Review: Synology DS-101 Disk Station

Under the Covers

Synology's documentation reveals that the processor in the unit is the Intel ixdp425 - the same as used in the Linksys NSLU2. However, the DS101 has 64 megaBytes of RAM, which is twice the amount in the NSLU2. It is also more than found in other similar products and should help out disk caching performance and large print job processing.

Figure 9: DS-101 board
(click image for larger view)

From the description of the box and its capabilities, it was pretty clear that it was running Linux - just like many of the other devices in its class. And a quick look into the flash image provided on the Synology web site confirmed my hypothesis. Extracting various components in the image revealed typical embedded Linux processes such as busybox, samba and thttpd. By the way, a search of the Synology web site had no references to available GPL source code, which is required by the GPL license, but Synology said they are in the process of correcting this.

One somewhat unusual item I noticed while exploring the DS-101 flash image was the use of the Reiser filesystem. I had not previously seen this filesystem used in a consumer NAS device, so decided to do some additional exploration. I took the IDE drive out of the DS-101 and hooked it to my Desktop Linux system so that I could explore the filesystem. My first attempt to simply mount one of the partitions failed, with an error indicating that the partition was not in Reiser format!

Next, I did a simple ASCII dump of the first section of the partition to see if I was mistaken about the filesystem used. This is where I saw something curious. A standard Linux filesystem will have an ASCII signature present to tell the operating system what kind of filesystem is present. For Reiser, this will normally be "ReIsEr2Fs", but on the DS-101 initialized disk, I saw an "SyNoRs2Fs". Reading between the lines, this indicated to me that perhaps this was a "Synology Reiser2 Filesystem." Had Synology created a custom Reiser2 Filesystem? Or maybe they had just put in a custom signature.

To test this theory (don't try this at home kids!) I bypassed the filesystem by writing directly to the raw hard drive and overwrote the signature with the standard one. After I did this, I was able to mount the filesystem without error and I could see the top-level directories on the disk. But when I tried to examine any of the files, I would get an error indicating that the file didn't exist. This told me that it really was a modified Reiser filesystem, and not just a change in signature. For the time being this was as far as I was willing to explore, so I put the original filesystem signature back and re-assembled the DS-101.