As the Internet of Things continues to expand (and it is expanding rapidly) the issue of access to applications is going to get a lot more messy.
One of the benefits of cloud, so I'm told, is it is "easy access" for everyone. Indeed, it matters not if you're on a mobile device or a tethered one, you can access stuff in the cloud easy peasy, lemon squeezy.
Which I've always thought while true, wasn't exactly what enterprises needed or wanted. Unfettered access (allow all) to applications is not exactly the security policy most organizations put into place. Authorized access, certainly, but an open door policy? Not that I've seen lately.
Which is why the Internet of Things and the migration of more applications (certainly not all, but many more) to publicly accessible cloud environments that provide very little in the way of access control is going to collide and make things messy for a while.
The problem is that because we're just in the early days of the Internet of Things, we haven't really got a comprehensive strategy for matching applications, access needs and deployment environments with the "things" we're going to be monitoring, measuring and messaging with. That means a single "thing" might need to communicate with an activation app in the data center, but manage data through an app in a public cloud. It might download updates through yet another service hosted in another public cloud. And it might need to connect with other "things" that might be located on the same desk, in the same house, or in completely different countries.
Don't try to do the math and come up with the possible combinations you'd need to lay out in order to put together a traditional application access control policy. It'll just make your head spin. Of course if you like that feeling go ahead, no one can see you you know*.
The thing is that our security controls are still based on an IP world, where every person or device needs to be matched to an IP address so we can match that against a list and decide "yes" or "no" to access.
If BYOD didn't teach us that's not feasible, the Internet of Things will.
Bob doesn't just have a laptop and a phone. Now he also has a smart key, a smart car, and a smart watch. They're all "him" but yet they aren't. One might like to assume that if they're all coming out of the same network (the house network) over WiFi, then they all map back to the same public IP address cause, NAT works that way. And maybe that works well enough when every app and service needed is behind the corporate firewall. But they're not anymore. They're in the cloud, too, and across the Internet.
We can't continue to craft firewall and access rules based on IP addresses. Not feasible and ultimately, it's not secure or accurate enough. We need ID-based access rules that not only consider who but what. Not just Bob, but Bob's phone. Not just Bill, but Bill's refrigerator. Not just Alice, but Alice's television.
Identity will have to expand to include the notion of "ownership". Each of us becomes a "group" unto ourselves, with individual smart things and apps being a part of that group but having their own sub-identity and thus, access rights and constraints.
An application world is both about the apps that run those smart things (just because you can't see them doesn't mean they aren't there) and the apps that manage them (just because you can't see them in the data center / cloud doesn't mean they aren't there). We're going to need more flexible and dynamic means of determining not only who but what can access each of them at any given time. The pressure on identity management and access services is going to be incredible, because it's going to have to be the new perimeter. A traditional IP-based perimeter just isn't going to be enough to meet the new requirements for application delivery.
* Or maybe they can, you never know. Did you leave your camera uncovered?