XetHub uses a patent-pending novel content-defined chunking algorithm to represent all the files in a Xet repository. Using this representation, XetHub is able to identify duplicate parts of files in the repository and only store them once. This reduces the total storage needed in XetHub server storage for the Xet repository and reduces the amount of data that is transmitted over the network when working with this Xet repository.

XetHub only charges for the deduplicated server storage & transfer. As a concrete example, imagine a 100GB CSV file that is updated every day. Every day 1GB in this file is changed (maybe the file keeps 90-day trailing information). This file is then split into 3 parts (train, val, test) as separate files. With XetHub only ~1GB is added to the Xet repository every day, and that is all that is billed.

You can read more about how XetHub deduplication works here or in our CIDR’23 paper Git is for Data.

XetHub only charges for GB transfers outside of our server data center(s). This is often called egress costs. XetHub measures every GB transferred out of our data center(s) for each Xet repository you own (you can see this in the Settings→Usage tab for each Xet repository). Once the transfer exceeds the plan limit, each GB is charged at the listed rate.

Volume discounts are applied in a tiered basis, so as your use of XetHub increases the price per GB stored & transferred is charged at a lower rate. Volume-based discounts are only available to Pro plan customers. As usage crosses a tier you will be charged the lower tier price for usage up to the next tier, and the next tier’s price for subsequent usage.

A concrete example: As a Pro plan customer, on May 1st you have 125GB stored with XetHub. Between May 1 and May 31 you add 130GB of storage with XetHub. So at the end of the month your total storage is 255GB. The Pro plan charges $0 for the first 50GB, then charges $0.08/GB for the next 200GB, and $0.05/GB over 250GB. So, for the month of May, XetHub would charge you $0 (50GB) + $16 (200GB*$0.10) + $0.25 (5GB*$0.05) = $16.25 for Data Storage.