for licenses for example). Currently, for every new research project the access-wheel is reinvented. Collaborations and research are delayed in the start-up phase because setting up secure access takes time (and IT-expertise). What if there was a plug and play service? SRAM is delivering just that.

We provide an access Authentication & Authorisation Infrastructure-as-a-Service focused on the needs of researchers, research projects and providers of resources for researchers. It takes care of user management. The service name: SURF Research Access Management (SRAM). On these pages we describe what the SCZ project and the resulting SRAM-service is about. 

Besides SRAM, which is tailored for Dutch-led research projects, similar AARC-based initiatives exist, like the EOSC-Life AAI and eduTEAMS.

How does SRAM provide a solution?

  • on the SRAM website researchers can create a collaboration, manage membership, connect services etc.
  • SRAM offers a tool to put the collaboration in control, while allowing others involved enough control as well, while everyone saves time. Who better knows who needs to be in (and out) of the collaboration at what moment? What services everyone needs to use for the research? What everyone should be allowed to do in connected services? Instead of a PI having to email with many people to allow or revoke access, zero hours contracts, many hours wasted on managing access instead of doing research: use SRAM.
  • allowing guests from other institutions access to a resource nowadays often leads to just creating temporary accounts in one or more places. SRAM leverages institutional accounts and the existing global educational federated identity landscape. This improves data security (GDPR!), as access is revoked much faster. We're reusing the global institutional account 'telephone book', eduGAIN, so almost everyone with an institutional account can sign in.
  • institutional accounts also offer better identity assurance: institutions check the identity of their researchers. So when someone logs in using their institutional account, everyone has a high level of confidence this is the intended person.
  • we offer people that for any reason can't use an institutional account, like from a company, guest identity providers so they can use that to sign in and collaborate.
  • we offer mechanisms that, together with federated identity, revokes access to data and resources as soon as possible. While nowadays, people often have access far to long, amongst others because no federated identity is used.
  • providers of services for researcher also save time because they technically have to connect their services to SRAM only once (using open standards like LDAP, OIDC and SAML) and thereafter can easily offer their service to unlimited collaborations and people. Providers can configure and offload simple repetitive fault sensitive user creation tasks, while still being in control over which collaborations are are allowed access etc.

  • institutions can decide how their researchers use SRAM, and are offered more insight in research collaborations their researchers are involved in. Institutions only have to join the SCZ once in order to give all its researchers (via one or more collaborations) access to the participating services and resources. Institutions can use granular authorisation options like authorisation rules and groups (as described in this blog) to limit who can access SRAM with their institutional account, and for instance have their research support office or data competence center create CO's etc
  • we make it easier for service providers to allow people from outside of the Netherlands to access resources.

To get an extra idea of what SCZ wants to offer, here we share the 'user stories' (in broad outline) for which we want to offer a solution with SCZ.

SCZ and Open Access / Open data

Open Access and access regulation mechanisms often go together. Possible scenario's:

  • The research team has every intention to publish lots of data and results at some point, but at the start or during the research, access has to be limited. SCZ can provide for this.
  • Certain data is available for open access, but for all kinds of reasons, certain other data is only available for authorised users (see for instance page 75, 13.2 and 13.4, in 'RDM Toolkit'). SCZ can provide for that as well. 

Schematic overview of the SCZ solution 

Schematically the SCZ can be drawn as follows:

The picture above shows that the research services are linked to the SCZ proxy: these services only have to make and maintain one link. The picture shows the features of the SCZ infrastructure:

  • Connects with eduGAIN so that research services are accessible for researchers at institutions outside the Netherlands.

  • Provides a mechanism (via COmanage) to invite users and manage groups and attributes (a so called 'Membership Management Service').

  • Provides a solution for people without an edu account to use services (such as via Google and / or other social accounts).

  • Link with SURFconext so that researchers at Dutch institutions can make use of the research services via SURFconext and the SCZ proxy.
  • Provides a solution to securely unlock non-web services.

Video and demo you can try yourself

Wondering how a flow of inviting a user to access via SSH looks like? See the below video, but know this is just to get an idea as the environment is developing continuously (if the video doesn't start playing, try opening it full-screen via the icon in the top right corner. The cow-sound at the start of the video is related to the name of the company involved in work on COmanage, Spherical Cow Group of which the name is based on the usage of spherical cow, a humorous metaphor for highly simplified scientific models of complex real life phenomena):

Another way of logging in is shown in a video at the bottom of PAM Module. We've made a connection to Azure AD VM's which we show in this video.

You can also try a demo yourself.

How SCZ aligns with GDPR/privacy

Many federated academic services require a few user attributes to successfully complete login, usually name, email, and a persistent user identifier (called the “R&S attribute bundle”). An international program called the Research & Scholarship Entity Category (R&S) was established to meet this need. This program enables federated services serving a research or scholarly purpose to request that their national R&E federation (as InCommon is for the US) “tag” them with the R&S entity category. It also specifies how R&E federation operators vet such requests to ensure that such tags are only applied to appropriate services.

The R&S program further provides a means by which an academic IdP can automatically release the R&S attribute bundle when users login to services that have been tagged R&S, and a corresponding R&S tag to be given to an IdP to signal that it participates in this global program. This is important because some R&S tagged services will only permit a login to proceed if the user’s IdP is also tagged R&S.

It’s worth noting that releasing R&S attributes under the R&S program contributes to good privacy practice under the European General Data Protection Regulation (GDPR). REFEDS, the international organization of Research and Education Federations, conducted a thorough analysis of how attribute release under the R&S Category addresses GDPR requirements to arrive at this conclusion. 

SCZ only connects services in the R&S category. So IdP's can connect to our proxy, knowing they are compliant to the GDPR in regards to authentication (for processing personally identifiable information (PII) in services connected to our hub, the involved institutions might need extra contractual agreements, which normally are taken care of in the startup phase of research project).

Involved collaborations and institutions

We have a https://wiki.surfnet.nl/display/SCZ/Pilot+partners listing (a part of) the institutions piloting within our project and what is being piloted.

The institutions involved in pilots are expected to participate in meetings and allow the right people within the institution to test the pilot environment, provide feedback to SURF and participate in talks about new features and requirements.

Apart from pilots, we also frequently present about the project, like for the Health-RI event of Dec 8th 2017, where a poster was crafted to show the value of COmanage for collaborations like BBMRI. A generic version:

.

Which technical components are used?

Interested in the components used? See Technical overview of SCZ .

Connecting services

Connecting Services to the SCZ environment describes how to services to the SCZ infrastructure. A list of connected services can be found at https://mdq.pilot.scz.lab.surf.nl/role/sp.html .

Why authenticate in a federated way?

Enabling a service / resource for federated authentication means users can 'login' (authenticate) with their institutional account: as soon as they want to access a service, they are automatically forwarded to the login screen of their institution (or other organisation where they have an account, if that can be used, such as a bank). Reasons to arrange this like this:

  • It provides more reliability
    • As a service / resource you have certainty about the identity
    • If an employee leaves an organisation and may therefore no longer have access to a service / resource, federative authentication ensures that access is no longer possible.
  • It ensures scalability
    • As a service / resource you have no / less work on creating an account, supporting users who forget their password etc
  • It increases security
    • Users can use their (strong) settings institutional password and do not have 'another' account and password to manage
    • Users only have to enter their password on the institutional-login screen known to them (the fewer deviating screens ask for passwords, the less sensitive users are for phishing)
  • It ensures user-friendliness
    • Users don't need to manage extra user accounts and passwords, they can re-use the already known institutional account 

UK JISC has created a video about federated identity:

The European AARC-project has a training-module on what a identity federation is and what its advantages are: 1. AAI Overview.pdf. More information can be found at these websites: Federation-101 and Training for service providers. See also the advantages for IdP's, SP's and users as listed for SURFconext.

Why institutions need to connect

The SCZ can work without institutions releasing attributes; the researcher can use social accounts etc to sign in to the SCZ. But the value for all parties increases when an institution connects its IdP to a service in the SURFconext dashboard; this makes it so the researcher can use the credentials of their home institution, which provides more certainty for the research group and resource providers.

In many Dutch institutions it takes a long time between the request from a researcher to be able to sign in to a research resource with their institutional account and the time sign in is enabled. Sometimes it never happens. Often, the researcher doesn't want to wait that long, and chooses a different solution to get access, which often turns out to be less secure, more costly etc in the long run.

The reason why it often takes that long, is that the employees tasked with connecting a service, want to make sure privacy is protected, IPR is taken care of, whether any financial flows need to be available, whether usage of the service will increase helpdesk calls etc.

Some institutions have decided researchers can carry part of the mentioned responsibility. They have started using a 'light' procedure for connecting research services:

Will this be a SURF service?

SURF has conducted pilots to also answer this question. In May 2019 representatives of institutions advised/voted to develop SCZ into a production ready service. Our team is working on that, and we hope we will have a production ready service in the 1st half of 2020.

Due to our international relations and activities, we know GÉANT has been gearing up a new service, eduTEAMS. Both our teams have been sharing a lot of knowledge, and there are a lot of similarities. We intend to use eduTEAMS as part of our service offering to Dutch research collaborations. A nice feature is eduTEAMS also offers Hexaa and Perun as alternative Membership Management Services to COmanage (GÉANT has a comparison of the 3 systems).

Mailinglist

We have a mailing list for this project. Feel free to sign up for that list via https://list.surfnet.nl/mailman/listinfo/projectscz-fiam . An archive of previously shared messages can be found via https://list.surfnet.nl/mailman/private/projectscz-fiam . Interested? Questions? Suggestions? Mail with Raoul Teeuwen ( raoul.teeuwen@surfnet.nl ). 

If you find the SURFnet SCZ mailinglist interesting, you might also be interested in the following:

"Following some community interest, a new (not COmanage specific) list has been established: cmp-discuss. This is a discussion group for any technologies, policies, or use cases associated with collaboration management platforms, and especially general (non-product specific) topics or topics crossing multiple technologies.

You can join and manage your subscription here: https://groups.google.com/forum/#!aboutgroup/cmp-discuss

(The list was set up as a Google Group to avoid associations with any particular project or community.)"

Planning / timeline / status

In June 2017 phase 1 of the project was completed. Phase 2 ended in May 2019 with institutions advising to develop the result of project SCZ to a production ready service. In phase 1, use cases were drawn up and coordinated with a number of cooperative organisations, an architecture was drawn up and needs were assessed. In phase 2 was dedicated to realising the various components and gaining experience through pilots. 

SCZ phase 3 focuses on creating a production ready services, which includes deciding on a software stack, setting up that stack, contracting sub contractors, having experts conduct code audits/penetration tests, draft relevant contract texts/AUPs etc, design and implement support processes etc

Schedule

  • Aug / Sep 2017 - Establish pilot environment
  • Oct / Nov 2017 - Connecting backend systems
  • Oct / Nov 2017 - Set up and test deployment flows
  • Oct-Dec 2017 - Set up and fine-tune access for external people / guests / etc
  • Dec 2017 - Jun 2019 - Pilot with the pilot environment:
    • Access for "ordinary" (pilot) users
    • Finetuning flows
    • Connect more services
    • Develop the platform
  • Jun 2019 - mid 2020 - SCZ phase 3 (service development)

Support

Support questions and emails can be directed to scz-support@surfnet.nl .

More information

  • No labels