Federation and eGovernment

Digital Identity is a kind of personal passport for the digital market space. But not only people require passports; machines, merchandise and even whole organisations or government agencies must be positively identified and described, too, or else core processes will break down. That why Identity Management is arguably the fastest-growing segment in today’s IT landscape. According to Frost & Sullivan it will grow by 45% a year until 2006, reaching a sales volume of some $1.1 billion in Western Europe alone by 2007.

Identity is central to any discussion about eGovernment. Unfortunately, though,  nobody seems to know exactly whose responsibility it is to manage the necessary complex and expensive infrastructure. Generally speaking, the debate seems to focus on two distinct and diametrically opposed strategies: centralism versus fragmentation.

Of course, this is nothing new. Political theorists have been discussing the division of power, between central governments and state or local authorities for centuries. And in fact almost every nation state has found its own more or less unique answer to the questions implied. France since the days of the Louis’ has been a highly centralized state with Paris as its hub. In the United States the disagreements about States’ Rights had to be settled by a Civil War and is still highly contentious. Politicians in Germany recently failed yet again to sort out the hopeless tangle of responsibilities and jurisdictions between Berlin and its 16 Laender, each jealously guarding its ancient prerogatives. If Charles de Gaulle once declared it impossible to govern a country that produces 246 kinds of cheese, then pity poor Gerhard Schroeder faced with a phalanx of powerful and self-governing regions and cities. Should you visit me in Munich one day you will actually pass border signs welcoming you to the “Free State of Bavaria”.

Kuppinger Cole + Partner would respectfully suggest that, in the context of eGovernment, both centralism and fragmentation are the wrong strategies. We believe the only way eGovernment will ever really fulfill its promise of cost reduction and improved quality of public services is through the concept of federation.

In the age of the Internet, the public and private sectors are more closely interlinked than ever before, and this is still early days. Like it or not, more and more information is being exchanged electronically between public administrations, individual citizens and corporations. The totally networked economy is rapidly giving way to the totally networked nation (itself of course just part of a totally networked planet). For politicians and civil servants the world over, the question of how to digitize internal and external processes is at least as acute as it is for managers. In fact, eGovernment today is one of the major growth areas and a favorite target for an IT industry still dodging the fallout of from the bursting dotcom bubble.

In Germany, for instance, the first halting steps towards Internet-based public services date all the way back to the early Nineties. Cities like Bremen, Mannheim and Munich were already keen on building themselves a “glass city hall” (“glaesernes Rathaus”) back in 1995. These were rather primitive, builetin-board type websites, and most of the information could be found in the local newspaper, too.

Today almost every eGovernment initiative involves some form of data transfer. More often than not, this data is sensitive, either in terms of security or privacy. Therefore they must comply with a raft of local, regional, national and international regulations. And to make matters worse, these rules is not only patchy but frequently contradictory or even mutually exclusive.

But eGovernment must comply with another set of statutes entirely, too – the rules of the market. That means they must be user-friendly, they must promise tangible and easily understandable benefits, and they must be very secure. In the end, eGovernment is all about trust, and that makes it so hard. Trust cannot be regulated into existence, it has to be earned.

Like early eCommerce initiatives, early eGovernment projects were marked by a sense of excitement or even euphoria. Finally we would really change the world, or at least that part of it concerned with serving the public.

And like their counterparts in business, the pioneers of eGovernment followed the famous Internet dictum: “Think global, act local”. These regional, local or even neighborhood projects were fragmented in the extreme, both in terms of content and technology.

Today, as we enter the second phase of eGovernment initiatives, the pendulum seems to be swinging back. There is much talk of standardization and transregional coordination, not to mention a desire for a national or pan-European approach. Thus, the eGovernment community today seems neatly divided between regionalists, supporters of the Old Order, on the one hand and centralists on the other. Their discussions are at time bitter and highly politicized. And in the meantime the confusion of mutually incompatible and/or redundant systems and applications grows and grows.

The  split between centrally organized and federalized states within the European Union is accurately mirrored in member states’ wish for more or less standardized and centrally organized eGovernment solutions, at least when regional, national or international concerns are being addressed. On the local and municipal level, its still everyone for himself and the devil take the hindmost.

But whether centralized or decentralized, every single eGovernment system has to cope somehow with the problem of identity. It must be provable exactly who did what, when, with whom and with which legitimization. If not, then, in the world of public administration at least,  such systems quickly bump up against the wall of feasibility – and legality. Purveyors of eGovernment, even more than their counterparts busy building eBusiness or eCommerce solutions, need to agree on open, interoperable and decentralized standards of identity management, and they need to do it fast.

Instead, we are witnessing an accelerating fragmentation in eGovernment. In Germany, central government initiatives such as “Deutschland online”, founded 2003 to replace the rather homely BundOnline portal, bills itself as the “common eGovernment strategy” of national, state and local authorities, but in practice each continues to “cook his own soup”, as the German saying goes. In reality most centralized systems fail to address, much less resolve key questions such as:

- how to improve interoperability while maintaining autonomy of systems run by different agencies at different levels and within different jurisdictions
- how to shorten response and processing times internally and externally while simultaneously reducing administration costs
- open up administrative processes so that citizens and companies can access them and input data directly without compromising security or confidentiality.

If, as is the case in Germany, anyone with a computer can access your tax files via the officially sanctioned online software “Elster” by simply entering your tax number and pressing “enter”, then something is seriously wrong.

The U.S. as the home of the Internet has more experience than most with identity management in eGovernment, most of it accumulated through trial and error. Last year, after long and painful deliberation, officials finally agreed not to agree on a single centralized authentication for users of eGovernment systems across the nation. Such a system would have required building a database of truly Orwell-like dimensions. Given that it took until this year for U.S. officials to finally agree that a driver’s license should actually bear a photograph of the bearer, and that any mention of introducing a national identity card invariably brings the National Rifle Association out shooting, the chances of establishing a centralized identity depository for eGovernments were recognized to be nil.

Instead, the United States may very well be the first nation to opt for a federation-based approach. The charm of federation in identity management lies in the fact that the identity information itself always remains where it was. Instead of creating mega/meta/hyper or whatever directories, federated identity relies on data standards and on sharing of information between peers.

This requires the organizations involved to form so-called “circles of trust” based on technical standards and common rules and contractual agreements. These circles of trust may involve companies, as say when Visa, Lufthansa and Avis agree to allow their customers access to their mutual online systems. They may also involve authorities from various levels of government or even across borders. And they can in fact be public/private partnerships with administrations and companies cooperating to better server their citizen customers.

To enter such a circle of trust, the citizen customer requires a method of authentication. This can be a digital signature if such a system exists. It could be biometric. It could also be a simple username and password system, although even Bill Gates now admits that all password systems are inherently unsafe. (Microsoft is currently shifting internally to a smartcard system).

In a federated system, the information about who is allowed to do what remains where it is – in the computer system of the authority that owns it. Whenever necessary, another administrative entity belonging to an established circle of trust can “borrow” the information, so to speak, but it may not store it, and in fact the system itself will foil any attempt to get around this provision.

Federated identity is still a new idea even in the U.S., and it has been greeted by many misconceptions. Sharing, it seems, is not something bureaucrats relate to naturally. In fact the concept is technically very far advanced, thanks mostly to the Liberty Alliance, a voluntary association of more than 150 large corporations from the IT and telecommunications industries as well as various national, international and transnational organizations such as the U.S. General Services Administration, various national postal services, the Austrian government, and the U.S. department of defense.

The Liberty Alliance does not aim at creating unified technical solutions for Federated Identity. Instead it is trying to achieve consensus on standards and descriptions for technologies devoted to enabling users to share identities in a timely, beneficial and legal manner.

A word about the competition: Microsoft as usual is trying to have things their way. Instead of joining the Liberty Alliance, Redmond has set up a rival consortium – a very exclusive one, I might add, seeing that it consists to date of Microsoft and a handful of other companies. Two major WS_* backers, Intel and IBM, recently joined the ranks of Liberty: IBM especially had been pressured by customers to offer Liberty-compliant products. Microsoft’s WS_*- approach is in fact in some ways more simple and direct than Liberty’s, but this mearly raises the question: Why do we need two competing federation standards anyway? Ultimately, Kuppinger Cole + Partner see some kind of marriage of the two system that will allow Microsoft to bring in parts of WS as a dowry, with Liberty maybe even agreeing to a change of name as part of a face-saving deal.

Whatever happens, Kuppinger Cole + Partner recommend: don’t wait for Microsoft! Go with Liberty now and see what Redmond may or may not have to offer a few years down the road. Chances are it will be Liberty-compliant anyway.

In any case, the secret of successful federation lies in interoperability: Customers who buy products from developers or manufacturers conforming to Liberty’s standards can rest assured that they will integrate with any other product similarly endowed. That means that a user entering a circle of trust through a Liberty compliant system will be duly recognized and authorized by systems from other vendors that also work along Liberty lines. Said user can access information and perform transactions within multiple systems without having to authenticate himself repeatedly.

At the same time, the user is in complete control over which items of information will be shared by whom and for what purpose. The user may chose to allow certain participants within the circle of trust free access to his or her personal information and restrict or exclude others. In an eGovernment environment, you may have to give your consent in order to use the system, or your consent may be mandated, but in every case the user will be able to monitor who is allowed to use his or her personal information or exchange authorization attributes. And of course each transaction will leave an audit trail.

Warnings that Identity Federation might lead to even worse violations of privacy are either frivolous or malicious. Federation is in fact the surest way to keep sensitive information out of the hands of hackers and hoodlums. And it provides a neat answer to the ancient worrying question: “Quis custodiet ipos custodes” – who will protect us from our protectors?

From an administrative perspective, Identity Federation offers many benefits ranging from cost reduction to process optimization and increased staff productivity. And last, not least, it translates into increased citizen participation and satisfaction. For government authorities suffering from deep budget cuts and growing criticism for their alleged unresponsiveness to the needs and wishes of the public, this would seem almost too good to be true.

In fact, many of these benefits can be demonstrated by simply adding up the numbers. In terms of IT costs, password resets are responsible for up to 80 percent of all support calls, according to a Gartner Group study. Since almost every government employee today needs to be able to access a growing number of computer applications, each one protected by its own little password system, keeping passwords current can be a major nightmare for your tech support people and a constant money drain.

In the longer term, benefits will derive especially from improved communication between different agencies and across borders as well as with citizens and private companies. Let’s quickly explore some of the more likely scenarios:

Communication between government agencies:

State and municipal organizations are increasingly forced to exchange information more quickly and seamlessly. Interoperability is therefore a central issue in choosing IT systems. A federation-based infrastructure allows exchange within a rapidly evolving and expanding coalition or circle of trust while maintaining the autonomy of each individual participant. This kind of improved communication can literally become a matter of life and death, for instance when a regional natural or man-made disaster such as a train or airplane crash – or a tidal wave – can require instant and improvised cooperation between authorities in different countries or even on different continents. When nerves are raw and time is short, highly sensitive identity information must be allowed to flow without compromising data security or individual privacy.

But it doesn’t need a tsunami to raise the issue of free information flow. The European Commission has identified a number of measures and goals in such diverse areas as eHealth, eLearning and eBusiness, which are to be addressed under the “eEurope” program. In each case, improving public service involves ensuring that identity information is shared in a timely and secure fashion.

Communication between authorities and citizens:

Customers expect manufacturers, retailers and service companies to react faster and more individually to their needs and demands than ever before. The same goes for the citizen customer in dealing with the suppliers of public services. Citizens in fact see themselves increasingly as the customer of their governments and expect to be treated as such. Complicated or incomprehensible forms, long waiting lines at public service outlets and being required to repeatedly enter the same information (or password) is not only annoying, it also translates into estrangement with the state and reduced civic participation. Also, lack of transparency and citizen disillusionment are grist for the mills of the critics decrying the alleged waste of public funds.

Governments in different countries have gained experience in eGovernment projects, but the level of maturity and development varies significantly. Germany was the first country in Europe to pass laws providing the framework for digital signatures, but lack of standards and bureaucratic inflexibility have defeated attempts to actually introduce them, but then there must be a reason that half of all legal textbooks in the world are written in German. Italy has pressed ahead with digital identity cards, but unfortunately the technology is proprietary, turning Italy for all intents and purposes into an identity island.

Giving citizens direct access to information and services provided by both state and semi-state agencies is a sure way of opening these systems up to misuse. Federation is a much better idea since the actual information remains in safekeeping; no physical transfer of sensitive data is allowed to occur.

In a federated eGovernment environment, one government agency can provide authentication and thus enable other agencies to provide personalized services such as tax returns, issuing of passports or drivers’ licenses any way up to online voter registration or even the electronic ballot without the hassle and cost of multiple authentication.

Communication between authorities and private companies

When business and government interact, interoperability and privacy invariably become central issues. This goes for approval processes as much as for applications and tax returns. Both sides stand to profit from even incremental improvements in these everyday processes. A federated infrastructure would provide companies with a single portal through which they could access a wide range of administrative services, making the business of dealing with the authorities faster, more reliable and cheaper. It would also enhance the security and trustworthiness of systems involving companies and authorities in different countries such as international procurement or research projects.

The Japanese provide an interesting example of how such a system might work. As part of their “EduMart” project, part of the “eJapan Policy Priority Program” aimed at improving cooperation between education authorities and the private sector, an open interface was created, using federation standards, to give students of more than 40,000 schools and universities access to valuable educational materials provided by both public and corporate sponsors. In this case, it was crucial to ensure that sensitive information such as school grades cannot be accessed without authorization, while other information must remain free for all. This was achieved in very short time with the help of a so-called single sign-on solution  based on Liberty Alliance standards and Identity Federation tools from multiple vendors.

Matching identity federation and eGovernment seems to be a non-brainer. However, there remain a number of unsolved issues to be addressed by both public and private stakeholders. For instance:

What is an identity? Depending on which public agency the citizen is dealing with, he will be perceived from a totally different perspective. Often, identity is connected to a set of numbers, such as a passport number, a social security number, or a tax number. Who should coordinate these different identities between agencies “A”, “B” and “C”? And aren’t we actually creating a backdoor to even more centralized systems that way? The only thing we know for certain is that is we unless we achieve a seamless user experience in dealing with multiple government agencies, fragmentation will prevail. From a citizen’s standpoint, things will have remained the way they always were.

Who can I trust? In a federated environment it seems natural for public and private organizations to cooperate in questions of identity. In Germany, the “Bündnis elektronische Signatur” or „Digital Signature Alliance“ has brought together federal ministries, major banks, IT corporations and the state controlled pension authority BfA to develop common standards and hopefully get the stalled process restarted. But while as a customer I may trust my bank more than I trust a government agency, and therefore may be more willing to give a bank unrestricted access to my personal data, the details can prove very tricky if personal information from government sources is involved or when government agencies – think: tax authorities – request access to information I give my bank.

Who do I sue? Given that we live in an imperfect world, things are bound to go wrong occasionally even in a federated environment. Liability issues have to be threshed out in detail before federation will work in practice. This is not a question of technology but of contracts. It remains to be seen what these contracts will look like – lawyers have no previous experience to draw on, but I’m sure that won’t stop them. Organizational issues have to be resolves, too. Who, for instance, will run (and pay for) the necessary helpdesk services for a federation single sign-on website? This is also probably not an issue to be dealt with on a purely national level, but within the context of the EU or even the United Nations or the World Trade Organisation.

At least initially, federated eGovernment initiatives will most likely remain purely internal, so only public servants will be involved. This is good because it will allows the administration to get the knots ironed out first before expanding the system to include private sector organizations. But bear in mind that the true value of federated systems will only show themselves when such public private cooperation is put in place.

So, if federated identity is the answer to a litany of prayers from eGovernment proponents, when can we expect things to really start happening?

Unfortunately, that’s hard to predict. Since eGovernment, as we have seen, is less about technology and more about people, contracts, circles of trust, identity and federation, it is essentially a political issue. And don’t hold your breath waiting for our politicos to jump on the bandwagon.

Instead, go out and start spreading the word. Federation will be a grass-roots movement. Circles of trust will flourish and grow almost organically.

Kuppinger Cole + Partner also strongly advise looking very closely at what the Liberty Alliance is doing. Public institutions and agencies should actually think about joining the Alliance as members in order to make their voices heard. I mean, the Austrian government is nice, but where is the German government? The Italians? The French? And how about our friends from Hungary?

Created: 12.04.05, modified: 25.05.07