EMC talks up the ix4-200d as a desktop NAS, but really, you wouldn't want it on your desk. The device's sporadic chatter can be distracting. While I was building a 100GB iSCSI LUN (which took hours), the ix4-200d sounded like it contained a horde of ravenous chimpmunks eating acorns, though deleting an iSCSI LUN was quick and quiet. The ix4-200d does has power management features, such as dimming the bright blue LED panel and powering down the disks while idle.
The device uses a simple Web-based user interface, and the purchase price includes one year of TZO's Dynamic DNS so that you can access the web interface and files remotely. The StorCenter walks you through the TZO Dynamic DNS registration, where you choose a name for your StorCenter, and it registers under the Iomegalink.com domain. When you are on the road, all you have to do is open HTTPS on your firewall/router to get access to the NAS.
The ix4-200d has simple file security features to manage access to folders. These are okay for most situations, but an advanced set of options would have more robust access controls. We could assign directories as read only, read/write, and no permissions to users, but we couldn't enable certain protocols for specific directories. For example, FTP users are able to access all directories. There isn't a way to set up anonymous FTP, either. Arguably, SMBs may not be running anonymous FTP servers to the Internet, but FTP is still used for file transfer. It would have been nice to have more control over what protocols were assigned to shares.
Performance wise, we found the is4-200d fairly fast for both CIFS and ftp file transfers. We used Iometer 2008.06.18 for performance testing. We connected the ix4-200d using a single 1Gb Ethernet to a Cisco Catalyst 3750G. Our client device was a Dell Optiplex 745 running Windows XP. We monitored the network traffic using a Fluke Networks OptiView Series III Network Integrated Analyzer. For the iSCSI testing, we created a 100GB LUN on the ix4-200d. For CIFS and FTP, we used an existing share. We used the 32Kb Read test from Iometer, which specified 32KB sequential reads and well as the 32KB 0% Read (all sequential writes). We also set the outstanding IO on Iometer to ten to maximize performance. Beyond ten, we didn't see much performance gain.