The Pansynchro Challenge

Pansynchro is designed from the ground up to be as fast as possible.

Now, anyone can say “our product is fast.” But at Pansynchro, we’re willing to stake our reputation on it. That’s why we’ve come up with the Pansynchro Challenge: if anyone can come up with a reproducible scenario where a competing product beats Pansynchro, we’ll post on our website that this product beat our times, and make it a top priority to fix that.

Rules

Just a few simple ground rules to keep everything fair and reasonable.

  1. This must be a data sync job involving between 1 GB and 100 GB of input data. If it’s just a small amount of data, other factors might be more significant than the sync performance; if it’s a huge amount of data, it’s likely that your point could be demonstrated on a somewhat smaller scale.
  2. It must be easily reproducible on our end. Share your input data source, or the scripts used to generate the data. Share your PanSQL scripts that you used to set up the Pansynchro job, and the setup instructions to prepare the competing job. And we need to be able to set up the competing system without any excessive monetary costs, of course.
  3. It must run on commodity hardware, with a minimal amount of instances required. We’re not interested in going head-to-head with things like SSIS’s “ETL world record,” which from their writeup relied far more on throwing a lot of carefully-configured hardware at the problem than on any inherent high performance characteristics of SSIS.
  4. Both jobs must perform the same work. They should start from the same source data and end with the same result data. The challenge is in how they get there.
  5. Both jobs should actually run to completion. If you can’t actually do the work in Pansynchro, that’s not a performance issue; that’s a bug report or a feature request. Feel free to file those on our Github page.
  6. The speed of the sync job is what’s being measured. The clock starts when the sync begins and stops when it ends. Setup time doesn’t matter, including factors like preparing the optimized Pansynchro data dictionary. (While this may seem at first glance like a special case designed to advantage our process, it makes sense when you consider that most jobs run many times on a regular schedule, while setup only has to be done once.)

And that’s it! If you can come up with a worthy challenger, we’d love to hear about it! Contact us with the information and we’ll look it over.