There was an excellent article at Dark Reading the other day about data leaks focusing on insider threats. It did all the right things by pointing out “insiders have access to critical company information, and there are dozens of ways for them to steal it” and “these attacks can have significant impact” even though “insider threats represent only a fraction of all attacks–just 4%, according to Verizon’s 2012 Data Breach Investigations Report.” The article goes on to discuss how you can use gateways, DLP for at rest and in flight data, behavioral anomaly detection, and a few other technologies in a “layered approach using security controls at the network, host, and human levels.” I agree with every word.
Yet, there is one aspect of the controls that somehow escapes mention – letting a potentially powerful ally in this fight off the hook from any action. There is not one mention of proactive controls inside the applications and platforms that can be placed there by IAM. A great deal of insider access is inappropriate. Either it’s been accrued over time or granted as part of a lazy “make them look like that other person” approach to managing entitlements. And app-dev teams build their own version of security into each and every little application they pump out. They repeat mistakes, build silos, and fail to consume common data or correctly reflect corporate policies. If these problems with entitlement management and policy enforcement could be fixed at the application level, the threats any insider could pose would be proactively reduced by cutting off access to data they might try to steal in the first place. It’s even possible to design a system where the behavioral anomaly detection systems could be consulted before even handing data over to a user when some thresholds are breached during a transaction – in essence, catching the potential thief red handed.
Why do they get let off the hook? Because it’s easier to build walls, post guards, and gather intelligence than it is to climb right inside of the applications and business processes to fix the root causes. It’s easier to move the levers you have direct control over in IT rather than sit with the business and have the value conversation to make them change things in the business. It’s cheaper now to do the perimeter changes, regardless of the payoff – or costs – later. Again, this is not to indict the content of the article. It was absolutely correct about how people can and very likely will choose to address these threats. But I think every knows there are other ways that don’t get discussed as much because they are harder. In his XKCD comic entitled “The General Problem,” Randall Munroe says it best: “I find that when someone’s taking time to do something right in the present, they’re a perfectionist with no ability to prioritize, whereas when someone took time to do something right in the past, they’re a master artisan of great foresight.” I think what we need right now are some master artisans who are willing to take the heat today for better security tomorrow.
No one knows how to make a big proclamation in the identity world like Kim Cameron. His keynote at #eic10, the Kuppinger Cole European Identity Conference for 2010, was no disappointment. Kim reviewed his ideas for the “Federated Interscaler Directory”, which was often misquoted as saying “Interstellar”. The basic idea was to “extend” the current ubiquitous Active Directory platform to hold a more flexible framework for relationship expression, policy enforcement and other elements that directories of today are missing. While adding all that, this new directory platform should also scale, in the sense that it could administer millions of identities, as well as support advanced features like federation, token translation and other things that are clearly becoming part of next gen identity.
On it’s surface, that all sounds nice. But it also sounds dangerous to me. One other theme at #eic10 throughout many talks, and something Kim even said during his, was that we shouldn’t want identity systems to be monolithic (he said so in reference to the ability to federate with other IdP’s outside the directory itself). But the system Kim described and the picture he used to illustrate it looked pretty monolithic to me. A lot of what he described is possible today already with a loose federation of platforms from many vendors and open source projects. You can enforce all the policy you need with a XACML authorization engine and properly tooled interfaces and proxies for applications and providers. You can manipulate schemas and the objects they serve up as needed with virtual directories. If Microsoft were to make AD into one big solution for all that, then the biggest differentiator would be having its monolithic status versus the loose coupling of many other components. I tend to be a fan of loose couplings, but I’ll keep the jury out until I see more from Kim.
One thing that I really liked was Kim’s call for everyone to work together on a common identity schema. It’s not the first time he’s done so. At PDC he made a great presentation that described the same idea in much greater detail [link to the PPTX Powerpoint file from PDC]. A project of this kind, if well done, could solve many, many interoperability and operational challenges in the identity world. So much time is spent now negotiating, either in research or in calls at run time, to figure out what attributes and properties of an identity are available. If there were a completely standard schema and a means to publish it easily, then that goes away.
I’ll have more thoughts from the conference later. For now I’m going to put on my space suit and leave the Microsoft ship and hope Kim hasn’t locked the bay doors when I get back.