Answers to the unanswered questions from the webinar

Introduction

Last Friday on Sept. 14, Pamela Dingle—Sr. Technical Architect from Ping Identity Corp.—and I conducted a free webinar about the much ballyhooed demise of SAML.

You can view the webinar in its entirety on the KuppingerCole website.

To us, the best measurement of interest in any given webinar is the drop off rate. Just how many people drop off during the presentation? We were very pleased in the interest of the topic for the number of attendees and for that fact that no one dropped off from the presentation and Q&A.

However, we did not have the time to answer all of the questions presented. The following is a sequence of questions and answers that were left open.

It could be a little disorienting to read this Q&A if you didn’t attend the webinar, I recommend watching the webinar first to avoid any confusion or misunderstanding.

Webinar Questions and Answers

Q: Since the organizations are still not migrated entirely to API, i.e. still we have web browser based applications. So my question is instead of implementing different solutions one for browser based applications and one for API. Do you suggest a common way to support both the users? Thanks

A: Using APIs does not preclude using the browser to access the information and resources provided by the API. In fact, using the browser for API access is quite common. The sub context of this presentation is that it is not limited to the request-response browser model that we know and love for traditional applications. We are now moving beyond the model to an interactive model.

Q: As a follow up these companies could help us “leap frog” to newer protocols very quickly much like some countries skip the notion of “land line” because it’s easier to deploy cellular.

A: Great metaphor. Indeed the combination of RESTful API interface (HTTP), OAuth, JSON, UMA, SCIM, and webhooks are the technologies that I think are the leapfrog technologies.

Q: Many companies are outsourcing IT functions to outside providers, at what point do we just take this to the n-th degree and just let an org like Google or Apple handle identity for us? Is that too scary?

A: I think the answer lies in a simple question, is it the vendor that manages your identity your customer, or are you their customer. If the answer is the latter, it is very scary indeed. As long as we have the expectation of having Identity Management be free, and act as customers of the vendors that provide that service, they will be monetizing our identities to pay for the service. It will be up to the corporation or individual to choose which direction to take.

Q: What about devices not directly linked to people? I.e. do you have numbers that include the Internet of Things?

A: I tried to keep the numbers focused and understandable. Including inanimate and non-digitized items just increases the whole argument. Look for more info on numbers in future conversations.

Q: Have you considered the impact of the availability of global identities on the problem you sketched?

A: I don’t think the availability of a global identity reduces any of the issues in the arguments. Global identities—assuming it will ever happen—just compounds the problem.

Q: Ok, Craig, how do you deal w/ 2.8B identities - who numbers them? Who vets them? What fraud is possible? What is the metasystem - and does it really matter whether it is OAuth/SAML/OpenID?

A: This is a multipart question and I will answer them in turn. First off it is 28 billion and not 2.8. 1). Different organizations—both open and private—will number these entities. 2). Some of them will be vetted and some not. This becomes a big problem we are still grappling with, especially when no single Identity Provider can even be considered to be the validation resource for even a fraction of the entities we are talking about. Look for more information on Trust Frameworks to understand more on this topic. 3). Yes, fraud is possible. Fraud will always be an issue. It needs to be minimized. I think we are on an encouraging course to resolve these matters. 4). The only Metasystem proposed so far is the Identity Management as a Service architecture being designed by Kim Cameron at Microsoft in the form of Azure Active Directory. 5). Finally, in the end the protocols won’t matter just as the argument of CSMA vs Token Ring no longer matters. We will simply moved up the stack. It gets a little more complicated at this level because there are no more layers in the stack to move up to. This is all layer 7 stuff. Layer 7.5?

Q: Will you to be talking about this at IIW 15?

A: I am registered for IIW 15 and plan to attend. I will coordinate with Pamela to see if we can repeat this session during the conference.

Q: Just want to echo Pam's point that the combinatorial explosion is over estimate. Not all users & devices will connect to all services. The real world ecosystems sees users congregate in niches.

A: I think the combinatorial explosion is an underestimate. Pam’s soft pedaling of the numbers are still staggering. If you recall, she thought that most organizations could look at the provisioning of devices in the 1000s or 10s of thousands. OK. To date, anything over 150 starts to create huge administrative overhead. This is not going to go away or be minimalized by downplaying what has already happened. 400M iOS devices alone. The numbers are staggering.

Conclusion

Thanks for the great questions and participation. I look forward to seeing people at IIW. I encourage anyone who attended this conference to attend IIW and the EIC next May in Munich.