Wednesday, 16 September 2009

iGovt and the Identity Verification Service

This promises some entertainment: the IVS project is not dead yet. A few months ago it looked like it was, but now the government is trying to get people inspired about its world-changing revolutionaryness.

This iGovt thing is in two parts:
  • A logon service: basically you get a single username / password, plus maybe a physical token, that you can use whenever you have to log in to any application using the service. Like the key you get when you buy an apartment.
  • An identity verification service: this lets you prove who you are as a new user of an application. Like the address and passport credentials you show to open a bank account.
The logon part (formerly the Government Logon Service, then rebranded 'iGovt Logon' almost as soon as it was live) has been running for a while, with disappointing takeup - even though you have to get a special ministerial exemption to NOT use it for a new government project.

The IVS part has been in the design stage for ages.

The logon part has limited takeup because it gives an application nothing on its own - there's a big implementation cost without any gain. This is because it's only useful once your physical identity has been established, which needs to be done when you first use each application. iGovt Logon only tells the app that you have an iGovt logon, not who you are - it lets the app see that you are the same user as last time. The app still has to verify, outside the iGovt Logon service, which person the user ID corresponds to, and what access (if any) they should be allowed. In fact, the system is explicitly designed NOT to provide, say, the IRD with an identifier that it can correlate with the same user's account at Housing NZ.

In other words, you only benefit from using iGovt Logon if your users access multiple apps using it, and even then, the benefit is to the users (convenience and security of using a single login), rather than the implementors and maintainers of the apps.

IVS is supposed to solve part of the 'who is this account' problem - it will 'prove' who you are for the initial registration with a new application, and let you share some of your personal details. Annoyingly, it STILL doesn't help with the 'which people should have access and how much' problem: you still need your existing provisioning system for anything that isn't 'everyone in NZ'.

The interesting twist: every project that launches with iGovt Logon in the meantime has a big incentive NOT to use IVS when it becomes available. This is because those projects have already implemented their own alternatives to IVS - typically whatever local user-provisioning system and manual checks those agencies use already. Switching to IVS will require a load of new development effort and will be risky for already-live projects.

There's also a big bootstrapping problem. Unless IVS provides seriously improved fraud detection or financial guarantees (which I doubt), it will never make economic sense for a new project to use IVS until it is fully available and everyone has it. You don't want to implement TWO identity / use provisioning systems for your app just so that some initial-trial IVS users can use their new ID.

Therefore projects will only use IVS if they are compelled to (read: bankrolled from the IVS development budget), or if they are being done by whatever big contractor (IBM, EDS) gets the tender and has a vested interest in pushing it. It will get mandated for government projects if it gets far enough, although that mandate for iGovt Logon doesn't seem to work so well.

No comments:

Post a Comment