Active Directory for Linux draws closer

David M Williams
Friday, 30 January 2009 18:56

IT Industry - Market

Then, the number of people in that office grows even just by a couple more people. Both you and they are concerned about loss of data from files stored on their local hard drives. The single USB printer plugged in to the reception computer and shared on the network is becoming a frustration because people forget that computer has to be turned on to print.

It’s time for a server in the local branch.

Now, you could buy a Windows server; you’re looking at over $1,000 on top of the hardware to pay the software license.

Some might suggest Microsoft Small Business Server (SBS.) Yet, that product is unrealistic for companies beyond a certain size, particularly if there are other servers on the network and especially if there’s already an SBS server anyway.

Frankly, you might as well suggest a desktop operating system like Windows XP, that’s how low SBS is in my estimation. Sadly, I have in fact seen companies where the “server” runs Windows XP. Pirated copies at that.

I have visited organisations that knew something was needed – a 24x7 always-on storage device that could be remotely backed up over the network and accessed by everyone in the office – yet they didn’t want to pay for a proper server. Instead, they went for a dual hard-drive Network Attached Storage (NAS) device like the D-Link DNS-323.

On the one hand, this device is cheap and it does offer RAID 1 to give some measure of redundancy and disaster recovery. It’s also small and easy to configure.

However, it doesn’t join to the company network. To use this device you need to make file shares on it that have their own entirely disparate set of usernames and passwords. The users in the office have to authenticate against this device using an entirely different set of credentials than they used to log in to Windows. If staff come and go you need to remember to manage this box as well as your central AD. Further, it’s trivial to reset a unit like this, even assuming the default admin password is changed.

I’m sure the problem here makes sense to you. Let’s scale it up; instead of a tiny NAS you’re back to a full-blown server system – but with Linux on it instead of Windows.

There’s a cost saving here, and Linux can effortlessly handle file and print operations. And DHCP and DNS and pretty much anything you want to throw at it. It’s also a snap to manage by your IT staff who are physically located elsewhere in the country.

Yet, we hit the same problem; unless the Linux installation can fit into the existing network directory structure your fileshares are just as cursed. Unless you can secure shares using groups defined within AD, unless your users can log in to Windows once – on any desktop – and access the fileshares with no ancillary login required, unless your DNS can replicate to and from this server you have not implemented a genuine solution. Instead you’ve gone from having one problem to having two problems!

There are no end to scenarios where you might rationally and reasonably insert Linux into an existing Windows infrastructure. You can use it as a firewall and anti-virus platform on the outer perimeter of your network where the Internet comes in. Linux is particularly attractive here because it is not affected by the legion of Windows malware that may attempt to make its way in.

So, what are we going to do about it? I'm glad you asked!