How versioning works
Understand how Crunch works with versioned objects.
Object versioning (from both Amazon S3 and Google Cloud Storage) is a great way to provide data protection in case of accidental deletion (or overwrites) as well as for use cases such as file sharing, collaboration etc.
However, each versioned object is a full object and consumes just as much storage space (and generates the same monthly cost) as if it was a non-versioned, net-new object. Your total storage bill is thus multiplied by the number of versions, e.g. 2 versions costs twice as much, 10 versions costs 10 times as much etc. As a result, object versioning can quickly become very expensive and many teams are understandably hesitant to use it.
Flattening the cost curve
Granica Crunch flattens the object versioning cost curve so you can add versions cost effectively. And the more versions you use, the more you save. Crunch supports versioning in AWS today*, with GCP support Coming Soon. Here’s a simple graph to illustrate the dramatic difference in cost scaling that Crunch delivers:
We see that without Crunch your monthly cloud object storage costs scale 1:1 with object versions. But with Crunch your monthly costs scale much more slowly (details below). With storage costs now under control you can finally take full advantage of object versioning to enhance your existing applications and/or build innovative new ones.
How it works
The main driver behind flattening the cost curve is that object versions tend to be pretty similar to each other, i.e. the true content delta between objects is often very small. In our customer deployments we’ve seen deltas as low as 5%, meaning versioned objects can be up to 20x more reducible than non-versioned objects. Of course the magnitude of the deltas vary based on your data and your use cases.
As as a result, versioned objects are ideal candidates for data reduction with Crunch. Up to 95% of the data is trivially reduced away, and the remaining (at least) 5% of delta content is then reduced using our patented machine ML-based reduction algorithms, which themselves typically yield 25-60% reduction rates. When you put this all together, with Crunch your object versions consume as little as 2% (5% * 40%) of their original storage.
No configuration required
Crunch detects when you are using object versioning and automatically adjusts accordingly. Crunch just works - there is no configuration or management required, just continue to read and write your data through using the Granica S3-compatible API.
Crunch will automatically disable recycle bin (if previously enabled) for a given bucket when you start using object versioning.
Summary - it’s time to go big
In our view cloud object versioning has been an underutilized capability given its cost challenges, and with those costs now mitigated with Granica Crunch we’re excited to see the cool things you’ll do with it.
* Lifecycle tiering policy is currently not applied to objects with versioning enabled. Support for concurrent lifecycle tiering and versioning policies is Coming Soon.