The second part to my SolidFire Element OS 8 update – code name Oxygen – is around our Protection enhancements.
Continuing the musical theme for these two blogs, todays feature album is the Massive Attack 1994 Album “Protection.”
5+1 ways to replicate your data
During our internal Oxygen training, one of the review questions was “How many different methods are there are replicating data in SolidFire ?”
The answer to that question is actually 6. The methods we support, with the new methods highlighted in bold are indicated below:
- Snapshot only replication
- Async Replication
- Async with Snapshots
- Synchronous Replication
- Synchronous Replication with Snapshots
- Bonus Answer : Bulk replication can be implemented between SolidFire systems : more here http://www.stackparty.com/the-death-of-backup-and-a-new-way-to-protect-your-data/
So how do these look in practice?
So only three options here – Snapshot only replication is triggered when a snapshot is taken and we will look at that in a moment. Both the other methods of replication are real-time with Synchronous having a full three way commit before we acknowledge the write. You can also move between all of the methods as needed – using API calls to change on a schedule if needed.
With SolidFire replication, clusters and then volumes are paired together to ensure access is secured. Above we have the cluster pairing already in place and we set up the initial volume with just snapshot replication.
Our way, your way : “Whatever’s comfortable”
For me the “so what, who cares” is the at you can choose SolidFire replication and we give you all of the methods likely you will need to protect your customers or your companies data. You can then easily move between them. All with no license costs.
If another methods comes along that suits better then thats great. If you database, application does a better job of being more manageable then you can take that method. Otherwise our methods are there for you.
You can productise these features and add them to your service catalog. They work fully alongside QoS, with APIs, with full efficiency features on. Change between the methods as you need. Even go open backup and replicate into S3 or SWIFT.
Ultimately, as per the Southern Comfort ad – “Whatever’s Comfortable”
In my dreams I used to see the old NetApp A-B-C A-B-C’ diagram from explaining snapshots so many times. I think the world is now pretty used to efficient pointer based snapshots so I will save you the trouble and me the dream.
Customers have however requested a proper scheduling & retention mechanism for snapshots. You can schedule flexibly and set retentions. You can de-couple snapshots and replication as needed then take snapshots off to open protocol backup stores when you want. How does this look? Firstly, we hit “Schedule Snapshot”
Then create both your schedule and set retention. Here you can also include the snapshot inside the replication process.
Finally you can view and edit your schedule:
Protection : your way – our way – Whatever’s Comfortable…