Well, it sounds great, but I'm really concerned about how easy this is going to be to work with.
I love the fact that there's an audit mode so I can discover what software is in use, but from everything I've seen and read, there are only two possible ways I can create a rule for an item. I can either install the software I want to authorize on my network admin computer, or I can install the remote server administration tools on the client machines that have the software installed.
One of the suggestions I've read from Microsoft (and which is mentioned in the video) is that I create template machines, and use those to configure AppLocker rules before deploying them to our network.
The problem is that simply doesn't scale as an admin model when it's applied to existing networks. We have over 143 applications in use on this network, on no fewer than 34 separate departmental setups. And that's not even counting the 120+ utilities we have on the IT computers to help us manage the network.
There is absolutely no way I'm going to install another 143 applications on this computer. Nor do I feel particularly inclined to create model clients for 34 separate departments. We already have group policy objects in place to deploy most of our software, and scripts for the few items that are tricky (namely DirectX and .NET 3.5). That work has already been done, why on earth would I want to repeat it?
Also, if I do go through the process of manually configuring all these programs, what if I miss one program? Instead of being able to simply authorise it, I have to rebuild a spare machine with the template for that department, install the remote admin tools on that computer, and then authorise it.
Contrast this to other programs I've seen, such as Winternals Protection Manager (now owned by Microsoft). You had the ability to react instantly to alerts from users, and immediately authorise that program. If AppLocker has an audit mode, that should log *all* the information the server needs for me to be able to create rules based on any audited application. There should be no need to do *anything* on the clients. Authorising a blocked program should be as simple as browsing to the alert on the server, right-clicking on it and creating a rule.
In short, if I'm going to run an AppLocker server, I want to be able to administer it from the server. Not from the clients.