While I was attending VMworld 2015, I attended the Reduxio presentation at Tech Field Day. I did some preliminary research on this company prior to the presentation and to be honest I did not completely understand their use case, so I was very interested in hearing more technical content from their engineers.
For the presentation, the company had Eyal Traltel (Director of Technical Marketing), Jacob Cherian (VP of Product Strategy and Product Management) as well as Nir Peleg who is the founder and CTO on site.
Reduxio has four primary backers. Seagate, Intel, JVP and Carmel (which is part of Viola Group).
At first I thought to myself: Ok, here’s another hybrid array but I quickly saw the value-add with the type of protection that they’ve implemented within this midrange storage system.
On the first slide of the presentation they had a statement that read, “A new way to protect data” which immediately sparked my interest. Here is a storage company that put data protection on the top of their list!
The Reduxio HX550 consumes 2U of rack space and has the ability to hold 24 drives. They mentioned that generally they have 16 HDD’s and 8 SSD’s in a standard build.
The HDD’s are 7200 RPM drives and the SSD’s are Western Digital eMLC SAS drives. The unit holds 40TB of raw data, but Reduxio said that capacity would be higher when data reduction is factored in. Also, it is a dual controller system with dual 10GB ports on each controller.
This is very different from what others are doing and is what makes this storage array so unique. TimeOS™ tracks every read/write operation (instead of allocating space) which enables the volumes for the following technologies:
This is a reactive tiering operation that operates at the block level, is automated and is continually running. It is designed to place data on flash first and moves infrequently accessed data to slower magnetic disks through the analysis of 8k blocks.
This is Reduxio’s method of crash consistent protection for applications on all LUN’s created and provisioned in the unit. It allows for recovery within one second of the failure anywhere in the volume. Essentially, giving you a view to the volume “the way it was” before the failure occurred. I did ask some specific questions revolving around always on clustering from Microsoft and even though they haven’t tested it to date, it is on their radar.
This is their deduplication & compression, it is a global setting and is always running in the array on all cache tiers, volumes and clones, as well as current and historical data. It is also real time and in-line!
The entire management interface is in HTML5 and during a demonstration of its operation, it was very smooth and fluid.
- All volumes are protected by default through policy based retention.
- The vCenter plugin has a nice look and feel to its operation.
- To date, the system does not scale out but they have that in mind for the future since it is designed for scale-out addressing.