How to manage “Bring your own device”
A few years back a customer’s CSO left the room when I said that this customer should start thinking about a scenario, where selected users bring their own devices – he called me “nuts”. Well, I think the smartphone area proofed me right. Basically the smartphones were the first Bring Your Own Device (BYOD) as far as I can tell. The CEO found the latest and coolest phone on Saturday and on Monday IT had to integrate it – not to say that the CEO definitely wanted to read the mail on his/her new toy.
This discussion is over since a long time and most people probably accepted the fact that the world changed – the cheese moved. BYOD, Consumerization of IT or however you want to call it at the end of the day is a reality. They might have different forms: In our case at Microsoft it might be officially a pre-stage as internally we get the hardware but we can set it up the way we want as long as we are following the policies. But even this is not the complete truth as there are a lot of people buying their own hardware and using it to work. I am currently not only running my notebook with Windows 7, I am using Windows 8 Developer Preview on a slate as well – and as I want to understand how we can make it happen – I did not join it to the domain as I want to run the Consumerization of IT scenario. This immediately raises questions on security.
We most probably need mail (Outlook in my case), Lync and some documents on a slate. So, I need to have Outlook installed and connected to Exchange (including RMS-protected mail), Lync as well as OneNote and some documents I want to have with me while I am travelling. What does this mean for IT? What about me connecting to the corporate network? Let’s look at some of the scenarios and functionalities. I know that there are answers to some of the problems but lets look at the questions first:
- Authentication: As it is not a device IT controls, how is the user authenticated? So we might want to require a PIN or a password to unlock the device. This makes sense anyway but there needs to be more than a “only” a paper policy. For those of you who have seen the build presentations on Windows 8 might have seen a new way to authenticate: A user can have a picture and store three gestures to unlock. A great way to authenticate to a slate but does the policy allow for that? Even if it is not a domain authentication, it is the authentication to the holy grail – the mail.
- Lost devices: Typically these devices are cool – that’s the reason why our users buy them – no? So, the risk of them getting stolen - or lost as they are small - is fairly high. How is the data and how are the credentials on the PC protected? So, we talk of disk encryption first, remote wipe second.
- Disk Encryption: There are devices like Windows Phone 7, which have a very sound security model and a very good device security but unfortunately no encryption, yet. There are others with “encryption” built in, which is broken in minutes as the device can be jail broken easily. What is the policy there? On the slate there will be a need for disk encryption as well. Which user will use something like this without being told? Yes, I know. You will but you are definitely not a representative sample as security people. On Windows we can switch Bitlocker on and will have at least the ability to securely protect the disk.
- Wipe: I would want my device to be wiped after a few unsuccessful authentication attempt or – if I lose it – I want to be able to remote-wipe the business data if I am IT.
- Network Access: Now the device comes on our network. What happens if the devices does not have any anti-malware protection? It might spread all the dirt on your network. Not something we typically enjoy. There are solutions to that – since a long time we talk about Server and Domain Isolation Using IPsec and Group Policy which at least separated the trusted and the untrusted devices. But we basically want the devices on the network and have them accessing the data – if they follow certain policies. Therefore we need a way to do policy enforcement and health checks with the ability to quarantine.
- VPN Access: This might be easier as we can enforce the policies as mentioned above much easier as the machines come through a well-defined channel where we can check them but are we allowed to? Think about privacy implications as well.
- Mail: Finally talking of mail. Access to e-mail is probably one of the crucial areas to enable and manage as a lot of confidential information is buried somewhere in mail. Additionally, to access mail, the keys will be needed if the mail is encrypted. Thus a lot of critical information is on such a device.
- Data: As a user I want my data (or at least key part of my data) synced between my devices. In my case between the business notebook and my slate. This should be done in a secure and safe way. Do we as IT want to allow the use of technologies like Live Mesh, which can either do a peer-to-peer synchronization or a peer-to-peer-to-Skydrive sync. In other words, a copy of the data can be hosted in the public cloud secured with a LiveID password.
So, a lot of different problems/questions. However, they are only partly new as I have seen a lot of people taking data home to their own private PC – the one the kids are gaming on – to do their work. Taking home means USB or even sending the data to the private mail account.
Protecting such an environment can have different approaches and I would be interested in what you think and what you need:
- First and foremost we need policies clarifying what can be done and what not. For severe violations, there needs to be disciplinary action.
- We want to have some policy enforcement. Basically, the key functionality the user is interested in is often e-mail and therefore Exchange might be one of your key management point for this. Exchange is basically able to enforce the following policy options to your device (from Understanding Exchange ActiveSync): Remote Wipe, Device Password Policies (minimum length, characters, alphanumeric, inactivity time, enforce history, enable recovery, wipe device after failed attempts), device encryption. Therefore, it can be expected that the key requirements can be met. But there is a fair chance as well that not all devices fulfill all the requirements. Or even worse: The active sync client could simply lie to the server.
- Would it be an option for an IT organization to require a client installation? Would the policy “if you want to use your own device, you have to let us install a piece of software” something which can be implemented? I am not completely sure are the user will look at the device as his/her own and will refuse interference. On the other hand it is the company’s data. A fairly interesting conflict. If we are allowed to install a client, all of a sudden technologies like Network Access Protection become feasible as we have a trusted piece of software being able to check the health of a computer
But what else is needed? Do you need management? Inventory? What else would you expect in such a scenario from your technology? Let me know – I am interested in this debate.
Roger