Shibboleth Service Provider V4 Alpha 1 now available

[Scott Cantor]

After roughly 18 months of design and coding, the Shibboleth Project is pleased to announce the first alpha release of the next generation SP software.

In addition to the new documentation available for the Agent and Hub, I have produced a "home page" of sorts for the testing and evaluation of the Alpha release here and it has links to all the important information.

https://shibboleth.atlassian.net/wiki/spaces/DEV/pages/5053906948/SP4+Prelease+Testing

That is your starting point, and lots of questions you might have should be answered there.

As many will be aware, this is a complex undertaking to coordinate as the SP is split into separate components now and they are being versioned and release-managed differently, so this is a little unusual for a prerelease.

We are calling this V4 because that's the easiest way to grasp it, but it really is a new "very slim" Agent implementation that replaces what was the SP (thus the V4) and then a separate set of new IdP plugins that turns that software into a "Hub" for the Agent to use to support SAML or OpenID Connect (or both). It is not a proxy, but it takes mental effort and using it to understand why. If you want a proxy, the idP already is.

We are formally releasing a 4.0.0alpha1 release of the Agent software and packages for it. We are not formally releasing the Hub plugins at this time because installing them as SNAPSHOT releases works, and it allows us to continue producing nightly updates with bug fixes and features.

We'll evolve that page as needed and when new prereleases are done.

We will certainly produce "official" beta versions of all the components in a fixed form in the future, but for the alpha, this was the best solution.

We "hope" to reach 4.0/1.0 this calendar year. That probably means Q1 2027.

Our goal for the alpha release is really evaluation. Finding bugs is great, but we have a long way yet, and the bigger issue is getting this in front of people in a real way so they can honestly understand what this looks like and whether it makes sense for them.

We would prefer most discussion, questions, etc. be kept on the dev list but obviously that's cat herding, it's just what we'd prefer for now.

As far as our willingness to respond to questions goes, obviously we will make some allowances at this stage, but as we get closer to 4.0/1.0, support from me will be more and more limited to members. I'm not apologizing in the least for that; they paid for this work to be done and it was not cheap.

We will be announcing some kind of vehicle for more dedicated help with planning/assessment for members in the Spring timeframe.

Shameless plug:
https://www.shibboleth.net/membership/join-the-shibboleth-consortium/

Anybody using the current SP has a choice to make:

* Seriously evaluate V4. Using it "well" in an organization will usually look quite different from the SP today, though it is possible for it to drop in very similarly for a single SP organization.

* Seriously evaluate something else.

* Fork it. I am not going to maintain that code for any amount of money in the future once it sunsets.

Our goal is to facilitate you all making that decision as soon as possible.

The alpha is quite usable for basic evaluation. It's certainly not done and it's not production worthy yet. It will crash (particularly on IIS).

But it will do SAML and OpenID Connect already, and that's enough to understand how this thing will work.

Enjoy,
-- Scott