There’s a question which we’re often asked here at Syneto: Isn’t Storage OS powered with Illumos, with the ZFS filesystem comparable to a Linux powered storage system with Logical Volume Manager, RAID protection and the Ext4 filesystem?
Frankly, no. And this is not just marketing talk. In this article we’ll study the technical feature differences that set Storage OS apart from a Linux data storage solution.
There’s a bunch of storage companies out there that are dishing out content telling you that a hardware RAID is better. While that might be true in some instances the software RAIDz is perfectly suited for a unified storage environment. Why? Because it’s far more cost effective. You don’t need to spend a ton of money on controller hardware which is very expensive. And secondly, it is far more easier to use. No need to spend a lot of time on configuring hardware controllers. It’s a win-win situation for both the vendor and the client.
RAIDz also eliminates the infamous RAID 5 write hole where data becomes inconsistent after a power loss. This means you don’t need to add expensive batteries to your storage.
The ZFS filesystem on Storage OS writes every data structure in the byte order of the machine using it. More importantly, and here’s the breakthrough, Storage OS flags the byte order it uses. This results in the ability to swap disks between machines while maintaining usability.
This just isn’t possible with the Linux Ext4. If the machine the disk is inserted into doesn’t have the same level of endianness the data won’t be usable.
With ZFS importing or exporting data, a big help when trying to achieve a consistent Disaster Recovery strategy, is extremely easy and fully automated. This means the sending or receiving data can be done in just a few clicks. Less time and money!
Truthfully, this can also be achieved on the comparable Linux device, but it’s incredibly complicated. All the required steps must be done manually.
Storage OS uses iostat and dtrace to natively provide automatic, accurate IO usage. This means you can get information on your storage’s activity by quickly glancing at a few interactive graphs
This important feature is not automated on Linux. Everything needs to be laboriously done from the terminal.
On-the-fly Backups & Recoveries
The copy-on-write transactional model used by the ZFS in Storage OS can execute data Backups and Recoveries while the system is running. This means scheduled downtime for backups is a thing of the past. Backups on Storage OS are incremental, an aspect which allows you to back up data in mere seconds.
Backups and Recoveries on the comparable Linux device requires the system to be offline. This means that you’ll inevitably loose time and money every time you need your data saved.
Space savings, Clones and Rollback
For an organisation that is required by law to maintain extensive data archive, the size of a Backup is very important. Storage OS backups, with the right configuration, can use less than 1% of the total space. Rollbacks and Clones are also easily achieved due to the incremental model and the copy-on-write transactional model. There’s no need for the system to go through every change ever made in the data to achieve point-in-time consistency.
With Linux, on the other hand, backups take a lot more space. Sometime as much as the saved data itself. The system also needs to run through all the actions the data has taken to rollback to a point, let’s say, 4 months in the past.
Disk failures & Corruption
Here’s were things get interesting. While both ZFS and Linux can handle whole disk failure, Storage OS checksum’s all the data. This means that it can also handle disk corruption, and fix it. On Linux, corrupt data is lost most of the time. This self-healing capability is unique to ZFS.
Automatic Fault Notification
If something happens the sys. admin doesn’t necessarily have to be in the same room as the appliance to notice the little red LED (if there is one). Storage OS is equipped with a notification system that automatically email’s you the problem. If it’s a data corruption issue, the system sometimes heals itself and just informs you that the event took place.
Built-in Storage Efficiency
Storage OS features Data Compression, Deduplication and Thin Provisioning as default. These will help you store more data in the same space, for the same money and with a far greater level of control.
Curios for more? Download our Storage OS vs. Linux paper to see a full feature comparison.Download document